Encyclopedia of Ocean Sciences, Second Edition, Volume 1: A-C

ENCYCLOPEDIA OF OCEAN SCIENCES SECOND EDITION Editor-in-chief JOHN H. STEELE Editors STEVE A. THORPE KARL K. TUREK...

Author: John H. Steele | Steve A. Thorpe | Karl K. Turekian

100 downloads 1655 Views 46MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

ENCYCLOPEDIA OF

OCEAN SCIENCES SECOND EDITION

Editor-in-chief

JOHN H. STEELE

Editors

STEVE A. THORPE KARL K. TUREKIAN

Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier

(c) 2011 Elsevier Inc. All Rights Reserved.

ENCYCLOPEDIA OF

OCEAN SCIENCES SECOND EDITION


Subject Area Volumes from the Second Edition Climate & Oceans edited by Karl K. Turekian Elements of Physical Oceanography edited by Steve A. Thorpe Marine Biology edited by John H. Steele Marine Chemistry & Geochemistry edited by Karl K. Turekian Marine Ecological Processes edited by John H. Steele Marine Geology & Geophysics edited by Karl K. Turekian Marine Policy & Economics guest edited by Porter Hoagland, Marine Policy Center, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts Measurement Techniques, Sensors & Platforms edited by Steve A. Thorpe Ocean Currents edited by Steve A. Thorpe The Coastal Ocean edited by Karl K. Turekian The Upper Ocean edited by Steve A. Thorpe


ENCYCLOPEDIA OF

OCEAN SCIENCES SECOND EDITION Volume 1: A - C Editor-in-chief

JOHN H. STEELE

Editors

STEVE A. THORPE KARL K. TUREKIAN

Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo Academic Press is an imprint of Elsevier


Academic Press is an imprint of Elsevier 32 Jamestown Road, London NW1 7BY, UK 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, CA 92101-4495, USA Copyright ^ 2009 Elsevier Ltd. All rights reserved

The following articles are US government works in the public domain and are not subject to copyright: Fish Predation and Mortality; International Organizations; Large Marine Ecosystems; Ocean Circulation: Meridional Overturning Circulation; Salt Marsh Vegetation; Satellite Passive-Microwave Measurements of Sea Ice; Satellite Oceanography, History and Introductory Concepts; Satellite Remote Sensing: Ocean Color; Science of Ocean Climate Models; Wind- and Buoyancy-Forced Upper Ocean. Fish Migration, Horizontal Crown Copyright 2001 Turbulence Sensors Canadian Crown Copyright 2001 No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher

Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44)(0) 1865 843830; fax (+44)(0) 1865 853333; email: [email protected]. Alternatively you can submit your request online by visiting the Elsevier website at (http://elsevier.com/locate/permissions), and selecting Obtaining permissions to use Elsevier material

Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein, Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

Library of Congress Control Number: 2009935908

ISBN: 978-0-12-375044-0

For information on all Elsevier publications visit our website at www.elsevierdirect.com

PRINTED AND BOUND IN ITALY 09 10 11 12 13 10 9 8 7 6 5 4 3 2 1


Editors

Editor-in-chief John H. Steele Marine Policy Center, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA

Editors Steve A. Thorpe National Oceanography Centre, University of Southampton Southampton, UK School of Ocean Sciences, University of Bangor, Menai Bridge, Anglesey, UK Karl K. Turekian Yale University, Department of Geology and Geophysics, New Haven, Connecticut, USA


v

Editorial Advisory Board John H. S. Blaxter Scottish Association for Marine Science Dunstaffnage Marine Laboratory Oban Argyll, UK Quentin Bone The Marine Biological Association of the United Kingdom Plymouth, UK Kenneth H. Brink Woods Hole Oceanographic Institution Woods Hole MA, USA Harry L. Bryden School of Ocean and Earth Science James Rennell Division University of Southampton Empress Dock Southampton, UK Robert Clark University of Newcastle upon Tyne Marine Sciences and Coastal Management Newcastle upon Tyne, UK J. Kirk Cochran State University of New York at Stony Brook Marine Sciences Research Center Stony Brook NY, USA Jeremy S. Collie Coastal Institute Graduate School of Oceanography University of Rhode Island South Ferry Road Narragansett RI, USA

Paul G. Falkowski Departments of Geological Sciences & Marine & Coastal Sciences Institute of Marine & Coastal Sciences School of Environmental & Biological Sciences Rutgers University New Brunswick NJ, USA Mike Fashamw Southampton Oceanography Centre University of Southampton Southampton UK John G. Field MArine REsearch (MA-RE) Institute University of Cape Town Rondebosch South Africa Michael Fogarty NOAA, National Marine Fisheries Service Woods Hole MA, USA Wilford D. Gardner Department of Oceanography Texas A&M University College Station TX, USA Ann Gargett Old Dominion University Center for Coastal Physical Oceanography Crittenton Hall Norfolk VA, USA

Peter J. Cook Australian Petroleum Cooperative Research Centre Canberra, Australia

vi

Robert A. Duce Departments of Oceanography and Atmospheric Sciences Texas A&M University College Station TX, USA

w

deceased


Editorial Advisory Board

Christopher Garrett University of Victoria Department of Physics Victoria British Columbia, Canada

Lindsay Lairdw Aberdeen University Zoology Department Aberdeen UK

W. John Gould Southampton Oceanography Centre University of Southampton Southampton UK

Peter S. Liss University of East Anglia School of Environmental Sciences Norwich, UK

John S. Grayw Institute of Marine Biology and Limnology University of Oslo Blindern Oslo, Norway

Ken Macdonald University of California Department of Geological Sciences Santa Barbara CA, USA

Gwyn Griffiths Southampton Oceanography Centre University of Southampton Southampton UK

Dennis McGillicuddy Woods Hole Oceanographic Institution Woods Hole MA, USA Alasdair McIntyre University of Aberdeen Department of Zoology Aberdeen UK

Stephen J. Hall World Fish Center Penang Malaysia Roger Harris Plymouth Marine Laboratory West Hoe Plymouth, UK Porter Hoagland Woods Hole Oceanographic Institution Woods Hole MA, USA George L. Hunt Jr. University of California, Irvine Department of Ecology and Evolutionary Biology Irvine CA, USA William J. Jenkins Woods Hole Oceanographic Institution Woods Hole MA, USA

w

deceased

W. Kendall Melville Scripps Institution of Oceanography UC San Diego La Jolla CA, USA John Milliman College of William and Mary School of Marine Sciences Gloucester Point VA, USA James N. Moum College of Oceanic and Atmospheric Sciences Oregon State University Corvallis OR, USA Michael M. Mullinw Scripps Institution of Oceanography Marine Life Research Group University of California San Diego La Jolla CA, USA


vii

viii

Editorial Advisory Board

Yoshiyuki Nozakiw University of Tokyo The Ocean Research Institute Nakano-ku Tokyo Japan

Ellen Thomas Yale University Department of Geology and Geophysics New Haven CT, USA

John Orcutt Scripps Institution of Oceanography Institute of Geophysics and Planetary Physics La Jolla CA, USA Richard F. Pittenger Woods Hole Oceanographic Institution Woods Hole MA, USA Gerold Siedler Universita¨t Kiel Institut fua¨r Meereskunde Kiel Germany

Peter L. Tyack Woods Hole Oceanographic Institution Woods Hole MA, USA Bruce A. Warren Woods Hole Oceanographic Institution Woods Hole MA, USA Wilford F. Weeks University of Alaska Fairbanks Department of Geology and Geophysics Fairbanks AK, USA

Robert C. Spindel University of Washington Applied Physics Laboratory Seattle WA, USA

Robert A. Weller Woods Hole Oceanographic Institution Woods Hole MA, USA

Colin P. Summerhayes Scientific Committee on Antarctic Research (SCAR) Scott Polar Institute Cambridge, UK

w

Stewart Turner Australian National University Research School of Earth Sciences Canberra Australia

James A. Yoder Woods Hole Oceanographic Institution Woods Hole MA, USA

deceased


Preface to Second Print Edition The first edition of the Encyclopedia of Ocean Sciences, published in print form in 2001, has proven to be a valuable asset for the marine science community – and more generally. The continuing rapid increase in electronic access to academic material led us initially to publish the second edition electronically. We have now added this print version of the second edition because of a demonstrated need for such a product. The encyclopedia can now be accessed in print or electronic format according to the preferences and needs of individuals and institutions. In this edition there are 54 new articles, 67 revisions of previous articles, and a completely revised and improved index. We are grateful to the members of the Editorial Advisory Board, nearly all of whom have stayed with us during the lengthy process of going electronic. The transition from Academic Press to Elsevier occurred between the two editions. We thank Dr. Debbie Tranter of Elsevier for her efforts to see this edition through its final stages.

Preface to First Edition In 1942, a monumental volume was published on The Oceans by H. U. Sverdrup, M. W. Johnson, and R. H. Fleming. It was comprehensive and covered the knowledge at that time of the scientific study of the oceans. This seminal book helped to initiate the tremendous burgeoning of marine research that occurred during the following decades. The Encyclopedia of Ocean Sciences aims to embody the great growth of knowledge in a major new reference work. There have been remarkable new approaches to the study of the oceans that blur the distinctions between the physical, chemical, biological, and geological disciplines. New theories and technologies have expanded our knowledge of ocean processes. For example, plate tectonics has revolutionized our view not only of the geology and geophysics of the seafloor but also of ocean chemistry and biology. Satellite remote sensing provides a global vision as well as detailed understanding of the close coupling of ocean physics and biology at local and regional scales. Exploration, fishing, warfare, and the impact of storms have driven the past study of the seas, but we now have a great public awareness of and concern with broader social and economic issues affecting the oceans. For this reason, we have invited articles explicitly on marine policy and environmental topics, as well as encouraged authors to address these aspects of their particular subjects. We believe the encyclopedia should be of use to those involved with policy and management as well as to students and researchers. Over 400 scientists have contributed to this description of what we now know about the oceans. They are distinguished researchers who have generously shared their knowledge of this ever-growing body of science. We are extremely grateful to all these authors, whose ability to write concisely on complex subjects has generated a perspective on our science that we, as editors, believe will enhance the appreciation of the oceans, their uses, and the research ahead. It has been a major challenge for the members of the Editorial Advisory Board to cover such a heterogeneous subject. Their knowledge of the diverse areas of research has guaranteed comprehensive coverage of the ocean sciences. The Board contributed significantly by suggesting topics, persuading authors to contribute, and reviewing drafts. Many of them wrote Overviews that give broad descriptions of major parts of the ocean sciences. Clearly, it was the dedicated involvement of the Editorial Advisory Board that made this venture successful. Such a massive enterprise as a multivolume encyclopedia would not be possible without the long-term commitment of the staff of the Major Reference Works team at Academic Press. In particular, we are very grateful for the consistent support of our Senior Developmental Editor, Colin McNeil, who has worked so well with us throughout the whole process. Also, we are very pleased that new technology permits enhanced search and retrieval through the Internet. We believe this will make the encyclopedia much more accessible to individual researchers and students.


ix

x

Preface to Second Print Edition

In Memoriam During the creation of the Encyclopedia of Ocean Sciences and also in several cases prior to the publication of the electronic Second Edition, several Associate Editors or designated Associate Editors died. We specifically acknowledge their role in making this work an effective publication. They are Mike Fasham, John S. Gray, Lindsay Laird, Michael Mullin and Yoshiyuki Nozaki. J. H. Steele, S. A. Thorpe, and K. K. Turekian Editors


Guide to Use of the Encyclopedia

Introductory Points In devising the vision and structure for the Encyclopedia, the Editors have striven to unite and interrelate all current knowledge that can be designated ‘‘Ocean Sciences’’. To aid users of the Encyclopedia, this new reference work offers intuitive searching and extensive cross-linking of content. These features are explained in more detail below.

Structure of the Encyclopedia The material in the Encyclopedia is arranged as a series of articles in alphabetical order. To help you realize the full potential of the material in the Encyclopedia we have provided three features to help you find the topic of your choice.

1. Contents Lists Your first point of reference will probably be the contents list. The contents list appearing in each volume will provide you with the page number of the article. Alternatively you may choose to browse through a volume using the alphabetical order of the articles as your guide. To assist you in identifying your location within the Encyclopedia a running headline indicates the current article.

2. Cross References All of the articles in the encyclopedia have heen extensively cross referenced. The cross references, which appear at the end of each article, have heen provided at three levels: i. To indicate if a topic is discussed in greater detail elsewhere.

ACOUSTICS, ARCTIC See also: Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Arctic Ocean Circulation. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Ocean Circulation: Meridional Overturning Circulation. Platforms: Autonomous Underwater Vehicles. Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.

ii. To draw the reader’s attention to parallel discussions in other articles. ACOUSTICS, ARCTIC See also: Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Arctic Ocean Circulation. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Ocean Circulation: Meridional Overturning Circulation. Platforms: Autonomous Underwater Vehicles. Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.


xi

xii

Guide to Use of the Encyclopedia

iii. To indicate material that broadens the discussion.

ACOUSTICS, ARCTIC See also: Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Arctic Ocean Circulation. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Ocean Circulation: Meridional Overturning Circulation. Platforms: Autonomous Underwater Vehicles. Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.

3. Index The index will provide you with the volume and page number where the material is to be located, and the index entries differentiate between material that is a whole article, is part of an article or is data presented in a table or figure. On the opening page of the index detailed notes are provided.

4. Appendices In addition to the articles that form the main body of the encyclopedia, there are a number of appendices which provide bathymetric charts and lists of data used throughout the encyclopedia. The appendices are located in volume 6, before the index.

5. Contributors A full list of contributors appears at the beginning of volume 1.


Contributors Volume 1 E E Adams

N Caputi

Massachusetts Institute of Technology, Cambridge, MA, USA

Fisheries WA Research Division, North Beach, WA, Australia

T Akal

C A Carlson

NATO SACLANT Undersea Research Centre, La Spezia, Italy

University of California, Santa Barbara, CA, USA H Chamley

R Arimoto New Mexico State University, Carlsbad, NM, USA

Universite´ de Lille 1, Villeneuve d’Ascq, France R Chester

J L Bannister The Western Australian Museum, Perth, Western Australia

Liverpool University, Liverpool, Merseyside, UK V Christensen University of British Columbia, Vancouver, BC, Canada

E D Barton University of Wales, Bangor, UK

J W Dacey

N R Bates Bermuda Biological Station for Research, St George’s, Bermuda, USA

Woods Hole Oceanographic Institution, Woods Hole, MA, USA R A Duce

A Beckmann

Texas A&M University, College Station, TX, USA

Alfred-Wegener-Institut fu¨r Polar und Meeresforschung, Bremerhaven, Germany

H W Ducklow

P S Bell

The College of William and Mary, Gloucester Point, VA, USA

Proudman Oceanographic Laboratory, Liverpool, UK I Dyer G Birnbaum

Marblehead, MA, USA

Alfred-Wegener-Institut fu¨r Polar und Meeresforschung, Bremerhaven, Germany

D W Dyrssen Gothenburg University, Go¨teborg, Sweden

B O Blanton The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA E A Boyle Massachusetts Institute of Technology, Cambridge, MA, USA

S M Evans Newcastle University, Newcastle, UK I Everson Anglia Ruskin University, Cambridge, UK

P Boyle

J W Farrington

University of Aberdeen, Aberdeen, UK

Woods Hole Oceanographic Institution, MA, USA

D M Bush

M Fieux

State University of West Georgia, Carrollton, GA, USA

Universite´ Pierre et Marie Curie, Paris, France

K Caldeira

R A Fine

Stanford University, Stanford, CA, USA

University of Miami, Miami, FL, USA


xiii

xiv

Contributors

K G Foote Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA L François University of Lie`ge, Lie`ge, Belgium M A M Friedrichs Old Dominion University, Norfolk, VA, USA T Gaston National Wildlife Research Centre, Quebec, Canada J Gemmrich University of Victoria, Victoria, BC, Canada Y Godde´ris University of Lie`ge, Lie`ge, Belgium D R Godschalk University of North Carolina, Chapel Hill, NC, USA A J Gooday Southampton Oceanography Centre, Southampton, UK A L Gordon Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA D A Hansell University of Miami, Miami FL, USA L W Harding Jr University of Maryland, College Park, MD, USA R Harris Plymouth Marine Laboratory, Plymouth, UK P J Herring Southampton Oceanography Centre, Southampton, UK B M Hickey University of Washington, Seattle, WA, USA M A Hixon Oregon State University, Corvallis, OR, USA E E Hofmann Old Dominion University, Norfolk, VA, USA S Honjo Woods Hole Oceanographic Institution, Woods Hole, MA, USA D J Howell Newcastle University, Newcastle, UK J M Huthnance CCMS Proudman Oceanographic Laboratory, Wirral, UK B Ja¨hne University of Heidelberg, Heidelberg, Germany F B Jensen SACLANT Undersea Research Centre, La Spezia, Italy A John Sir Alister Hardy Foundation for Ocean Science, Plymouth, UK

C D Jones University of Washington, Seattle, WA, USA P F Kingston Heriot-Watt University, Edinburgh, UK W Krauss Institut fu¨r Meereskunde an der Universita¨t Kiel, Kiel, Germany W A Kuperman Scripps Institution of Oceanography, University of California, San Diego, CA, USA D Lal Scripps Institute of Oceanography, University of California San Diego, La Jolla, CA, USA C S Law Plymouth Marine Laboratory, The Hoe, Plymouth, UK W J Lindberg University of Florida, Gainesville, FL, USA J R E Lutjeharms University of Cape Town, Rondebosch, South Africa P Malanotte-Rizzoli Massachusetts Institute of Technology, Cambridge, MA, USA W R Martin Woods Hole Oceanographic Institution, Woods Hole, MA, USA R P Matano Oregon State University, Corvallis, OR, USA J W McManus University of Miami, Miami, FL, USA G M McMurtry University of Hawaii at Manoa, Honolulu, HI, USA R Melville-Smith Fisheries WA Research Division, North Beach, WA, Australia P N Mikhalevsky Science Applications International Corporation, McLean, VA, USA W D Miller University of Maryland, College Park, MD, USA D Monahan University of New Hampshire, Durham, NH, USA J C Moore University of California at Santa Cruz, Santa Cruz, CA, USA A Morel Universite´ Pierre et Marie Curie, Villefranche-sur-Mer, France R Narayanaswamy The University of Manchester, Manchester, UK


Contributors W J Neal Grand Valley State University, Allendale, MI, USA D Pauly University of British Columbia, Vancouver, BC, Canada J W Penn Fisheries WA Research Division, North Beach, WA, Australia L C Peterson University of Miami, Miami, FL, USA S G Philander Princeton University, Princeton, NJ, USA N J Pilcher Universiti Malaysia Sarawak, Sarawak, Malaysia O H Pilkey Duke University, Durham, NC, USA

xv

D H Shull Western Washington University, Bellingham, WA, USA D K Steinberg College of William and Mary, Gloucester Pt, VA, USA L Stramma University of Kiel, Kiel, Germany R N Swift NASA Goddard Space Flight Center, Wallops Island, VA, USA T Takahashi Lamont Doherty Earth Observatory, Columbia University, Palisades, NY, USA P D Thorne Proudman Oceanographic Laboratory, Liverpool, UK

A R Piola Universidad de Buenos Aires, Buenos Aires, Argentina J M Prospero University of Miami, Miami, FL, USA S Rahmstorf Potsdam Institute for Climate Impact Research, Potsdam, Germany P C Reid SAHFOS, Plymouth, UK G Reverdin LEGOS, Toulouse Cedex, France S R Rintoul CSIRO Antarctic Climate and Ecosystems Cooperative Research Centre, Hobart, TAS, Australia J M Roberts Scottish Association for Marine Science, Oban, UK P A Rona Rutgers University, New Brunswick, NJ, USA T C Royer Old Dominion University, Norfolk, VA, USA B Rudels Finnish Institute of Marine Research, Helsinki, Finland

P L Tyack Woods Hole Oceanographic Institution, Woods Hole, USA T Tyrrell National Oceanography Centre, Southampton, UK F E Werner The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA E A Widder Harbor Branch Oceanographic Institution, Fort Pierce, FL, USA D J Wildish Fisheries and Oceans Canada, St. Andrews, NB, Canada A J Williams, III Woods Hole Oceanographic Institution, Woods Hole, MA, USA D K Woolf Southampton Oceanography Centre, Southampton, UK

W Seaman University of Florida, Gainesville, FL, USA

C W Wright NASA Goddard Space Flight Center, Wallops Island, VA, USA

F Sevilla, III, University of Santo Tomas, Manila,The Philippines

J D Wright Rutgers University, Piscataway, NJ, USA

L V Shannon University of Cape Town, Cape Town, South Africa

J R Young The Natural History Museum, London, UK

G I Shapiro University of Plymouth, Plymouth, UK

H J Zemmelink University of Groningen, Haren, The Netherlands

A D Short University of Sydney, Sydney, Australia

W Zenk Universita¨t Kiel, Kiel, Germany


xvi

Contributors

Volume 2 G P Arnold Centre for Environment, Fisheries & Aquaculture Science, Suffolk, UK

K Dyer University of Plymouth, Plymouth, UK

K M Bailey Alaska Fisheries Science Center, Seattle, WA, USA

M Elliott Institute of Estuarine and Coastal Studies, University of Hull, Hull, UK

J G Baldauf Texas A&M University, College Station, TX, USA

D M Farmer Institute of Ocean Sciences, Sidney, BC, Canada

J Bascompte CSIC, Seville, Spain

A V Fedorov Yale University, New Haven, CT, USA

A Belgrano Institute of Marine Research, Lysekil, Sweden

M J Fogarty Northeast Fisheries Science Center, National Marine Fisheries Service, Woods Hole, MA, USA

O A Bergstad Institute of Marine Research, Flødevigen His, Norway J H S Blaxter Scottish Association for Marine Science, Argyll, UK

R Fonteyne Agricultural Research Centre, Ghent, Oostende, Belgium

Q Bone The Marine Biological Association of the United Kingdom, Plymouth, UK

D J Fornari Woods Hole Oceanographic Institution, Woods Hole, USA

I Boyd University of St. Andrews, St. Andrews, UK

A E Gargett Old Dominion University, Norfolk, VA, USA

K M Brander DTU Aqua, Charlottenlund, Denmark and International Council for the Exploration of the Sea (ICES), Copenhagen, Denmark

C H Gibson University of California, San Diego, La Jolla, CA, USA

J N Brown Yale University, New Haven, CT, USA T K Chereskin University of California San Diego, La Jolla, CA, USA J S Collie Danish Institute for Fisheries Research, Charlottenlund, Denmark and University of Rhode Island, Narragansett, RI, USA G Cresswell CSIRO Marine Research, Tasmania, Australia

J D M Gordon Scottish Association for Marine Science, Argyll, UK J F Grassle Rutgers University, New Brunswick, New Jersey, USA S J Hall Flinders University, Adelaide, SA, Australia N Hanson University of St. Andrews, St. Andrews, UK P J B Hart University of Leicester, Leicester, UK

J Davenport University College Cork, Cork, Ireland

K R Helfrich Woods Hole Oceanographic Institution, Woods Hole, MA, USA

R H Douglas City University, London, UK

D M Higgs University of Windsor, Windsor, ON, Canada

S Draxler Karl-Franzens-Universita¨t Graz, Graz, Austria

N G Hogg Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J T Duffy-Anderson Alaska Fisheries Science Center, Seattle, WA, USA J A Dunne Santa Fe Institute, Santa Fe, NM, USA and Pacific Ecoinformatics and Computational Ecology Lab, Berkely, CA, USA

E D Houde University of Maryland, Solomons, MD, USA V N de Jonge Department of Marine Biology, Groningen University, Haren, The Netherlands


Contributors K Katsaros Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, FL, USA J M Klymak University of Victoria, Victoria, BC, Canada M Kucera Eberhard Karls Universita¨t Tu¨bingen, Tu¨bingen, Germany R S Lampitt University of Southampton, Southampton, UK J R N Lazier Bedford Institute of Oceanography, NS, Canada J R Ledwell Woods Hole Oceanographic Institution, Woods Hole, MA, USA P F J Lermusiaux Harvard University, Cambridge, MA, USA M E Lippitsch Karl-Franzens-Universita¨t Graz, Graz, Austria

xvii

T J Pitcher University of British Columbia, Vancouver, Canada A N Popper University of Maryland, College Park, MD, USA J F Price Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA R D Prien Southampton Oceanography Centre, Southampton, UK A-L Reysenbach Portland State University, Portland, OR, USA P L Richardson Woods Hole Oceanographic Institution, Woods Hole, MA, USA A R Robinson Harvard University, Cambridge, MA, USA M D J Sayer Dunstaffnage Marine Laboratory, Oban, Argyll, UK

B J McCay Rutgers University, New Brunswick, NJ, USA

R W Schmitt Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J D McCleave University of Maine, Orono, ME, USA

J Scott DERA Winfrith, Dorchester, Dorset, UK

D Minchin Marine Organism Investigations, Killaloe, Republic of Ireland

M P Sissenwine Northeast Fisheries Science Center, Woods Hole, MA, USA

C M Moore University of Essex, Colchester, UK K Moran University of Rhode Island, Narragansett, RI, USA G R Munro University of British Columbia, Vancouver, BC, Canada J D Nash Oregon State University, Corvallis, Oregon, OR, USA A C Naveira Garabato University of Southampton, Southampton, UK

T P Smith Northeast Fisheries Science Center, Woods Hole, MA, USA P V R Snelgrove Memorial University of Newfoundland, St John’s, NL, Canada M A Spall Woods Hole Oceanographic Institution, Woods Hole, MA, USA A Stigebrandt University of Gothenburg, Gothenburg, Sweden D A V Stow University of Southampton, Southampton, UK

J D Neilson Department of Fisheries and Oceans, New Brunswick, Canada

D J Suggett University of Essex, Colchester, UK

Y Nozakiw University of Tokyo, Tokyo, Japan

U R Sumaila University of British Columbia, Vancouver, BC, Canada

R I Perry Department of Fisheries and Oceans, British Columbia, Canada S G Philander Princeton University, Princeton, NJ, USA w

Deceased.

K S Tande Norwegian College of Fishery Science, Tromsø, Norway S A Thorpe National Oceanography Centre, Southampton, UK R S J Tol Economic and Social Research Institute, Dublin, Republic of Ireland


xviii

Contributors

K E Trenberth National Center for Atmospheric Research, Boulder, CO, USA J J Videler Groningen University, Haren, The Netherlands

R S Wells Chicago Zoological Society, Sarasota, FL, USA D C Wilson Institute for Fisheries Management and Coastal Community Development, Hirtshals, Denmark

Volume 3 S Ali Plymouth Marine Laboratory, Plymouth, UK

K H Coale Moss Landing Marine Laboratories, CA, USA

J T Andrews University of Colorado, Boulder, CO, USA

M F Coffins University of Texas at Austin, Austin, TX, USA

M A de Angelis Humboldt State University, Arcata, CA, USA

P J Corkeron James Cook University, Townsville, Australia

A J Arp Romberg Tiburon Center for Environment Studies, Tiburon, CA, USA

B C Coull University of South Carolina, Columbia, SC, USA

T Askew Harbor Branch Oceanographic Institute, Ft Pierce, FL, USA

R Cowen University of Miami, Miami, FL, USA

R D Ballard Institute for Exploration, Mystic, CT, USA

G Cresswell CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia

G Barnabe´ Universite´ de Montpellier II, France

D S Cronan Royal School of Mines, London, UK

R S K Barnes University of Cambridge, Cambridge, UK

J Csirke Food and Agriculture Organization of the United Nations, Rome, Italy

E D Barton University of Wales, Bangor, Menai Bridge, Anglesey, UK

G A Cutter Old Dominion University, Norfolk, VA, USA

D Bhattacharya University of Iowa, Iowa City, IA, USA

D J DeMaster North Carolina State University, Raleigh, NC, USA

F von Blanckenburg Universita¨t Bern, Bern, Switzerland

T D Dickey University of California, Santa Barbara, CA, USA

D R Bohnenstiehl North Carolina State University, Raleigh, NC, USA

D Diemand Coriolis, Shoreham, VT, USA

H L Bryden University of Southampton, Southampton, UK J Burger Rutgers University, Piscataway, NJ, USA S M Carbotte Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA G T Chandler University of South Carolina, Columbia, SC, USA M A Charette Woods Hole Oceanographic Institution, Woods Hole, MA, USA

C S M Doake British Antarctic Survey, Cambridge, UK C M Domingues CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia C J Donlon Space Applications Institute, Ispra, Italy F Doumenge Museé Oceánographique de Monaco, Monaco R A Dunn University of Hawaii at Manoa, Honolulu, HI, USA


Contributors R P Dziak Oregon State University/National Oceanic and Atmospheric Administration, Hatfield Marine Science Center, Newport, OR, USA O Eldholm University of Oslo, Oslo, Norway

xix

S K Hooker University of St. Andrews, St. Andrews, UK H Hotta Japan Marine Science & Technology Center, Japan G R Ierley University of California San Diego, La Jolla, CA, USA

A E Ellis Marine Laboratory, Aberdeen, Scotland, UK C R Engle University of Arkansas at Pine Bluff, Pine Bluff, AR, USA C C Eriksen University of Washington, Seattle, WA, USA V Ettwein University College London, London, UK S Farrow Carnegie Mellon University, Pittsburgh, PA, USA M Fieux Universite´ Pierre et Marie Curie, Paris Cedex, France N Forteath Inspection Head Wharf, TAS, Australia J D Gage Scottish Association for Marine Science, Oban, UK S M Garcia Food and Agriculture Organization of the United Nations, Rome, Italy

G Ito University of Hawaii at Manoa, Honolulu, HI, USA J Jacoby Woods Hole Oceanographic Institution, Woods Hole, MA, USA M J Kaiser Bangor University, Bangor, UK A E S Kemp University of Southampton, Southampton Oceanography Centre, Southampton, UK W M Kemp University of Maryland Center for Environmental Science, Cambridge, MD, USA V S Kennedy University of Maryland, Cambridge, MD, USA P F Kingston Heriot-Watt University, Edinburgh, UK G L Kooyman University of California San Diego, CA, USA

C Garrett University of Victoria, VIC, Canada

W Krijgsman University of Utrecht, Utrecht, The Netherlands

R N Gibson Scottish Association for Marine Science, Argyll, Scotland

J B Kristoffersen University of Bergen, Bergen, Norway

M Gochfeld Environmental and Community Medicine, Piscataway, NJ, USA

K Lambeck Australian National University, Canberra, ACT, Australia

H O Halvorson University of Massachusetts Boston, Boston, MA, USA

R S Lampitt University of Southampton, Southampton, UK

B U Haq Vendome Court, Bethesda, MD, USA

M Landry University of Hawaii at Manoa, Department of Oceanography, Honolulu, HI, USA

G R Harbison Woods Hole Oceanographic Institution, Woods Hole, MA, USA

C G Langereis University of Utrecht, Utrecht, The Netherlands A Lascaratos University of Athens, Athens, Greece

R M Haymon University of California, CA, USA

S Leibovich Cornell University, Ithaca, NY, USA

D L Hebert University of Rhode Island, RI, USA J E Heyning The Natural History Museum of Los Angeles County, Los Angeles, CA, USA P Hoagland Woods Hole Oceanographic Institution, Woods Hole, MA, USA

W G Leslie Harvard University, Cambridge, MA, USA C Llewellyn Plymouth Marine Laboratory, Plymouth, UK R A Lutz Rutgers University, New Brunswick, NJ, USA


xx

Contributors

K C Macdonald Department of Geological Sciences and Marine Sciences Institute, University of California, Santa Barbara, CA, USA F T Mackenzie University of Hawaii, Honolulu, HI, USA L P Madin Woods Hole Oceanographic Institution, Woods Hole, MA, USA M Maslin University College London, London, UK G A Maul Florida Institute of Technology, Melbourne, FL, USA M McNutt MBARI, Moss Landing, CA, USA M G McPhee McPhee Research Company, Naches, WA, USA A D Mclntyre University of Aberdeen, Aberdeen, UK J Mienert University of Tromsø, Tromsø, Norway G E Millward University of Plymouth, Plymouth, UK H Momma Japan Marine Science & Technology Center, Japan J H Morison University of Washington, Seattle, WA, USA A E Mulligan Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J E Petersen Oberlin College, Oberlin, OH, USA M Phillips Network of Aquaculture Centres in Asia-Pacific (NACA), Bangkok, Thailand B Qiu University of Hawaii at Manoa, Hawaii, USA F Quezada Biotechnology Center of Excellence Corporation, Waltham, MA, USA N N Rabalais Louisiana Universities Marine Consortium, Chauvin, LA, USA R D Ray NASA Goddard Space Flight Center, Greenbelt, MD, USA M R Reeve National Science Foundation, Arlington VA, USA R R Reeves Okapi Wildlife Associates, QC, Canada A Reyes-Prieto University of Iowa, Iowa City, IA, USA P B Rhines University of Washington,Seattle, WA, USA A R Robinson Harvard University, Cambridge, MA, USA H T Rossby University of Rhode Island, Kingston, RI, USA H M Rozwadowski Georgia Institute of Technology, Atlanta, Georgia, USA

W Munk University of California San Diego, La Jolla, CA, USA

A G V Salvanes University of Bergen, Bergen, Norway

E J Murphy British Antarctic Survey, Marine Life Sciences Division, Cambridge, UK

R Schlitzer Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany

P D Naidu National Institute of Oceanography, Dona Paula, India

M E Schumacher Woods Hole Oceanographic Institution, Woods Hole, MA, USA

N Niitsuma Shizuoka University, Shizuoka, Japan

M I Scranton State University of New York, Stony Brook, NY, USA

D B Olson University of Miami, Miami, FL, USA G-A Paffenho¨fer Skidaway Institute of Oceanography, Savannah, GA, USA C Paris University of Miami, Miami, FL, USA M R Perfit Department of Geological Sciences, University of Florida, Gainsville, FL, USA

K Sherman Narragansett Laboratory, Narragansett, RI, USA M D Spalding UNEP World Conservation Monitoring Centre and Cambridge Coastal Research Unit, Cambridge, UK J Sprintall University of California San Diego, La Jolla, CA, USA J H Steele Woods Hole Oceanographic Institution, MA, USA


Contributors C A Stein University of Illinois at Chicago, Chicago, IL, USA

S M Van Parijs Norwegian Polar Institute, Tromsø, Norway

C Stickley University College London, London, UK

L M Ver University of Hawaii, Honolulu, HI, USA

U R Sumaila University of British Columbia, Vancouver, BC, Canada

F J Vine University of East Anglia, Norwich, UK

S Takagawa Japan Marine Science & Technology Center, Japan

K L Von Damm University of New Hampshire, Durham, NH, USA

P K Taylor Southampton Oceanography Centre, Southampton, UK

R P Von Herzen Woods Hole Oceanographic Institution, Woods Hole, MA, USA

A Theocharis National Centre for Marine Research (NCMR), Hellinikon, Athens, Greece

xxi

D Wartzok Florida International University, Miami, FL, USA

P C Ticco Massachusetts Maritime Academy, Buzzards Bay, MA, USA R P Trask Woods Hole Oceanographic Institution, Woods Hole, MA, USA

W F Weeks Portland, OR, USA R A Weller Woods Hole Oceanographic Institution, Woods Hole, MA, USA

A W Trites University of British Columbia, British Columbia, Canada

J A Whitehead Woods Hole Oceanographic Institution, Woods Hole, MA, USA

A Turner University of Plymouth, Plymouth, UK

J C Wiltshire University of Hawaii, Manoa, Honolulu, HA, USA

P L Tyack Woods Hole Oceanographic Institution, Woods Hole, MA, USA

C Woodroffe University of Wollongong, Wollongong, NSW, Australia

G J C Underwood University of Essex, Colchester, UK

C Wunsch Massachusetts Institute of Technology, Cambridge, MA, USA

C L Van Dover The College of William and Mary, Williamsburg, VA, USA

H S Yoon University of Iowa, Iowa City, IA, USA

Volume 4 A Alldredge University of California, Santa Barbara, CA, USA D M Anderson Woods Hole Oceanographic Institution, Woods Hole, MA, USA O R Anderson Columbia University, Palisades, NY, USA

J M Bewers Bedford Institute of Oceanography, Dartmouth, NS, Canada N V Blough University of Maryland, College Park, MD, USA W Bonne

P G Baines CSIRO Atmospheric Research, Aspendale, VIC, Australia

Federal Public Service Health, Food Chain Safety and Environment, Brussels, Belgium

J M Baker Clock Cottage, Shrewsbury, UK

University of Cape Town, Cape Town, Republic of South Africa

J G Bellingham Monterey Bay Aquarium Research Institute, Moss Landing, CA, USA

R D Brodeur

G M Branch

Northwest Fisheries Science Center, Newport, OR, USA


xxii

Contributors

H Burchard Baltic Sea Research Institute Warnemu¨nde, Warnemu¨nde, Germany P H Burkill Plymouth Marine Laboratory, West Hoe, Plymouth, UK Francois Carlotti C.N.R.S./Universite´ Bordeaux 1, Arachon, France K L Casciotti Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J S Grayw University of Oslo, Oslo, Norway A G Grottoli University of Pennsylvania, Philadelphia, PA, USA N Gruber Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, Switzerland K C Hamer University of Durham, Durham, UK D Hammond University of Southern California, Los Angeles, CA, USA

A Clarke British Antarctic Survey, Cambridge, UK

W W Hay Christian-Albrechts University, Kiel, Germany

M B Collins National Oceanography Centre, Southampton, UK

J W Heath Coastal Fisheries Institute, CCEER Louisiana State University, Baton Rouge, LA, USA

J J Cullen Department of Oceanography, Halifax, NS, Canada D H Cushing Lowestoft, Suffolk, UK

D Hedgecock University of Southern California, Los Angeles, CA, USA C Hemleben Tu¨bingen University, Tu¨bingen, Germany

K L Denman University of Victoria, Victoria, BC, Canada S C Doney Woods Hole Oceanographic Institution, Woods Hole, MA, USA

T D Herbert Brown University, Providence, RI, USA I Hewson University of California Santa Cruz, Santa Cruz, CA, USA

J F Dower University of British Columbia, Vancouver, BC, Canada

Richard Hey University of Hawaii at Manoa, Honolulu, HI, USA

K Dysthe University of Bergen, Bergen, Norway

P Hoagland Woods Hole Oceanographic Institution, Woods Hole, MA, USA

H N Edmonds University of Texas at Austin, Port Aransas, TX, USA

N Hoepffner Institute for Environment and Sustainability, Ispra, Italy

L Føyn Institute of Marine Research, Bergen, Norway

M Hood Intergovernmental Oceanographic Commission, Paris, France

J Fuhrman University of Southern California, Los Angeles, CA, USA

M J Howarth Proudman Oceanographic Laboratory, Wirral, UK

C P Gallienne Plymouth Marine Laboratory, West Hoe, Plymouth, UK

M Huber Purdue University, West Lafayette, IN, USA

E Garel CIACOMAR, Algarve University, Faro, Portugal

J W Hurrell National Center for Atmospheric Research, Boulder, CO, USA

D M Glover Woods Hole Oceanographic Institution, Woods Hole, MA, USA S L Goodbred Jr State University of New York, Stony Brook, NY, USA J D M Gordon Scottish Association for Marine Science, Oban, Argyll, UK

D R Jackett CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia R A Jahnke Skidaway Institute of Oceanography, Savannah, GA, USA w

Deceased.


Contributors

xxiii

A Jarre University of Cape Town, Cape Town, South Africa

A F Michaels University of Southern California, Los Angeles, CA, USA

J Joseph La Jolla, CA, USA

J D Milliman College of William and Mary, Gloucester, VA, USA

D M Karl University of Hawaii at Manoa, Honolulu, HI, USA

C D Mobley Sequoia Scientific, Inc., WA, USA

K L Karsh Princeton University, Princeton, NJ, USA

M M Mullinw Scripps Institution of Oceanography, La Jolla, CA, USA

J Karstensen Universita¨t Kiel (IFM-GEOMAR), Kiel, Germany

P Mu¨ller University of Hawaii, Honolulu, HI, USA

R M Key Princeton University, Princeton, NJ, USA

L A Murray The Centre for Environment, Fisheries and Aquaculture Sciences, Lowestoft, UK

P D Killworth Southampton Oceanography Centre, Southampton, UK B Klinger Center for Ocean-Land-Atmosphere Studies (COLA), Calverton, MD, USA H E Krogstad NTNU, Trondheim, Norway I Laing Centre for Environment Fisheries and Aquaculture Science, Weymouth, UK

T Nagai Tokyo University of Marine Science and Technology, Tokyo, Japan K H Nisancioglu Bjerknes Centre for Climate Research, University of Bergen, Bergen, Norway Y Nozakiw University of Tokyo, Tokyo, Japan

G F Lane-Serff University of Manchester, Manchester, UK

K J Orians The University of British Columbia, Vancouver, BC, Canada

A Longhurst Place de I’Eglise, Cajarc, France

C A Paulson Oregon State University, Corvallis, OR, USA

R Lukas University of Hawaii at Manoa, Hawaii, USA

W G Pearcy Oregon State University, Corvallis, OR, USA

M Lynch University of California Santa Barbara, Santa Barbara, CA, USA

W S Pegau Oregon State University, Corvallis, OR, USA

M Macleod World Wildlife Fund, Washington, DC, USA E Maranõń University of Vigo, Vigo, Spain S Martin University of Washington, Seattle, WA, USA S M Masutani University of Hawaii at Manoa, Honolulu, HI, USA I N McCave University of Cambridge, Cambridge, UK T J McDougall CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia C L Merrin The University of British Columbia, Vancouver, BC, Canada

T Platt Dalhousie University, NS, Canada J J Polovina National Marine Fisheries Service, Honolulu, HI, USA D Quadfasel Niels Bohr Institute, Copenhagen, Denmark J A Raven Biological Sciences, University of Dundee, Dundee, UK G E Ravizza Woods Hole Oceanographic Institution, Woods Hole, MA, USA A J Richardson University of Queensland, St. Lucia, QLD, Australia M Rubega University of Connecticut, Storrs, CT, USA w

Deceased.


xxiv

Contributors

K C Ruttenberg Woods Hole Oceanographic Institution, Woods Hole, MA, USA

K K Turekian Yale University, New Haven, CT, USA T Tyrrell University of Southampton, Southampton, UK

A G V Salvanes University of Bergen, Bergen, Norway

O Ulloa Universidad de Concepcioń, Concepcioń, Chile

S Sathyendranath Dalhousie University, NS, Canada

C M G Vivian The Centre for Environment, Fisheries and Aquaculture Sciences, Lowestoft, UK

R Schiebel Tu¨bingen University, Tu¨bingen, Germany F B Schwing NOAA Fisheries Service, Pacific Grove, CA, USA

J J Walsh University of South Florida, St. Petersburg, FL, USA

M P Seki National Marine Fisheries Service, Honolulu, HI, USA

R M Warwick Plymouth Marine Laboratory, Plymouth, UK

L J Shannon Marine and Coastal Management, Cape Town, South Africa

N C Wells Southampton Oceanography Centre, Southampton, UK

K Shepherd Institute of Ocean Sciences, Sidney, BC, Canada

J A Whitehead Woods Hole Oceanographic Institution, Woods Hole, MA, USA

D Siegel-Causey Harvard University, Cambridge MA, USA D M Sigman Princeton University, Princeton, NJ, USA A Soloviev Nova Southeastern University, FL, USA J H Steele Woods Hole Oceanographic Institution, MA, USA P K Takahashi University of Hawaii at Manoa, Honolulu, HI, USA L D Talley Scripps Institution of Oceanography, La Jolla, CA, USA E Thomas Yale University, New Haven, CT, USA J R Toggweiler NOAA, Princeton, NJ, USA

M Wilkinson Heriot-Watt University, Edinburgh, UK R G Williams University of Liverpool, Oceanography Laboratories, Liverpool, UK C A Wilson III Department of Oceanography and Coastal Sciences, and Coastal Fisheries Institute, CCEER Louisiana State University, Baton Rouge, LA, USA H Yamazaki Tokyo University of Marine Science and Technology, Tokyo, Japan B deYoung Memorial University, St. John’s, NL, Canada G Zibordi Institute for Environment and Sustainability, Ispra, Italy

Volume 5 D G Ainley H.T. Harvey Associates, San Jose CA, USA W Alpers University of Hamburg, Hamburg, Germany J R Apelw Global Ocean Associates, Silver Spring, MD, USA w

Deceased.

A B Baggeroer Massachusetts Institute of Technology, Cambridge, MA, USA L T Balance NOAA-NMFS, La Jolla, CA, USA R Batiza Ocean Sciences, National Science Foundation, VA, USA W H Berger Scripps Institution of Oceanography, La Jolla, CA, USA


Contributors J L Bodkin US Geological Survey, AK, USA

I Everson British Antarctic Survey Cambridge, UK

I L Boyd Natural Environment Research Council, Cambridge, UK

I Fer University of Bergen, Bergen, Norway

A C Brown University of Cape Town, Cape Town, Republic of South Africa

M Fieux Universite´-Pierre et Marie Curie, Paris, France

xxv

J Burger Rutgers University, Piscataway, NJ, USA

R A Flather Proudman Oceanographic Laboratory, Bidston Hill, Prenton, UK

C J Camphuysen Netherlands Institute for Sea Research, Texel, The Netherlands

G S Giese Woods Hole Oceanographic Institution, Woods Hole, MA, USA

D C Chapman Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J M Gregory Hadley Centre, Berkshire, UK

R E Cheney Laboratory for Satellite Altimetry, Silver Spring, Maryland, USA T Chopin University of New Brunswick, Saint John, NB, Canada J A Church Antarctic CRC and CSIRO Marine Research, TAS, Australia J K Cochran State University of New York, Stony Brook, NY, USA P Collar Southampton Oceanography Centre, Southampton, UK R J Cuthbert University of Otago, Dunedin, New Zealand L S Davis University of Otago, Dunedin, New Zealand K L Denman University of Victoria, Victoria BC, Canada R P Dinsmore Woods Hole Oceanographic Institution, Woods Hole, MA, USA G J Divoky University of Alaska, Fairbanks, AK, USA

S M Griffies NOAA/GFDL, Princeton, NJ, USA G Griffiths Southampton Oceanography Centre, Southampton, UK A Harding University of California, San Diego, CA, USA W S Holbrook University of Wyoming, Laramie, WY, USA G L Hunt, Jr University of Washington, Seattle, WA, USA and University of California, Irvine, CA, USA P Hutchinson North Atlantic Salmon Conservation Organization, Edinburgh, UK K B Katsaros Atlantic Oceanographic and Meteorological Laboratory, NOAA, Miami, FL, USA H L Kite-Powell Woods Hole Oceanographic Institution, Woods Hole, MA, USA M A Kominz Western Michigan University, Kalamazoo, MI, USA

L M Dorman University of California, San Diego, La Jolla, CA, USA

R G Kope Northwest Fisheries Science Center, Seattle, WA, USA

J F Dower University of British Columbia, Vancouver, BC, Canada

G S E Lagerloef Earth and Space Research, Seattle, WA, USA

J B Edson Woods Hole Oceanographic Institution, Woods Hole, MA, USA

L M Lairdw Aberdeen University, Aberdeen, UK

T I Eglinton Woods Hole Oceanographic Institution, Woods Hole, MA, USA

M Leppa¨ranta University of Helsinki, Helsinki, Finland w

Deceased.


xxvi

Contributors

E J Lindstrom NASA Science Mission Directorate, Washington, DC, USA

C T Roman University of Rhode Island, Narragansett, RI, USA

A K Liu NASA Goddard Space Flight Center, Greenbelt, MD, USA

M Sawhney University of New Brunswick, Saint John, NB, Canada

C R McClain NASA Goddard Space Flight Center, Greenbelt, MD, USA

G Shanmugam The University of Texas at Arlington, Arlington, TX, USA

D J McGillicuddy Jr Woods Hole Oceanographic Institution, Woods Hole, MA, USA

J Sharples Proudman Oceanographic Laboratory, Liverpool, UK

W K Melville Scripps Institution of Oceanography, La Jolla CA, USA

J H Simpson Bangor University, Bangor, UK R K Smedbol Dalhousie University, Halifax, NS, Canada

D Mills Atlantic Salmon Trust, UK

L B Spear H.T. Harvey Associates, San Jose, CA, USA

P J Minnett University of Miami, Miami, FL, USA W A Montevecchi Memorial University of Newfoundland, NL, Canada W S Moore University of South Carolina, Columbia, SC, USA S J Morreale Cornell University, Ithaca, NY, USA K W Nicholls British Antarctic Survey, Cambridge, UK T J O’Shea Midcontinent Ecological Science Center, Fort Collins, CO, USA T E Osterkamp University of Alaska, Alaska, AK, USA F V Paladino Indiana-Purdue University at Fort Wayne, Fort Wayne, IN, USA C L Parkinson NASA Goddard Space Flight Center, Greenbelt, MD, USA A Pearson Woods Hole Oceanographic Institution, Woods Hole, MA, USA J T Potemra SOEST/IPRC, University of Hawaii, Honolulu, HI, USA J A Powell Florida Marine Research Institute, St Petersburg, FL, USA T Qu SOEST/IPRC, University of Hawaii, Honolulu, HI, USA

R L Stephenson St. Andrews Biological Station, St. Andrews, NB, Canada J M Teal Woods Hole Oceanographic Institution, Rochester, MA, USA K K Turekian Yale University, New Haven, CT, USA P Wadhams University of Cambridge, Cambridge, UK W F Weeks Portland, OR, USA G Wefer Universita¨t Bremen, Bremen, Germany W S Wilson NOAA/NESDIS, Silver Spring, MD, USA M Windsor, North Atlantic Salmon Conservation Organization, Edinburgh, UK S Y Wu NASA Goddard Space Flight Center, Greenbelt, MD, USA L Yu Woods Hole Oceanographic Institution, Woods Hole, MA, USA


Contributors

xxvii

Volume 6 A V Babanin Swinburne University of Technology, Melbourne, VIC, Australia

S E Humphris Woods Hole Oceanographic Institution, Woods Hole, MA, USA

R T Barber Duke University Marine Laboratory, Beaufort, NC, USA

W J Jenkins University of Southampton, Southampton, UK

J Bartram World Health Organization, Geneva, Switzerland

D R B Kraemer The Johns Hopkins University, Baltimore, MD, USA

A Beckmann Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung, Bremerhaven, Germany M C Benfield Louisiana State University, Baton Rouge, LA, USA P S Bogden Maine State Planning Office, Augusta, ME, USA J A T Bye The University of Melbourne, Melbourne, VIC, Australia M F Cronin NOAA Pacific Marine Environmental Laboratory, Seattle, WA, USA A R J David Bere Alston, Devon, UK W Deuser Woods Hole Oceanographic Institution, Woods Hole, MA, USA J Donat Old Dominion University, Norfolk, VA, USA C Dryden Old Dominion University, Norfolk, VA, USA A Dufour United States Environmental Protection Agency, OH, USA C A Edwards University of Connecticut, Groton, CT, USA W J Emery University of Colorado, Boulder, CO, USA E Fahrbach Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung, Bremerhaven, Germany

S Krishnaswami Physical Research Laboratory, Ahmedabad, India E L Kunze University of Washington, Seattle, WA, USA T E L Langford University of Southampton, Southampton, UK J R Ledwell Woods Hole Oceanographic Institution, Woods Hole, MA, USA P L-F Liu Cornell University, Ithaca, NY, USA M M R van der Loeff Alfred-Wegener-Institut fu¨r Polar und Meereforschung Bremerhaven, Germany R Lueck University of Victoria, Victoria, BC, Canada J E Lupton Hatfield Marine Science Center, Newport, OR, USA L P Madin Woods Hole Oceanographic Institution, Woods Hole, MA, USA M E McCormick The Johns Hopkins University, Baltimore, MD, USA M G McPhee McPhee Research Company, Naches, WA, USA J H Middleton The University of New South Wales, Sydney, NSW, Australia P J Minnett University of Miami, Miami, FL, USA E C Monahan University of Connecticut at Avery Point, Groton, CT, USA

A M Gorlov Northeastern University, Boston, Massachusetts, USA

C Moore WET Labs Inc., Philomath, OR, USA

I Helmond CSIRO Marine Research, TAS, Australia

J H Morison University of Washington, Seattle, WA, USA

R A Holman Oregon State University, Corvallis, OR, USA

J N Moum Oregon State University, Corvallis, OR, USA


xxviii

Contributors

N S Oakey Bedford Institute of Oceanography, Dartmouth, NS, Canada D T Pugh University of Southampton, Southampton, UK D L Rudnick University of California, San Diego, CA, USA H Salas CEPIS/HEP/Pan American Health Organization, Lima, Peru L K Shay University of Miami, Miami, FL, USA W D Smyth Oregon State University, Corvallis, OR, USA J Sprintall University of California San Diego, La Jolla, CA, USA

L St. Laurrent University of Victoria, Victoria, BC, Canada W G Sunda National Ocean Service, NOAA, Beaufort, NC, USA M Tomczak Flinders University of South Australia, Adelaide, SA, Australia A J Watson University of East Anglia, Norwich, UK P H Wiebe Woods Hole Oceanographic Institution, Woods Hole, MA, USA P F Worcester University of California at San Diego, La Jolla, CA, USA


Contents Volume 1 Abrupt Climate Change

S Rahmstorf

1

Absorbance Spectroscopy for Chemical Sensors Abyssal Currents

R Narayanaswamy, F Sevilla, III

W Zenk

Accretionary Prisms

15

J C Moore

31

Acoustic Measurement of Near-Bed Sediment Transport Processes Acoustic Noise

Acoustic Scintillation Thermography Acoustics In Marine Sediments

K G Foote

62

P A Rona, C D Jones

71

T Akal

75

P N Mikhalevsky

Acoustics, Deep Ocean

92

W A Kuperman

101

F B Jensen

112

Acoustics, Shallow Water R Chester

Agulhas Current

120

J R E Lutjeharms

Aircraft Remote Sensing

128

L W Harding Jr, W D Miller, R N Swift, C W Wright

Air–Sea Gas Exchange

38 52

Acoustic Scattering by Marine Organisms

Aeolian Inputs

P D Thorne, P S Bell

I Dyer

Acoustics, Arctic

7

B Ja¨hne

138 147

Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens J W Dacey, H J Zemmelink

157

Air–Sea Transfer: N2O, NO, CH4, CO

163

Alcidae

C S Law

T Gaston

171

Antarctic Circumpolar Current Antarctic Fishes

S R Rintoul

I Everson

191

Anthropogenic Trace Elements in the Ocean Antifouling Materials

E A Boyle

211

W Seaman, W J Lindberg

234

R A Duce

Atmospheric Transport and Deposition of Particulate Material to the Oceans R Arimoto Authigenic Deposits

226

S G Philander

Atmospheric Input of Pollutants

Baleen Whales

203

B Rudels

Atlantic Ocean Equatorial Currents

Bacterioplankton

195

D J Howell, S M Evans

Arctic Ocean Circulation Artificial Reefs

178

G M McMurtry H W Ducklow

J L Bannister

238 J M Prospero, 248 258 269 276


xxix

xxx

Contents

Baltic Sea Circulation Bathymetry

W Krauss

288

D Monahan

297

Beaches, Physical Processes Affecting Benguela Current

Benthic Foraminifera

316 D J Wildish

328

A J Gooday

Benthic Organisms Overview

336

P F Kingston

348

P L Tyack

357

Biogeochemical Data Assimilation

E E Hofmann, M A M Friedrichs

Biological Pump and Particle Fluxes Bioluminescence

Bioturbation

S Honjo

376

A Morel

385

D H Shull

Black Sea Circulation

395

G I Shapiro

Bottom Water Formation

401

A L Gordon

415

Brazil and Falklands (Malvinas) Currents

A R Piola, R P Matano

Breaking Waves and Near-Surface Turbulence

J Gemmrich

D K Woolf

Calcium Carbonates

L C Peterson

E D Barton

Carbon Dioxide (CO2) Cycle

467

T Takahashi

Cenozoic Climate – Oxygen Isotope Evidence Cenozoic Oceans – Carbon Cycle Models

J D Wright L François, Y Godde´ris

R A Fine W R Martin

J W Farrington

Coastal Zone Management

514

539 551 563

F E Werner, B O Blanton

Coastal Topography, Human Impact on Coastal Trapped Waves

502

531

H Chamley

Coastal Circulation Models

495

524

Chemical Processes in Estuarine Sediments

Coccolithophores

E E Adams, K Caldeira

P Boyle

Chlorinated Hydrocarbons

477 487

Carbon Sequestration via Direct Injection into the Ocean

Clay Mineralogy

455

C A Carlson, N R Bates, D A Hansell, D K Steinberg

CFCs in the Ocean

431

445

B M Hickey, T C Royer

Canary and Portugal Currents

Cephalopods

422

439

California and Alaska Currents

Carbon Cycle

364 371

P J Herring, E A Widder

Bio-Optical Models

Bubbles

305

L V Shannon

Benthic Boundary Layer Effects

Bioacoustics

A D Short

D M Bush, O H Pilkey, W J Neal

J M Huthnance D R Godschalk

T Tyrrell, J R Young


572 581 591 599 606

Contents

Cold-Water Coral Reefs Conservative Elements

J M Roberts

615

D W Dyrssen

626

Continuous Plankton Recorders Copepods

A John, P C Reid

R Harris

Coral Reefs

630 640

Coral Reef and Other Tropical Fisheries Coral Reef Fishes

xxxi

V Christensen, D Pauly

M A Hixon

655

J W McManus

660

Corals and Human Disturbance Cosmogenic Isotopes

N J Pilcher

671

D Lal

678

Coupled Sea Ice–Ocean Models Crustacean Fisheries

651

A Beckmann, G Birnbaum

688

J W Penn, N Caputi, R Melville-Smith

699

CTD (Conductivity, Temperature, Depth) Profiler Current Systems in the Atlantic Ocean Current Systems in the Indian Ocean

A J Williams, III

L Stramma M Fieux, G Reverdin

Current Systems in the Southern Ocean

A L Gordon

Current Systems in the Mediterranean Sea

P Malanotte-Rizzoli

708 718 728 735 744

Volume 2 Data Assimilation in Models Deep Convection

A R Robinson, P F J Lermusiaux

J R N Lazier

Deep Submergence, Science of

13 D J Fornari

22

K Moran

37

Deep-Sea Drilling Methodology Deep-Sea Drilling Results

1

J G Baldauf

45

Deep-Sea Fauna

P V R Snelgrove, J F Grassle

55

Deep-Sea Fishes

J D M Gordon

67

Deep-Sea Ridges, Microbiology

A-L Reysenbach

73

Deep-Sea Sediment Drifts

D A V Stow

80

Demersal Species Fisheries

K Brander

90

Determination of Past Sea Surface Temperatures Differential Diffusion

A E Gargett

Dispersion from Hydrothermal Vents Diversity of Marine Species Dolphins and Porpoises

R W Schmitt, J R Ledwell

K R Helfrich

P V R Snelgrove R S Wells

Double-Diffusive Convection

98 114

Dispersion and Diffusion in the Deep Ocean

Drifters and Floats

M Kucera

R W Schmitt

P L Richardson


122 130 139 149 162 171

xxxii

Contents

Dynamics of Exploited Marine Fish Populations East Australian Current

M J Fogarty

G Cresswell

179 187

Economics of Sea Level Rise

R S J Tol

197

Ecosystem Effects of Fishing

S J Hall

201

Eels

J D McCleave

208

Effects of Climate Change on Marine Mammals Ekman Transport and Pumping

T K Chereskin, J F Price

El Ninõ Southern Oscillation (ENSO)

Electrical Properties of Sea Water

Energetics of Ocean Mixing

228

S G Philander

R D Prien

Elemental Distribution: Overview

Y Nozaki

255

A C Naveira Garabato

261

Eutrophication

271

J M Klymak, J D Nash

Estuarine Circulation

288

K Dyer

299

V N de Jonge, M Elliott

Evaporation and Humidity

Fiord Circulation

306

K Katsaros

Exotic Species, Introduction of Expendable Sensors

241 247

w

A V Fedorov, J N Brown

Estimates of Mixing

218 222

K E Trenberth

El Ninõ Southern Oscillation (ENSO) Models

Equatorial Waves

I Boyd, N Hanson

324

D Minchin

332

J Scott

345

A Stigebrandt

353

Fiordic Ecosystems

K S Tande

359

Fish Ecophysiology

J Davenport

367

Fish Feeding and Foraging Fish Larvae

P J B Hart

E D Houde

Fish Locomotion

381

J J Videler

Fish Migration, Horizontal Fish Migration, Vertical

Fish Reproduction

Fish Vision

392

G P Arnold

402

J D Neilson, R I Perry

Fish Predation and Mortality

Fish Schooling

374

411

K M Bailey, J T Duffy-Anderson

J H S Blaxter

425

T J Pitcher

432

R H Douglas

445

Fish: Demersal Fish (Life Histories, Behavior, Adaptations) Fish: General Review

O A Bergstad

Q Bone

458 467

Fish: Hearing, Lateral Lines (Mechanisms, Role in Behavior, Adaptations to Life Underwater) A N Popper, D M Higgs w

417

Deceased.


476

Contents

Fisheries and Climate

K M Brander

Fisheries Economics

483

U R Sumaila, G R Munro

Fisheries Overview

491

M J Fogarty, J S Collie

Fisheries: Multispecies Dynamics Fishery Management

499

J S Collie

505

T P Smith, M P Sissenwine

Fishery Management, Human Dimension

513

D C Wilson, B J McCay

Fishery Manipulation through Stock Enhancement or Restoration Fishing Methods and Fishing Fleets Floc Layers

xxxiii

M D J Sayer

R Fonteyne

522 528 535

R S Lampitt

548

Florida Current, Gulf Stream, and Labrador Current Flow through Deep Ocean Passages Flows in Straits and Channels

P L Richardson

N G Hogg

554 564

D M Farmer

572

Fluid Dynamics, Introduction, and Laboratory Experiments

S A Thorpe

578

Fluorometry for Biological Sensing

D J Suggett, C M Moore

581

Fluorometry for Chemical Sensing

S Draxler, M E Lippitsch

589

Food Webs

A Belgrano, J A Dunne, J Bascompte

Forward Problem in Numerical Models Fossil Turbulence

596

M A Spall

604

C H Gibson

612

Volume 3 Gas Exchange in Estuaries

M I Scranton, M A de Angelis

Gelatinous Zooplankton

L P Madin, G R Harbison

General Circulation Models

Geomorphology

20

C G Langereis, W Krijgsman

C Woodroffe

Geophysical Heat Flow

C A Stein, R P Von Herzen

40 K Lambeck

C C Eriksen A D Mclntyre

Grabs for Shelf Benthic Sampling

67

P F Kingston

70

M McNutt

80

Groundwater Flow to the Coastal Ocean Habitat Modification

49 59

Global Marine Pollution

Gravity

25 33

Glacial Crustal Rebound, Sea Levels, and Shorelines Gliders

9

G R Ierley

Geomagnetic Polarity Timescale

1

A E Mulligan, M A Charette

M J Kaiser

Heat and Momentum Fluxes at the Sea Surface Heat Transport and Climate History of Ocean Sciences

88 99

P K Taylor

H L Bryden H M Rozwadowski


105 114 121

xxxiv

Contents

Holocene Climate Variability Hydrothermal Vent Biota

M Maslin, C Stickley, V Ettwein R A Lutz

125 133

Hydrothermal Vent Deposits

R M Haymon

144

Hydrothermal Vent Ecology

C L Van Dover

151

Hydrothermal Vent Fauna, Physiology of

A J Arp

159

Hydrothermal Vent Fluids, Chemistry of

K L Von Damm

164

Hypoxia

N N Rabalais

172

Icebergs

D Diemand

181

Ice-Induced Gouging of the Seafloor Ice–Ocean Interaction

W F Weeks

J H Morison, M G McPhee

191 198

Ice Shelf Stability

C S M Doake

209

Igneous Provinces

M F Coffins, O Eldholm

218

Indian Ocean Equatorial Currents Indonesian Throughflow

M Fieux

J Sprintall

237

Inherent Optical Properties and Irradiance Internal Tidal Mixing Internal Tides

T D Dickey

W Munk

258

C Garrett

266

International Organizations Intertidal Fishes

M R Reeve

274

R N Gibson

Intra-Americas Sea

244 254

R D Ray

Internal Waves

Intrusions

226

280

G A Maul

286

D L Hebert

295

Inverse Modeling of Tracers and Nutrients

R Schlitzer

300

Inverse Models

C Wunsch

312

IR Radiometers

C J Donlon

319

K H Coale

331

Iron Fertilization Island Wakes Krill

E D Barton

343

E J Murphy

349

Kuroshio and Oyashio Currents

B Qiu

Laboratory Studies of Turbulent Mixing Lagoons

358 J A Whitehead

R S K Barnes

Lagrangian Biological Models Land–Sea Global Transfers

377 D B Olson, C Paris, R Cowen F T Mackenzie, L M Ver

Langmuir Circulation and Instability Large Marine Ecosystems

S Leibovich

K Sherman

Laridae, Sternidae, and Rynchopidae Law of the Sea

371

389 394 404 413

J Burger, M Gochfeld

P Hoagland, J Jacoby, M E Schumacher (c) 2011 Elsevier Inc. All Rights Reserved.

420 432

Contents

Leeuwin Current

G Cresswell, C M Domingues

Long-Term Tracer Changes Macrobenthos Magnetics

444

F von Blanckenburg

455

J D Gage

467

F J Vine

478

Manganese Nodules Mangroves

xxxv

D S Cronan

488

M D Spalding

496

Manned Submersibles, Deep Water

H Hotta, H Momma, S Takagawa

Manned Submersibles, Shallow Water

T Askew

505 513

Mariculture Diseases and Health

A E Ellis

519

Mariculture of Aquarium Fishes

N Forteath

524

Mariculture of Mediterranean Species Mariculture Overview

G Barnabe´, F Doumenge

M Phillips

537

Mariculture, Economic and Social Impacts Marine Algal Genomics and Evolution Marine Biotechnology

532

C R Engle

545

A Reyes-Prieto, H S Yoon, D Bhattacharya

H O Halvorson, F Quezada

552 560

Marine Chemical and Medicine Resources

S Ali, C Llewellyn

567

Marine Fishery Resources, Global State of

J Csirke, S M Garcia

576

Marine Mammal Diving Physiology

G L Kooyman

Marine Mammal Evolution and Taxonomy

J E Heyning

Marine Mammal Migrations and Movement Patterns Marine Mammal Overview

582

P J Corkeron, S M Van Parijs

P L Tyack

Marine Mammal Trophic Levels and Interactions Marine Mammals and Ocean Noise

A W Trites

Marine Policy Overview Marine Protected Areas Marine Silica Cycle

635 643

654 G-A Paffenho¨fer

656

P Hoagland, P C Ticco

664

P Hoagland, U R Sumaila, S Farrow D J DeMaster

R S Lampitt

Maritime Archaeology

622

651

J H Steele

Marine Plankton Communities

Mediterranean Sea Circulation

672 678 686

R D Ballard

Meddies and Sub-Surface Eddies

Meiobenthos

S K Hooker

A E S Kemp

Marine Mesocosms

615

628

R R Reeves

Marine Mammals: Sperm Whales and Beaked Whales

Marine Snow

P L Tyack

D Wartzok

Marine Mammals, History of Exploitation

596 605

Marine Mammal Social Organization and Communication

Marine Mats

589

H T Rossby A R Robinson, W G Leslie, A Theocharis, A Lascaratos

B C Coull, G T Chandler (c) 2011 Elsevier Inc. All Rights Reserved.

695 702 710 726

xxxvi

Contents

Mesocosms: Enclosed Experimental Ecosystems in Ocean Science Mesopelagic Fishes

J E Petersen, W M Kemp

A G V Salvanes, J B Kristoffersen

Mesoscale Eddies

748

P B Rhines

Metal Pollution

755

G E Millward, A Turner

Metalloids and Oxyanions

732

768

G A Cutter

776

Methane Hydrates and Climatic Effects

B U Haq

784

Methane Hydrate and Submarine Slides

J Mienert

790

Microbial Loops

M Landry

Microphytobenthos

799

G J C Underwood

807

Mid-Ocean Ridge Geochemistry and Petrology Mid-Ocean Ridge Seismic Structure Mid-Ocean Ridge Seismicity

M R Perfit

815

S M Carbotte

826

D R Bohnenstiehl, R P Dziak

Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology

837 K C Macdonald

Mid-Ocean Ridges: Mantle Convection and Formation of the Lithosphere Millennial-Scale Climate Variability

J T Andrews

Mineral Extraction, Authigenic Minerals Molluskan Fisheries Monsoons, History of Moorings

G Ito, R A Dunn

852 867 881

J C Wiltshire

890

V S Kennedy

899

N Niitsuma, P D Naidu

910

R P Trask, R A Weller

919

Volume 4 Nekton

W G Pearcy, R D Brodeur

Nepheloid Layers

1

I N McCave

Network Analysis of Food Webs

8 J H Steele

Neutral Surfaces and the Equation of State Nitrogen Cycle

19 T J McDougall, D R Jackett

D M Karl, A F Michaels

Nitrogen Isotopes in the Ocean Noble Gases and the Cryosphere Non-Rotating Gravity Currents North Atlantic Oscillation (NAO) North Sea Circulation

25 32

D M Sigman, K L Karsh, K L Casciotti

40

M Hood

55

P G Baines

59

J W Hurrell

65

M J Howarth

73

Nuclear Fuel Reprocessing and Related Discharges

H N Edmonds

82

Ocean Biogeochemistry and Ecology, Modeling of

N Gruber, S C Doney

89

Ocean Carbon System, Modeling of Ocean Circulation

S C Doney, D M Glover

N C Wells

105 115

Ocean Circulation: Meridional Overturning Circulation

J R Toggweiler


126

Contents

Ocean Gyre Ecosystems

M P Seki, J J Polovina

Ocean Margin Sediments Ocean Ranching

132

S L Goodbred Jr

138

A G V Salvanes

146

R G Williams

156

Ocean Subduction

Ocean Thermal Energy Conversion (OTEC) Ocean Zoning

S M Masutani, P K Takahashi

M Macleod, M Lynch, P Hoagland

Offshore Sand and Gravel Mining Oil Pollution

Okhotsk Sea Circulation

E Garel, W Bonne, M B Collins

200

H Yamazaki, H Burchard, K Denman, T Nagai

Open Ocean Convection

A Soloviev, B Klinger

Open Ocean Fisheries for Deep-Water Species

Optical Particle Characterization

P H Burkill, C P Gallienne

265 272 274

R Lukas

287

E Thomas

295 W W Hay

Paleoceanography: Orbitally Tuned Timescales Paleoceanography: the Greenhouse World Particle Aggregation Dynamics Past Climate from Corals

T D Herbert

M Huber, E Thomas

A Alldredge

A G Grottoli

K L Denman, J F Dower

Pelagic Biogeography

A Longhurst

D H Cushing

Pelecaniformes

Peru–Chile Current System

C A Paulson, W S Pegau

J Karstensen, O Ulloa

319 330 338 348 356

379 385 393

K C Ruttenberg

Photochemical Processes

311

370

M Rubega

Phosphorus Cycle

303

364

D Siegel-Causey

Penetrating Shortwave Radiation

252 261

I Laing

Paleoceanography, Climate Models in

Phytobenthos

R A Jahnke

K K Turekian

Pacific Ocean Equatorial Currents

Phalaropes

243

G F Lane-Serff

Oysters – Shellfish Farming

Pelagic Fishes

234

K K Turekian

Oxygen Isotopes in the Ocean

Paleoceanography

226

J Joseph

Organic Carbon Cycling in Continental Margin Environments

Overflows and Cascades

208 218

J D M Gordon

Open Ocean Fisheries for Large Pelagic Species

Origin of the Oceans

182 191

L D Talley

One-Dimensional Models

167 174

J M Baker

Patch Dynamics

xxxvii

N V Blough

M Wilkinson

401 414 425


xxxviii

Contents

Phytoplankton Blooms

D M Anderson

Phytoplankton Size Structure Plankton

M M Mullin

Plankton and Climate Plankton Viruses

432

E Maranõń

445

w

453

A J Richardson

455

J Fuhrman, I Hewson

465

Platforms: Autonomous Underwater Vehicles Platforms: Benthic Flux Landers

J G Bellingham

R A Jahnke

485

Platinum Group Elements and their Isotopes in the Ocean

G E Ravizza

Plio-Pleistocene Glacial Cycles and Milankovitch Variability Polar Ecosystems

K H Nisancioglu

A Clarke

Pollution, Solids

494 504 514

C M G Vivian, L A Murray

Pollution: Approaches to Pollution Control Pollution: Effects on Marine Communities Polynyas

473

519

J S Grayw, J M Bewers R M Warwick

526 533

S Martin

540

Population Dynamics Models

Francois Carlotti

Population Genetics of Marine Organisms Pore Water Chemistry

546

D Hedgecock

556

D Hammond

Primary Production Distribution

563

S Sathyendranath, T Platt

572

Primary Production Methods

J J Cullen

578

Primary Production Processes

J A Raven

585

Procellariiformes

K C Hamer

590

Propagating Rifts and Microplates

Richard Hey

597

Protozoa, Planktonic Foraminifera

R Schiebel, C Hemleben

606

Protozoa, Radiolarians

O R Anderson

Radiative Transfer in the Ocean Radioactive Wastes Radiocarbon

C D Mobley

619

L Føyn

629

R M Key

637

Rare Earth Elements and their Isotopes in the Ocean Red Sea Circulation Redfield Ratio Refractory Metals

Y Nozaki

w

D Quadfasel

677

K J Orians, C L Merrin

Regime Shifts, Physical Forcing Regime Shifts: Methods of Analysis

653 666

T Tyrrell

Regime Shifts, Ecological Aspects

w

613

L J Shannon, A Jarre, F B Schwing F B Schwing B deYoung, A Jarre

Deceased.


687 699 709 717

Contents

Regional and Shelf Sea Models

J J Walsh

Remote Sensing of Coastal Waters

Rigs and offshore Structures River Inputs

722

N Hoepffner, G Zibordi

Remotely Operated Vehicles (ROVs)

xxxix

732

K Shepherd

742

C A Wilson III, J W Heath

748

J D Milliman

754

Rocky Shores

G M Branch

762

Rogue Waves

K Dysthe, H E Krogstad, P Mu¨ller

770

Rossby Waves

P D Killworth

Rotating Gravity Currents

781

J A Whitehead

790

Volume 5 Salmon Fisheries, Atlantic

P Hutchinson, M Windsor

Salmon Fisheries, Pacific Salmonid Farming Salmonids

1

R G Kope

L M Laird

12

w

23

D Mills

29

Salt Marsh Vegetation

C T Roman

Salt Marshes and Mud Flats Sandy Beaches, Biology of Satellite Altimetry

39

J M Teal

43

A C Brown

49

R E Cheney

58

Satellite Oceanography, History, and Introductory Concepts J R Apel w

W S Wilson, E J Lindstrom, 65

Satellite Passive-Microwave Measurements of Sea Ice

C L Parkinson

80

Satellite Remote Sensing of Sea Surface Temperatures

P J Minnett

91

Satellite Remote Sensing SAR

A K Liu, S Y Wu

Satellite Remote Sensing: Ocean Color

C R McClain

Satellite Remote Sensing: Salinity Measurements Science of Ocean Climate Models Sea Ice

103

G S E Lagerloef

S M Griffies

P Wadhams

114 127 133 141

Sea Ice Dynamics

M Leppa¨ranta

159

Sea Ice: Overview

W F Weeks

170

Sea Level Change

J A Church, J M Gregory

179

Sea Level Variations Over Geologic Time Sea Otters

w

M A Kominz

J L Bodkin

185 194

Deceased.


xl

Contents

Sea Surface Exchanges of Momentum, Heat, and Fresh Water Determined by Satellite Remote Sensing L Yu

202

Sea Turtles

212

F V Paladino, S J Morreale

Seabird Conservation

J Burger

Seabird Foraging Ecology Seabird Migration

220

L T Balance, D G Ainley, G L Hunt Jr

L B Spear

227 236

Seabird Population Dynamics

G L Hunt Jr

247

Seabird Reproductive Ecology

L S Davis, R J Cuthbert

251

Seabird Responses to Climate Change Seabirds and Fisheries Interactions

David G Ainley, G J Divoky C J Camphuysen

Seabirds as Indicators of Ocean Pollution Seabirds: An Overview Seals

265

W A Montevecchi

274

G L Hunt, Jr

279

I L Boyd

285

Seamounts and Off-Ridge Volcanism Seas of Southeast Asia

R Batiza

292

J T Potemra, T Qu

Seaweeds and their Mariculture Sediment Chronologies

305

T Chopin, M Sawhney

317

J K Cochran

327

Sedimentary Record, Reconstruction of Productivity from the Seiches

Seismic Structure

I Fer, W S Holbrook

L M Dorman

367

K B Katsaros

375

Sensors for Micrometeorological and Flux Measurements Shelf Sea and Shelf Slope Fronts

J B Edson

J Sharples, J H Simpson

H L Kite-Powell

Single Point Current Meters

T I Eglinton, A Pearson

P Collar, G Griffiths

436

Slides, Slumps, Debris Flows, and Turbidity Currents Small Pelagic Species Fisheries Small-Scale Patchiness, Models of

419 428

T J O’Shea, J A Powell G Shanmugam

R L Stephenson, R K Smedbol D J McGillicuddy Jr

Small-Scale Physical Processes and Plankton Biology

J F Dower, K L Denman

M Fieux

447 468 474 488 494

A B Baggeroer

Southern Ocean Fisheries

391

409

Single Compound Radiocarbon Measurements

Sonar Systems

382

401

R P Dinsmore

Somali Current

351 361

Sensors for Mean Meteorology

Shipping and Ports

333 344

A Harding

Seismology Sensors

Sirenians

G Wefer, W H Berger

D C Chapman, G S Giese

Seismic Reflection Methods for Study of the Water Column

Ships

257

504

I Everson


513

Contents

Sphenisciformes

L S Davis

520

Stable Carbon Isotope Variations in the Ocean Storm Surges

K K Turekian

529

R A Flather

530

Sub Ice-Shelf Circulation and Processes Submarine Groundwater Discharge Sub-Sea Permafrost Surface Films

xli

K W Nicholls

541

W S Moore

551

T E Osterkamp

559

W Alpers

570

Surface Gravity and Capillary Waves

W K Melville

573

Volume 6 Temporal Variability of Particle Flux Thermal Discharges and Pollution

W Deuser

1

T E L Langford

10

Three-Dimensional (3D) Turbulence Tidal Energy Tides

W D Smyth, J N Moum

A M Gorlov

26

D T Pugh

Tomography

32

P F Worcester

Topographic Eddies Towed Vehicles

40

J H Middleton

57

I Helmond

Trace Element Nutrients

65

W G Sunda

Tracer Release Experiments

75

A J Watson, J R Ledwell

Tracers of Ocean Productivity

Transmissometry and Nephelometry Tritium–Helium Dating

87

W J Jenkins

93

Transition Metals and Heavy Metal Speciation

Tsunami

18

J Donat, C Dryden

100

C Moore

109

W J Jenkins

119

P L-F Liu

127

Turbulence in the Benthic Boundary Layer Turbulence Sensors

R Lueck, L St. Laurrent, J N Moum

N S Oakey

Under-Ice Boundary Layer

148

M G McPhee, J H Morison

Upper Ocean Heat and Freshwater Budgets Upper Ocean Mean Horizontal Structure Upper Ocean Mixing Processes

155

P J Minnett

163

M Tomczak

175

J N Moum, W D Smyth

185

Upper Ocean Structure: Responses to Strong Atmospheric Forcing Events Upper Ocean Time and Space Variability Upper Ocean Vertical Structure Upwelling Ecosystems

141

L K Shay

192

D L Rudnick

211

J Sprintall, M F Cronin

217

R T Barber

Uranium-Thorium Decay Series in the Oceans: Overview

225 M M R van der Loeff


233

xlii

Contents

Uranium-Thorium Series Isotopes in Ocean Profiles Vehicles for Deep Sea Exploration

S E Humphris

Viral and Bacterial Contamination of Beaches Volcanic Helium

J Bartram, H Salas, A Dufour

285

Water Types and Water Masses

W J Emery

291

M E McCormick, D R B Kraemer

Waves on Beaches

267 277

E L Kunze

Wave Generation by Wind

244 255

J E Lupton

Vortical Modes

Wave Energy

S Krishnaswami

300

J A T Bye, A V Babanin

304

R A Holman

310

Weddell Sea Circulation

E Fahrbach, A Beckmann

318

Wet Chemical Analyzers

A R J David

326

Whitecaps and Foam

E C Monahan

Wind- and Buoyancy-Forced Upper Ocean Wind Driven Circulation

331 M F Cronin, J Sprintall

P S Bogden, C A Edwards

Zooplankton Sampling with Nets and Trawls

337 346

P H Wiebe, M C Benfield

355

Appendix 1. SI Units and Some Equivalences

373

Appendix 2. Useful Values

376

Appendix 3. Periodic Table of the Elements

377

Appendix 4. The Geologic Time Scale

378

Appendix 5. Properties of Seawater

379

Appendix 6. The Beaufort Wind Scale and Seastate

384

Appendix 7. Estimated Mean Oceanic Concentrations of the Elements

386

Appendix 8. Abbreviations

389

Appendix 9. Taxonomic Outline Of Marine Organisms

L P Madin

401

Appendix 10. Bathymetric Charts of the Oceans

412

Index

421


ABRUPT CLIMATE CHANGE S. Rahmstorf, Potsdam Institute for Climate Impact Research, Potsdam, Germany & 2009 Elsevier Ltd. All rights reserved.

Introduction High-resolution paleoclimatic records from ice and sediment cores and other sources have revealed a number of dramatic climatic changes that occurred over surprisingly short times – a few decades or in some cases a few years. In Greenland, for example, temperature rose by 5–10 1C, snowfall rates doubled, and windblown dust decreased by an order of magnitude within 40 years at the end of the last glacial period. In the Sahara, an abrupt transition occurred around 5500 years ago from a relatively green shrubland supporting significant populations of animals and humans to the dry desert we know today. One could define an abrupt climate change simply as a large and rapid one – occurring faster than in a given time (say 30 years). The change from winter to summer, a very large change (in many places larger than the glacial–interglacial transition) occurring within 6 months, is, however, not an abrupt change in climate (or weather), it is rather a gradual transition following the solar forcing in its near-sinusoidal path. The term ‘abrupt’ implies not just rapidity but also reaching a breaking point, a threshold – it implies a change that does not smoothly follow the forcing but is rapid in comparison to it. This physical definition thus equates abrupt climate change with a strongly nonlinear response to the forcing. In this definition, the quaternary transitions from glacial to interglacial conditions and back, taking a few hundred or thousand years, are a prime example of abrupt climate change, as the underlying cause, the Earth’s orbital variations (Milankovich cycles), have timescales of tens of thousands of years. On the other hand, anthropogenic global warming occurring within a hundred years is not as such an abrupt climate change as long as it smoothly follows the increase in atmospheric carbon dioxide. Only if global warming triggered a nonlinear response, like a rapid ocean circulation change or decay of the West Antarctic Ice Sheet (WAIS), would one speak of an abrupt climate change.

Paleoclimatic Data A wealth of paleoclimatic data has been recovered from ice cores, sediment cores, corals, tree rings, and

other sources, and there have been significant advances in analysis and dating techniques. These advances allow a description of the characteristics of past climatic changes, including many abrupt ones, in terms of geographical patterns, timing, and affected climatic variables. For example, the ratio of oxygen isotopes in ice cores yields information about the temperature in the cloud from which the snow fell. Another way to determine temperature is to measure the isotopic composition of the nitrogen gas trapped in the ice, and it is also possible to directly measure the temperature in the borehole with a thermometer. Each method has advantages and drawbacks in terms of time resolution and reliability of the temperature calibration. Dust, carbon dioxide, and methane content of the prehistoric atmosphere can also be determined from ice cores. On long timescales, climatic variability throughout the past 2 My at least has been dominated by the Milankovich cycles in the Earth’s orbit around the sun – the cycles of precession, obliquity, and eccentricity with periods of roughly 23 ky, 41 ky, and 100 ky, respectively. Since the middle Pleistocene transition 1.2 Ma, the regular glaciations of our planet follow the 100-ky eccentricity cycle; even though this has only a rather weak direct influence on the solar radiation reaching the Earth, it modulates the much stronger other two cycles. The prevalence of the 100-ky cycle in climate is thus apparently a highly nonlinear response to the forcing that is likely linked to the nonlinear continental ice sheet and/or carbon cycle dynamics. The terminations of glaciations occur rather abruptly (Figure 1). Greenland ice cores show that the transition from the last Ice Age to the warm Holocene climate took about 1470 years, with much of the change occurring in only 40 years. The local Greenland response is not typical for the global response; however, since Greenland temperatures can be strongly affected by Atlantic Ocean circulation, which went through rapid changes during deglaciation. Globally, the transition from full Ice Age to Holocene conditions took around 5 ky. The ice ages were not just generally colder than the present climate but were also punctuated by abrupt climatic transitions. The best evidence for these transitions, known as Dansgaard–Oeschger (D/O) events, comes from the last ice age (Figure 2). D/O events typically start with an abrupt warming by up to 12 1C within a few decades or less, followed by gradual cooling over several hundred or thousand


1

2

ABRUPT CLIMATE CHANGE

18

O

LGM

I

II

III

IV

V

Eemian _ 500

_ 450

_ 400

_ 350

_ 300

_ 250

_ 200

_ 150

Holocene _ 50 0

_100

Age (1000 yr BP) Figure 1 Record of d18O from marine sediments (arbitrary units), reflecting mainly the changes in global ice volume during the past 50 ky. Note the rapid terminations (labeled with roman numbers) of glacial periods.

18

O (ppm)

_ 34

1 20 19 18

_ 38

17 161514

12 13

11 8 76 43 10 5 9

0

2

ΔT

_ 42

100

_ 20 80

60

40

20

0

Age (1000 yr BP) Figure 2 Record d18O from the GRIP ice core, a proxy for atmospheric temperature over Greenland (approximate temperature range DT (1C) is given on the right). Note the relatively stable Holocene climate during the past 10 ky and before that the much colder glacial climate punctuated by Dansgaard–Oeschger warm events (numbered). The timing of Heinrich events 1 to 6 is marked by black dots.

years. The cooling phase often ends with an abrupt final temperature drop back to cold (‘stadial’) conditions. Although first seen in the Greenland ice cores, the D/O events are not a local feature of Greenland climate. Figure 3 shows that subtropical sea surface temperatures in the Atlantic closely mirror the sequence of events in Greenland. Similar records have been found near Santa Barbara, California, in the Cariaco Basin off Venezuela, and off the coast of India. D/O climate change is centered on the North Atlantic and regions with strong atmospheric response to changes in the North Atlantic, with little response in the Southern Ocean or Antarctica. The ‘waiting time’ between successive D/O events is most often around 1470 years or, with decreasing probability, multiples of this period. This suggests the existence of an as yet unexplained 1470year cycle that often (but not always) triggers a D/O event. The second major type of abrupt event in glacial times is the Heinrich (H) event. H events involve surging of the Laurentide Ice Sheet through Hudson Strait, occurring in the cold stadial phase of some D/O cycles. They have a variable spacing of several thousand years. The icebergs released to the

North Atlantic during H events leave telltale dropstones in the ocean sediments when they melt, the socalled Heinrich layers. Sediment data suggest that H events shut down or at least drastically reduce the formation of North Atlantic Deep Water (NADW). Records from the South Atlantic and parts of Antarctica show that the cold H events in the North Atlantic were associated with unusual warming there (a fact sometimes referred to as ‘bipolar seesaw’). At the end of the last glacial, a particularly interesting abrupt climatic change took place, the so-called Younger Dryas event (12 800–11 500 years ago). Conditions had already warmed to near-interglacial conditions and continental ice sheets were retreating, when within decades the climate in the North Atlantic region switched back to glacial conditions for more than a thousand years. It has been speculated that the cooling resulted from a sudden influx of fresh water into the North Atlantic through St. Lawrence River, when an ice barrier holding back a huge meltwater lake on the North American continent broke. This could have shut down the Atlantic thermohaline circulation (i.e., the circulation driven by temperature and salinity differences), but evidence is controversial.


3

22 20 18 _ 35

16

12 13

11 10

8 9

_ 40

GISP2 18O (ppm)

Temperature (°C)


_ 45 60

50

40

30

Age (1000 yr BP) Figure 3 Sea-surface temperatures derived from alkenones in marine sediments from the subtropical Atlantic (Bermuda Rise, upper curve) compared to d18O values from the GISP2 ice core in Greenland (lower curve).

Alternatively, the Younger Dryas may simply have been the last cold stadial period of the glacial following a temporary D/O warming event. Does abrupt climate change occur only during glacial times? Early evidence for the last interglacial, the Eemian, suggested abrupt changes there, but has since been refuted. During the present interglacial, the Holocene, climate was much more stable than during the last glacial. However, two abrupt events stand out. One is the 8200-year event that shows up as a cold spike in Arctic ice cores and affected the North Atlantic region. The second major change is the abrupt desertification of the Sahara 5500 years ago. There is much evidence from cave paintings, fire remains, bones, ancient lake sediments, and the like that the Sahara was a partly swampy savannah before this time. The best evidence for the abrupt ending of this benign climate comes from Atlantic sediments off northeastern Africa, which show a sudden and dramatic step-function increase in windblown dust, witnessing a drying of the adjacent continent.

Mechanisms of Abrupt Climate Change The increased spatial coverage, quality, and time resolution of paleoclimatic data as well as advances in computer modeling have led to a greater understanding of the mechanisms of abrupt climate change, although many aspects are still in dispute and not fully understood. The simplest concept for a mechanism causing abrupt climatic change is that of a threshold. A gradual change in external forcing (e.g., the change in insolation due to the Milankovich cycles) or in an internal climatic parameter (e.g., the slow buildup or melting of continental ice) continues until a specific

threshold value is reached where some qualitative change in climate is triggered. Various such critical thresholds are known to exist in the climate system. Continental ice sheets may have a stability threshold where they start to surge; the thermohaline ocean circulation has thresholds where deep-water formation shuts down or shifts location; methane hydrates in the seafloor have a temperature threshold where they change into the gas phase and bubble up into the atmosphere; and the atmosphere itself may have thresholds where large-scale circulation regimes (such as the monsoon) switch. For the D/O events, H events, and the Younger Dryas event discussed above, the paleoclimatic data clearly point to a crucial role of Atlantic Ocean circulation changes. Modeling and analytical studies of the Atlantic thermohaline circulation (sometimes called the ‘conveyor belt’) show that there are two positive feedback mechanisms leading to threshold behavior. The first, called advective feedback, is caused by the large-scale northward transport of salt by the Atlantic currents, which in turn strengthens the circulation by increasing density in the northern latitudes. The second, called convective feedback, is caused by the fact that oceanic convection creates conditions favorable for further convection. These (interconnected) feedbacks make convection and the large-scale thermohaline circulation self-sustaining within certain limits, with well-defined thresholds where the circulation changes to a qualitatively different mode. Three main circulation modes have been identified both in sediment data and in models (Figure 4): (1) a warm or interglacial mode with deep-water formation in the Nordic Seas and large oceanic heat transport to northern high latitudes (Figure 4(a)); (2) a cold or stadial mode with deep-water formation south of the shallow sill between Greenland, Iceland, and Scotland and with greatly reduced heat transport


4


(a) 0

Depth (km)

1 2 3 4

‘Warm’

5 (b) 0

Depth (km)

1 2 3 4

‘Cold’

5 (c) 0

Depth (km)

1 2 3 4 5 30° S

‘Off’ 0°

30° N

60° N

90° N

Figure 4 Schematic of three major modes of Atlantic Ocean circulation. (a) ‘Warm’ or interglacial mode; (b) ‘cold’ or stadial mode; (c) ‘off’ or Heinrich mode. In the warm mode the Atlantic thermohaline circulation reaches north over the Greenland– Iceland–Scotland ridge into the Nordic Seas, while in the cold mode it stops south of Iceland. Switches between circulation modes at certain thresholds can pace and amplify climatic changes.

to high latitudes (Figure 4(b)); and (3) a ‘switched off’ or ‘Heinrich’ mode with practically no deepwater formation in the North Atlantic (Figure 4(c)). In the last mode, the Atlantic deep circulation is dominated by inflow of Antarctic Bottom Water (AABW) from the south. Many features of abrupt glacial climate can be explained by switches between these three circulation modes. Model simulations suggest that the cold stadial mode is the only stable mode in a glacial climate; it prevails during the cold stadial periods of the last glacial. D/O events can be interpreted as temporary incursions of warm Atlantic waters into the Nordic Seas and deep-water formation there, that is, a switch to the warm mode causing abrupt climatic warming in the North Atlantic region. As this mode is not stable in glacial conditions, the

circulation starts to gradually weaken and temperatures start to decline again immediately after the incursion, until the threshold is reached where convection in the Nordic Seas stops and the system reverts to the stable stadial mode. H events can be interpreted as a switch from the stadial mode to the H mode, that is, a shutdown of North Atlantic deepwater formation. As this mode is probably also unstable in glacial conditions, the system spontaneously reverts to the stadial or to the warm mode after a waiting time of centuries, the timescale being determined partly by slow oceanic mixing processes. This interpretation is consistent with the observed patterns of surface temperature change. The warming during D/O events is centered on the North Atlantic because this is where the change in oceanic heat transport occurs; the warm mode delivers heat to much higher latitudes than does the cold mode. A switch to the H mode, on the other hand, strongly reduces the interhemispheric heat transport from the South Atlantic to the North Atlantic. This cools the Northern Hemisphere while warming the Southern Hemisphere, explaining the ‘bipolar seesaw’ response in climate. It should also be noted that the initial transient response can differ from the equilibrium response as the oceanic heat storage capacity is large. The patterns of these abrupt changes differ from the longer-timescale (many thousands of years) response to the Milankovich cycles because, for the latter, the slow changes in atmospheric greenhouse gases (e.g., CO2) and continental ice cover act to globally synchronize and amplify climatic change. While the threshold behavior of the Atlantic ocean can dramatically shape and amplify climatic change, the question remains what triggers the mode switches. As mentioned above, D/O switches appear to be paced by an underlying 1470-year cyclicity that is as yet unexplained. This could either be an external (astronomical or solar) cycle or an internal oscillation of the climate system, perhaps also involving the Atlantic thermohaline circulation. A superposition of two major shorter solar cycles can, in climate model simulations, trigger events spaced 1470 years apart. The irregularity in D/O event timing is probably the result of the presence of stochastic variability in the climate system as well as the presence of longer-term trends such as the slow buildup of large continental ice sheets. The ocean circulation change during H events, on the other hand, can be explained by the large amounts of fresh water entering the North Atlantic at these times in the form of icebergs. Simulations show that the observed amounts of fresh water are sufficient to shut down deep-water formation in the North Atlantic. The nonlinear dynamics of ice sheets



provide a plausible trigger mechanism. Ice sheets may grow for many thousands of years until their base melts owing to geothermal heating, when the ice sheet becomes unstable and surges. Thresholds in ocean circulation and continental or sea ice dynamics are not the only mechanisms that can cause abrupt climatic changes. In the desertification of the Sahara in the mid-Holocene, probably neither of these mechanisms were involved. Rather, an unstable positive feedback between vegetation cover (affecting albedo and evapotranspiration) and monsoon circulation in the atmosphere appears to have been responsible. It is almost certain that there are further nonlinearities in the climate system that could have caused abrupt climatic changes in the past or may do so in the future. We are only beginning to understand abrupt climate change, and the interpretations presented here – while consistent with data and model results – are not the only possible interpretations. Reflecting the state of this science, they are current working hypotheses rather than established and welltested theory.

Risk of Future Abrupt Changes The prevalence of abrupt nonlinear (rather than smooth and gradual) climatic change in the past naturally leads to the question whether such changes can be expected in the future, either by natural causes or by human interference. The main outside driving forces of past climatic changes are the Milankovich cycles. Close inspection of these cycles as well as modeling results indicate that we are presently enjoying an unusually quiet period in the climatic effect of these cycles, owing to the present minimum in eccentricity of the Earth’s orbit. The next large change in solar radiation that could trigger a new ice age is probably at least 30 ky away. If this is correct, it makes the Holocene an unusually long interglacial, comparable to the Holstein interglacial that occurred around 400 ka when the Earth’s orbit went through a similar pattern. This stable orbital situation leaves unpredictable events (such as meteorite impacts or a series of extremely large volcanic eruptions) and anthropogenic interference as possible causes for abrupt climatic changes in the lifetime of the next few generations of humans. Significant anthropogenic warming of the lower atmosphere and ocean surface will almost certainly occur in this century, raising concerns that nonlinear thresholds in the climate system could be exceeded and abrupt changes could be triggered at some point. Processes that have been (rather speculatively)

5

mentioned in this context include a collapse of the WAIS, a strongly enhanced greenhouse effect due to melting of permafrost or triggering of methane hydrate deposits at the seafloor, a large-scale wilting of forests when drought-tolerance thresholds are exceeded, nonlinear changes in monsoon regimes, and abrupt changes in ocean circulation. Of those possibilities, the risk of a change in ocean circulation is probably the best understood and perhaps also the least unlikely. Two factors could weaken the circulation and bring it closer to a threshold: the warming of the surface and a dilution of high-latitude waters with fresh water. The latter could result from an enhanced atmospheric water cycle and precipitation as well as meltwater runoff from Greenland and other glaciers. Both warming and fresh water input reduce surface density and thereby inhibit deep-water formation. Model simulations of global-warming scenarios so far suggest three possible responses: a shutdown of convection in the Labrador Sea, one of the two main NADW formation sites; a complete shutdown of NADW formation (i.e., similar to a switch to the H mode); and a shutdown of AABW formation. A transition to the stadial circulation mode has so far not been simulated, perhaps because convection in the Nordic Seas is strongly wind driven and is more effectively switched off by increased sea ice cover than by warming. A shutdown of Labrador Sea convection would be a significant qualitative change in the Atlantic Ocean circulation, but would probably affect only the surface climate of a smaller region surrounding the Labrador Sea. Effects on ecosystems and fisheries have not been investigated but could be severe. A complete shutdown of NADW formation would have wider climatic repercussions. Temperatures in northwestern Europe could initially rise several degrees in step with global warming, then abruptly drop back to near present values (the competing effects of raised atmospheric CO2 and reduced oceanic heat transport almost balancing). If CO2 levels decline again in future centuries as expected, European temperatures could remain several degrees below present as the Atlantic thermohaline circulation is not expected to recover perhaps for millennia. Further effects of a shutdown of deep-water renewal include reduced oceanic uptake of CO2 (enhancing the greenhouse effect), shifts in tropical rainfall belts, accelerated global sea level rise (due to a faster warming of the deep oceans), and rapid regional sea level rise in the northern Atlantic. The probability of major climatic thresholds being crossed in the coming centuries is difficult to establish and largely unknown. Currently, this possibility


6


lies within the (still rather large) uncertainty range for future climate projections, so the risk cannot be ruled out. The IPCC 4th assessment report assigns a probability of up to 10% to a shutdown of the Atlantic overturning circulation within this century.

See also Authigenic Deposits. Deep-Sea Drilling Results. Past Climate from Corals. Ocean Circulation: Meridional Overturning Circulation.

Further Reading Abrantes F and Mix A (eds.) (1999) Reconstructing Ocean History – A Window Into the Future. New York: Plenum.

Broecker W (1987) Unpleasant surprises in the greenhouse? Nature 328: 123. Clark PU, Webb RS, and Keigwin LD (eds.) (1999) Mechanisms of Global Climate Change at Millennial Time Scales. Washington, DC: American Geophysical Union. Clark PU, Alley RB, and Pollard D (1999) Northern Hemisphere ice sheet influences on global climate change. Science 286: 1104--1111. Houghton JT, Meira Filho LG, and Callander BA (1995) Climate Change 1995. Cambridge, UK: Cambridge University Press. Sachs JP and Lehman SJ (1999) Subtropical North Atlantic temperatures 60,000 to 30,000 years ago. Science 286: 756--759. Stocker T (2000) Past and future reorganisations in the climate system. Quarterly Science Review (PAGES Special Issue) 19: 301--319. Taylor K (1999) Rapid climate change. American Scientist 87: 320.


ABSORBANCE SPECTROSCOPY FOR CHEMICAL SENSORS R. Narayanaswamy, The University of Manchester, Manchester, UK F. Sevilla, III, University of Santo Tomas, Manila, The Philippines & 2009 Elsevier Ltd. All rights reserved.

the optical fiber to the phototransducer. The phototransducer generates an electrical signal that is related to the concentration of the analyte. This article focuses on chemical sensors that are based on absorbance spectroscopy. The various principles involved are reviewed, and some applications are presented.

Introduction

Absorbance and Reflectance Chemical sensors have introduced an alternative Spectroscopy technology for chemical measurements. These analytical devices generate an electrical signal in response to the presence of a specific substance, the signal being related to the concentration of the analyte. The application of these sensors has simplified chemical analysis, since the need for obtaining a laboratory sample is eliminated and a real-time and on-site measurement can be carried out. Furthermore, the configuration of these devices enables the detection with great sensitivity of low concentrations of chemical species. A number of analytical principles have been exploited in the development of chemical sensors. Among them are the optical methods of chemical analysis, which rely on the interaction of electromagnetic radiation with matter for the quantitation of a large number of substances. These methods provide a rapid and nondestructive tool for the measurement of chemical species. Optical methods have indeed played an important role, and continue to do so, in various field of chemical analysis. One group of chemical sensors that are based on the optical methods of chemical analysis is called optodes (or optrodes). Other types of optical sensing techniques that utilize waveguides, for example, surface plasmon resonance, also exist but are not reviewed here. The basic concept of optodes is depicted diagrammatically in Figure 1. An optode consists of an optical fiber, an optoelectronic instrumentation that incorporates a light source and a phototransducer, and a solid-phase molecular recognition element. Light from a suitable source is launched into the optical fiber and directed to the sensing zone which contains the molecular recognition element. The molecular recognition element reacts with the analyte, resulting in a modification of its optical property. This change is probed by the supplied radiation which is subsequently guided via

Ultraviolet (UV)/visible absorbance spectroscopy has been employed extensively in analytical chemistry. The basis for analysis here is the ability of the analyte or its derivative to absorb radiation that impinges into the measuring system. The radiation can have a wavelength occurring in the UV (200–400 nm), visible (400–780 nm), near-infrared (780–3000 nm), or infrared (3–50 mm) region. The absorption of radiation causes a reduction in the intensity of the radiation after it has passed through the system (Figure 2). This phenomenon is mathematically described by the Beer–Lambert law, as expressed by the following equation: I0 A ¼ log ¼ ecl It

½1

where A is the absorbance, c is the concentration of the absorbing species, I0 and It are the intensity of incident and transmitted light, respectively, e is the

Optoelectronic instrumentation Light source

Phototransducer

Optical fiber system

Molecular recognition element Figure 1 Basic design of optical fiber chemical sensor.


7

8

ABSORBANCE SPECTROSCOPY FOR CHEMICAL SENSORS

Cladding

(a)

(b)

Transmitted radiation (It)

Incident radiation (Io)

Core

Analyte solution

Figure 3 Structure of an optical fiber. (a) Longitudinal cross-section view, and (b) latitudinal cross-section view.

Figure 2 Absorption of radiation.

molar absorptivity of the species, and l is the optical path length of the absorbing species. Absorbance spectroscopy is applicable only when the measurand system containing the analyte is transparent. However, if the medium is optically dense or even opaque, in which case the absorbance measurements would produce a high background, a technique that is complementary to absorptiometry (viz., reflectometry) can be employed for analytical measurements. In this case, the radiation infringes on the boundary interface of two media having different dielectric constants and reflection occurs. Two distinct types of reflection are possible, namely (1) specular (or mirror-type) reflection and (2) diffuse reflection. Specular reflection occurs at the interface of a medium with no transmission through it, and reflection is at the same angle as the incident light; whereas, in diffuse reflection, the light penetrates the medium and subsequently reappears at the surface after partial absorption and multiple scattering within the medium. Of these two processes, diffuse reflection has found to be useful in chemical measurements. Specular reflection is minimized or eliminated through appropriate sample preparation and optical engineering. The distribution of diffusely reflected light is rather homogeneous and largely independent of size and shape of the particles. The optical characteristics of diffuse reflectance are dependent on the composition of the system. Among several theoretical models that have been proposed for diffuse reflectance, the most widely used is the Kubelka–Munk theory. Here, the scattering layer is assumed to be infinitely thick, which may be effectively the case with molecular recognition element utilized in optical sensors, and the reflectance (R) is related to the absorption coefficient (K) and the scattering coefficient (S), as follows: FðRÞ ¼

ð1 RÞ2 K ¼ 2R S

½2

where F(R) is known as the Kubelka–Munk function. The absorption coefficient K can be expressed in

terms of the molar absorptivity (e) and the concentration (c) of the absorbing species. Thus, eqn [2] can be rewritten as FðRÞ ¼

ð1 RÞ2 ec ¼ ¼ kc 2R S

½3

where k ¼ e/S, and S is assumed to be independent of concentration. Equation [3] is analogous to the Beer– Lambert relationship (eqn [1]) and holds true within a range of concentrations for solid solutions in which the absorber is adsorbed onto the surface of a scattering particle. The reflectance values (R) are generally evaluated relative to the reflectance of standard reference materials such as barium sulfate.

Optical Fibers Almost all optical sensors employ optical fibers to transmit light to and from the molecular recognition element. The fiber couples the optoelectronic instrumentation to the molecular recognition element, resulting in an integrated analytical system. This integration has contributed to the simplification of the chemical measurement process. Optical fibers consist of a cylinder (known as the ‘core’) of transparent dielectric with a certain refractive index (n1), surrounded by a thin film of another dielectric (called the ‘cladding’) of a lower refractive index (n2). Most optical fibers are then covered with a protective jacket that has no influence on the wave-guiding properties of the optical fiber. The basic structure of an optical fiber is shown in Figure 3. The common materials used in optical fibers include plastic (poly(methyl methacrylate)), glass, and quartz. Incident light is transmitted through the fiber when it impinges the core–cladding interface at an angle greater than the critical angle, so that there is total internal reflection at the core–cladding interface. A series of total internal reflection takes place until the light reaches the other end of the fiber (Figure 4). The optical fiber light transmission characteristics are described by its numerical aperture



Figure 4 Transmission of a light beam inside an optical fiber through total internal reflection.

Evanescent wave

Figure 5 Evanescent wave due to a light beam being transmitted inside an optical fiber.

(NA), which is directly proportional to the sine of the half angle (a) of the acceptance cone of light entering it which, in turn, is related to n1 and n2, as in eqn [4]: NA ¼ sin a ¼ ðn21 n22 Þ1=2

½4

where n0 is the refractive index of the surrounding medium, for example, air. Basically there are three kinds of optical fibers in use for sensing purposes. These are the multimode step index, the multimode graded index, and the single-mode step index fibers. These fibers differ in the number of light beams that travel through the length of the waveguide. The light rays accepted into the fiber interact, and only those which undergo constructive propagation traverse through the fiber. A small portion of light that is transmitted through an optical fiber or a waveguide by total internal reflection extends outside core and is referred to as the ‘evanescent wave’. In an optical fiber, the evanescent wave penetrates the cladding material (Figure 5). The intensity (I) of the evanescent wave decreases exponentially with increasing distance (d) from the surface according to eqn [5]: I ¼ I0 ed=dp

½5

where I0 is the electric field intensity at the core– cladding interface and dp, the depth of penetration, is the distance from the interface of the point at which the electric field intensity has reduced to 1/e of its value at the interface. This characteristic depth of penetration is dependent on wavelength of the propagating light, refractive indices of the core and the cladding, and the acceptance angle of the light

9

beam entering the fiber core. The distance dp is typically of the order of a fraction of a wavelength of light. The incorporation of optical fibers in chemical sensors imparts a number of advantages to optical sensors over the conventional devices in many areas of application. Optical sensors are electrically passive and immune to electromagnetic disturbances. They are geometrically flexible, corrosion-resistant, and capable of being miniaturized. They are compatible with telemetry and capable of operation in remote and hostile environments. They can be of low-cost, of rugged construction, and intrinsically safe. The optical fibers used in these sensors are capable of transmission of optical signals over great distances with low attenuation of optical power. Thus optical sensors are capable of measurements of samples in their dynamic environment, no matter how distant, difficult to reach, or harsh that environment is. Intrinsic safety aspects are imparted to these sensing devices by the low optical power utilized in them. Furthermore, the chemical sensing process itself is nonelectrical. With these sensors only very small sample volumes are needed for analysis, which has the advantages of nonperturbation of samples in real-time monitoring applications. However, optical sensing devices possess certain limitations, such as interference from ambient light, limited dynamic range, long response times, limited specificity, and nonreversibility. Many of these limitations can be eliminated or reduced by the use of appropriate instrumentation and sensing phases, and thus the sensor devices can be used advantageously in specific applications.

Optoelectronic Instrumentation Optical fiber sensors involve similar instrumentation as those involved in spectrophotometers. These sensors require both optical and electrical components, including a light source; a wavelength selector; a photodetector; and a readout device. A block diagram of the basic instrumentation associated with optical sensors is presented in Figure 6. The radiation supplied by the light source is first made monochromatic by a wavelength selector before it enters into optical fiber sensor system. Then, the radiation emanating from the sensor is directed to the photodetector which subsequently generates an electrical signal. This magnitude of the final signal is displayed in the readout device. Several types of sources, including incandescent lamps (tungsten and quartz–halogen), lasers, lightemitting diodes (LEDs), and laser diodes, have been


10


Light source

Signal modulator

Monochromator

Photodetector

Lock-in amplifier

Optical fiber Readout device

Measurand system

Figure 6 Schematic diagram of a typical instrumentation system associated with optical fiber chemical sensors.

used. Each of these types of source has its own advantages and disadvantages. The detection of light is carried out using a photocounting device, which converts optical signals into electrical signals that can be amplified electronically. The photodetectors used in optical sensors include photomultiplier tubes, p–i–n photodiodes, avalanche photodiodes, and photodiode arrays. A wavelength selector, such as a filter or a monochromator, isolates the desired wavelength for the measurement. Optical couplers and lenses are used to focus the beam of light to the optical fiber and to direct light to the detector. It should be emphasized that efficient coupling of light ensures the attainment of high sensitivity of sensor signals. It is also common practice to exclude extraneous light reaching the detector by suitable modulation of the light source and by synchronizing the detector to this modulation frequency so as to detect only those source signals. Instrumental drift may be eliminated or reduced by the use of a suitable referencing system. Drifts in sensor response due to aging of the sensing phase and variable light source may be encountered. But these contributions may not be significant due to short periods of measurement periods used with the sensing devices.

Molecular Recognition Element The molecular recognition element constitutes the primary sensing unit of a chemical sensor. It interfaces the chemical sensor with the measurand system, generating an optical signal in response to the presence of the analyte. It transforms the analyte into another substance with distinct optical properties, so that it can be considered as a chemical transduction unit. The transformation can be detected and quantified by the instrumentation system via the optical fiber. However, if the analyte itself possesses a detectable optical characteristic, then that property can be directly measured using optical fibers and quantified to the concentrations of the analyte.

For absorbance- and reflectance-based chemical sensors, a number of reagents employed in colorimetry can serve as a molecular recognition element. Chromogenic reagents, such as pH colorimetric indicators and chelating reagents, have been used as chemical transduction elements in optical sensors. Novel materials, such as conducting polymers, molecularly imprinted polymers, and nanoparticles, have also been exploited for molecular recognition in optical fiber chemical sensors. The molecular recognition element is often employed in the solid state. The reagent(s) is usually immobilized onto an inert and stable solid material by physical methods (adsorption, entrapment, and electrostatic attraction) or by chemical means (covalent bond formation). Among the solid supports that have been used to immobilize the molecular recognition element are glass, silica gel, or organic polymers. The physical methods are simple and economical to carry out, but they do not necessarily produce stable reagent matrices. On the other hand, the chemical means of immobilization produce a strongly bound reagent, but this is often achieved after several reaction steps in the synthesis or modification of the reagent and/or the support material in order to realize stable chemical bond(s) between them. The sensor response characteristics in chemical transducer-based sensors depend on the manner in which the analyte–reagent interaction takes place. In a simple system, where direct indicators are employed, the analyte concentrations can be correlated to the optical changes that occur in the reagent phase or in the product or both. The correlated sensor signal may also be dependent on the equilibrium constant of the analyte–reagent reaction. For example, pH can be measured by monitoring the changes in optical property of acid-base indicators. In many sensors, reversible reactions are preferred in the chemical transduction process because they can be used in continuous monitoring applications. In this case, the response time of these sensors (i.e., the



time to reach the reaction equilibrium) is dependent on mass transfer processes. Irreversible reactions may also be employed in sensors, which can result in ‘one-shot’ devices. Although such sensors would be of limited merit, measurements with high sensitivity can be attained here. In certain cases, the reagent phases can be regenerated by the use of another chemical reaction and the sensor reused. Indirect chemical reactions involving two or more reagents and/or reactions can also be adapted as optical chemical transducers. Many enzyme-based reactions fall into this category.

Sensor Design A variety of configurations have been utilized in sensors based on absorbance and other spectroscopic measurements. The sensor designs can be classified into two categories: extrinsic and intrinsic sensors. In extrinsic sensors, the optical fiber acts only as a light guide between the light source and chemical sensor and between the sensor and the detector. In intrinsic sensors, the optical fiber becomes a part of the transducer. In absorbance-based sensors, light is usually fed at one end of an optical fiber, guided to a sensing cell, collected through a second optical fiber, and detected at the other end of this fiber. If no chemical

(a)

From light source

11

transduction is employed (Figure 7(a)), the light interacts directly with the measurand system. This type of sensors has been described as ‘spectroscopic sensors’, since they are based on the spectral properties of the analyte. In sensors involving a molecular recognition element, the reagent phase is often translucent and is set at one end of the feed fiber and of the collector fiber (Figure7(b)). The analyte reacts with the immobilized reagent, and the product absorbs the light that passes through the reagent phase. In reflectance-based sensors, chemical transduction is often employed and the molecular recognition element is placed at one end of an optical fiber system. A single optical fiber or optical fiber bundles can be used in this type of optical sensors. In single-fiber sensors (Figure 8(a)), the source light and detected light travel through the same optical fiber, and are discriminated either temporally or by wavelength with the aid of a beam splitter. Optical fiber bundles are often configured as a bifurcated system (Figure 8(b)), wherein the incident radiation travels through one branch and the reflected light is directed to the photodetector through the other branch. Evanescent wave interactions have been exploited in optical fiber chemical sensors. In these devices, the cladding material of the optical fiber is removed and replaced with the analyte system itself (Figure 9(a))

Flow cell

To photodetector Optical fiber

Optical fiber

Analyte

(b)

From light source

Reagent film

Optical fiber

Optical fiber

To photodetector

Analyte inlet Figure 7 Typical configuration of absorbance-based chemical sensors: (a) employing no chemical transduction, and (b) with chemical transduction.


12


(a)

(b)

From light source

From light source

To photodetector

Beam splitter

To photodetector

Optical fiber

Optical fiber

Molecular recognition element

Membrane

Molecular recognition element

Membrane

Figure 8 Typical configuration of reflectance-based chemical sensors: (a) employing a single fiber, and (b) employing a bifurcated optical fiber bundle.

(a)

Cladding

Optical fiber

From light source

To photodetector

Flow cell Analyte inlet (b)

From light source

To photodetector

Cladding

Reagent phase

Optical fiber

Figure 9 Typical configuration of evanescent-wave chemical sensors: (a) not employing chemical transduction, and (b) with chemical transduction.

or with a thin layer of the molecular recognition element (Figure 9(b)). The presence of the analyte affects the evanescent wave interaction and generates changes in the optical signal. In optical fiber absorptiometry or reflectometry, reference signals are employed for correcting

variations in the background caused by source, detector, and also intrinsic absorption of light by optical fibers. This can be done by subtracting the blank signal from the sensor signal electronically, or by the use of two identical optical fibers, one of which is used as a reference.



Sensor Applications A number of optical fiber pH sensors have been developed based on the absorption characteristics of a colorimetric acid-base indicator. The reagent is immobilized on a finely powdered solid support or on a membrane that is deposited on the tip of the optical fiber and held in place by an enclosing membrane. One of the earliest reported optical fiber sensor for pH was based on the measurement of the absorbance of phenol red covalently bound to polyacrylamide mixed with polystyrene microspheres to scatter the light within the reagent phase. The sensor was configured as a probe employing a bifurcated optical fiber, with one arm for the probing radiation and the other arm for the scattered radiation from the reagent phase. The sensor measures pH in the range 7.0–7.4 to 0.01 pH units. This example illustrates the feasibility of the approach though the sensors may not be directly applicable to pH measurements in seawater. Different indicators may be employed for such applications. Absorbance-based optical sensing has been applied for the measurement of metal ions. These ion sensors involve the use of immobilized metal-ion-selective reagents interfaced to optical fibers. These sensors rely on the fact that the metal ion (M) reacts with the immobilized reagent (R) to form a metal complex (MR), accompanied by either an enhancement or change of color of the immobilized reagent. This change can be correlated to metal ion concentrations. Thus, an optical sensor for micromolar amounts of cobalt was developed employing a reagent phase consisting of pyrogallol red held on a cellulose acetate film. Likewise, a fiber optic sensor for copper ions occurring in ppm levels was devised based on a-benzoinoxime, a highly selective colorimetric reagent for copper ion, immobilized on hydrophobic reagent Amberlite XAD-2 microspheres. Optical sensors based on absorbance measurement have been developed for organic compounds, such as pesticides, occurring in environmental water systems. In the case of pesticide sensors, more than one type of transduction system is involved – an enzyme reagent and a pH indicator co-immobilized on suitable polymeric material. In the absence of pesticides, the reaction of the immobilized enzyme, such as acetylcholine esterase, with its specific substrate will be accompanied by a change in pH that can then be measured by the pH sensor using absorbance or reflectance. However, in the presence of the pesticides, the substrate/immobilized enzyme reaction is inhibited and the degree of inhibition, transduced in the pH sensor, can be correlated to the pesticide

13

concentration. Organophosphate and carbamate pesticides, and also many toxic metal ions, have been quantified using this principle. Absorbance measurement has also been used in gas-phase sensing. A nitrogen dioxide sensor has been constructed using optical fibers and measurement of absorbance of the gas at 0.5 mm using an argon-ion laser source. Real-time measurement in the lower ppm concentration range has been conducted at remote locations that are 20 km away. A similar device has been used for the measurement of methane at low concentrations by recording absorbance at 1.33 mm. Many sensors for gases and vapors such as ammonia, carbon dioxide, humidity, hydrogen cyanide, hydrogen sulfide, etc., have developed based on the use of analyte-specific immobilized reagents and measurements of absorbance or reflectance. For instance, the optical sensing of ammonia can be carried out through the change in color of polyaniline deposited on a polystyrene substrate. An optical sensor for gaseous hydrogen sulfide was based on the development of a grayish color on a cellulose membrane impregnated with lead acetate.

Conclusions As described above, many types of optical sensor designs have been studied for a variety of analytes using different types of transduction reactions in the development of absorbance- and reflectance-based sensors. A current trend in this development is to construct multianalyte sensing systems based on the use of single or a few reagent phases, together with the employment of appropriate signal-processing techniques such as pattern recognition and artificial neural networks. Most of the optical sensors described above can be designed for use in oceanographic measurements in the analysis of heavy metal ions, dissolved gases, and other species. Applications of absorbance-based sensors to ocean sciences have yet to demonstrate their potential. Most of the applications published in the literature describe ‘proof-of-concept’ studies with the sensors and with very little or no practical demonstration in such areas. Some of the recent studies in this area have been focused on the monitoring of dissolved CO2 in seawater using absorbance and fluorescence measurements with transducers that incorporate pH-sensitive indicators. The use of infrared absorption measured through the evanescent waves in optical fibers for subsea monitoring of organic compounds (at ppm levels) has also been demonstrated. Parameters such as sensitivity, specificity,


14


lifetime, aging, etc., need to be investigated and addressed before the sensors can be used for seawater measurements. These studies would present great challenges in order to realize practical sensors. However, there is substantial interest in new sensors for oceanographic applications including monitoring of nutrients and pollutants, and optical sensors have great potential here. The devices are clearly attractive in concept and require expertise from several scientific disciplines including analytical chemistry, polymer chemistry, environmental chemistry, fiber optics, and opto-electronics.

See also Fluorometry for Chemical Sensing. Inherent Optical Properties and Irradiance. Wet Chemical Analysers.

Further Reading Andres R, Kuswandi B, and Narayanaswamy R (2001) Optical fiber biosensors based on immobilized enzymes – a review. The Analyst 126: 1469--1491. Eggins BR (2002) Chemical Sensors and Biosensors. London: Wiley. Hales B, Burgess L, and Emerson S (1997) An absorbancebased fiber-optic sensor for CO2(aq) measurement in pure waters of sea floor sediments. Marine Chemistry 59: 51--62. Harmer AL and Narayanaswamy R (1988) Spectroscopic and fibre-optic transducers. In: Edmunds TE (ed.) Chemical Sensors, ch. 13. Glasgow: Blackie. Lieberzeit PA and Dickert FL (2007) Sensor technology and its applications in environmental analysis. Analytical and Bioanalytical Chemistry 387: 237--247. Mizaikoff B (1999) Mid-infrared evanescent wave sensor – a novel approach for sub-sea monitoring. Measurement Science and Technology 10: 1185--1194.

Narayanaswamy R (1991) Current developments in optical biochemical sensors. Biosensors and Bioelectronics 6: 467--475. Narayanaswamy R (1993) Chemical transducers based on fibre optics for environmental monitoring. The Science of the Total Environment 135: 103--113. Narayanaswamy R (1993) Optical chemical sensors: Transduction and signal processing. Analyst 118: 317--322. Narayanaswamy R and Sevilla FS, III (1988) Optical fibre sensors for chemical species. Journal of Physics E: Scientific Instruments 21: 10--17. Narayanaswamy R and Wolfbeis OS (2004) Springer Series on Chemical Sensors and Biosensors, Vol. 1: Optical Sensors – Industrial Environmental and Diagnostic Applications. Berlin: Springer. Oehme I and Wolfbeis OS (1997) Optical sensors for determination of heavy metal ions. Microchimica Acta 126: 177--192. Orellana G and Moreno-Bondi MC (2005) Springer Series on Chemical Sensors and Biosensors, Vol. 3: Frontiers in Chemical Sensors – Novel Principles and Techniques. Berlin: Springer. Rogers KR and Poziomek EJ (1996) Fiber optic sensors for environmental monitoring. Chemosphere 33: 1151--1174. Seitz WR (1988) Chemical sensors based on immobilised indicators and fiber optics. CRC Critical Reviews in Analytical Chemistry 19: 135--171. Sevilla F, III and Narayanaswamy R (2003) Optical chemical sensors and biosensors. In: Alegret S (ed.) Integrated Analytical Systems, ch. 9. Amsterdam: Elsevier. Tokar JM and Dickey TDN (2000) Chemical sensor technology. Current and future applications. Ocean Science and Technology 1: 303--329. Varney MS (2000) Chemical Sensors in Oceanography. London: Taylor and Francis. Wise DL and Wingard LB, Jr. (1991) Biosensors with Fiberoptics. Clifton, NJ: Humana. Wolfbeis OS (1991) Fiber Optic Chemical Sensors and Biosensors, vols. I and II. Boca Raton, FL: CRC Press.


ABYSSAL CURRENTS W. Zenk, Universita¨t Kiel, Kiel, Germany Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 12–28, & 2001, Elsevier Ltd.

Introduction Historically the term ‘abyss’ characterizes the dark, apparently bottomless ocean under extreme static pressure far beyond coastal and shelf areas. Today this ancient definition remains still rather unfocused in earth sciences. Geographers, marine biologists, and geologists use abyss for deep-sea regions with water depths exceeding 1000 or 4000 m. In physical oceanography a widely accepted definition of the abyss denotes the water column that ranges from the base of the main thermocline down to the seabed. The main thermocline itself – occasionally also called warm-water sphere – extends laterally between the polar frontal zones of both hemispheres. In contrast to deep strata the surface exposition of the thermocline allows the direct exchange of heat, substances, and kinetic energy with the atmosphere. This

wind-driven part of the water column and its fluctuations are consequently an immediate subject of weather and climatic conditions. The base of the thermocline at about 1000–1200 m represents the lower boundary of the warm-water sphere with temperatures well 451C. The abyss or cold-water sphere below, is clearly colder. Below 2000 m potential temperatures o41C are found virtually everywhere. Below 4000 m values of 0–21C are more characteristic. Until the advent of modern self-recording instrumentation abyssal currents were believed to be very slow (o2 cm s1) and negligible in comparison with rather vigorous and variable surface currents (sometimes 4100 cm s1). Only subsurface passages in rises and ridges that subdivide ocean basins (Figure 1) seemed to allow for more energetic deep interior currents funneling through gaps and channels. Until a few decades ago practically all knowledge about abyssal currents had to be inferred from the perspective of property fields like potential temperature, salinity, or potential density. Since these fields seem to be much more stable than velocity fields this method still yields reasonable general

60 60

N 30

N

0 30 30 S (B) 120 E

0

120 W

180

90 0

30

30 60 60 60

S

S (A)

60

30

W

0

(C)

30 E 60

90

120

Figure 1 Major subsurface fracture zones, passages, and sills of the world’s ocean. (A) Atlantic, (B) Pacific, (C) Indian Ocean. These topographic features represent important constraints for deep and bottom water spreading paths (after http:// www.soc.soton.ac.uk/JRD/OCCAM/sills.html).


15

16

ABYSSAL CURRENTS

large-scale flow patterns even with scarce and nonsynoptic data sets. However, for any resolution beyond the low-frequency band of the current spectrum direct current observations by floating objects, moored instruments, or remotely measuring methods are essential. In low latitudes of all oceans the permanent presence of the enormous, almost invariant temperature gradient between the surface at 4251C and the floor at o21C requires a deep advection of cold water. The latter circulates freshly ventilated cold water from polar latitudes towards the Equator. For reasons of continuity a compensating poleward flow of water heated from the surface through the thermocline must occur above the abyssal layer of the tropics and the subtropics. Hence, the cold water drift propelled by sinking of cooled water masses (convection in selected areas) accelerates an endless global circulation cell known as the meridional overturning circulation (MOC), occasionally called the global conveyor belt (Figure 2). Its bottomnearest limb, and sometimes also one or two layers in motion above it are characterized by the abyssal circulation.

The Stommel-Arons Concept and Diffusivity in the Interior The concept that rising water from a flat deep-sea bottom without density stratification has to be replaced by convectively formed water is a key element of the modern Stommel-Arons theory of abyssal circulation. While the freshly ventilated water sinks in only a few selected semi-enclosed polar regions, the rising process itself occurs over broad lateral scales everywhere at lower latitudes. As a consequence it is possible to distinguish between two different dynamical regimes of the abyssal circulation (Figure 3). (1) A small (typically only 100– 200 km wide) corridor at the continental rises and near the slopes on the western margins of the ocean’s basins is occupied by deep western boundary currents (DWBCs). (2) The huge remaining interiors of the basins, fed laterally by DWBCs, are ruled by a uniform broad-scale upwelling regime. The inherent vertical velocities imply vortex stretching on top of the abyss. Conservation of potential vorticity then requires the poleward return drift in the interior. Not long after the concept of DWBC was hypothesized in 1958, the first observations confirming its existence were made beneath the Gulf Stream in the north-western Atlantic using neutrally buoyant Swallow floats. Shortly afterwards, the newly detected

deep equatorward drift led to a supplementary experiment to find the slow poleward countercurrent of the interior flow. The test failed at the time (1962) in so far as no flow with a slow, persistently northward velocity component could be proved. Instead, the mesoscale eddy phenomenon at great depths was discovered. In steady state the vertical flux of heat at the base of the thermocline in lower latitudes can be formulated as the balance of temperature advection and its vertical diffusion: @T @2T ! ¼k 2 u rT þ w @z @z

½1

-

where u is the lateral current vector, T temperature, w vertical velocity component, z vertical coordinate, and k eddy diffusivity. This temperature equation implies for the interface between the base of the thermocline and the top of the abyss the generation of a vertical velocity to balance the upwelled cold water and the downward diffusion of heat. An assumed sinking rate of 20 106 m3 s1 in high latitudes and an active lower thermocline interface of 3 108 km6 yields a global integral upwelling speed of O(0.1 dm d1). An adjoined downward diffusion with a diffusivity of O(1 cm2 s1) can be estimated under the assumption of a vertical scale of 1 km within the abyssal upwelling regime. Figure 4 depicts a sketch of the integrated form of [1] under the assumption of negligible lateral currents along isotherms. w T ¼ kd

@T @z

½2

where kd is the cross-isothermal diffusivity. Observations of kd are difficult to conduct. On a global scale numerical values of kd fluctuate in a wide range (1–500 cm2 s1). A summary of the available sparse estimates from the Atlantic is reproduced in Figure 5. Figure 6 gives a rare example of the heterogeneity of cross-isopycnal diffusivity in the interior Brazil Basin. The diffusivity is quantified from uniform microstructure observations at all depths above the smooth abyssal plains and west of the Mid-Atlantic Ridge. Estimates range from very weak (o0.1 cm2 s1) to moderate (45 cm2 s1) turbulent signals. The latter enhancement is observed above the rough flanks of the ridge. The response is especially prominent within 500 m above the bottom. The abyss contains recirculation cells. Significant amounts of cold water do not participate directly in the global meridional overturning cell. They circulate horizontally in response to the upwelling process.



90°W 0°

C

90°E

INDIAN OCEAN

90°E

PACIFIC

Warm surface current

180°

Recirculated deep water

Cold saline bottom current

18°0

Figure 2 The Atlantic thermocline circulation as a key element of the global oceanic circulation. (After Broecker (1991), modified by Meier-Reimer.)

90°S

0°

TI

60°S

Heat release to atmosphere

AT N LA

30°S

0°

30°N

60°N

90°N

90°W

90°S

60°S

30°S

0°

30°N

60°N

90°N

ABYSSAL CURRENTS 17

18

ABYSSAL CURRENTS

Figure 3 The Stommel-Arons (1958) concept of the abyssal circulation in the Atlantic. A system of western boundary currents feeds a slow broad-scale upwelling regime in the remaining interior of the basins. (Reprinted from Deep-Sea Research 5, Stommel H, The abyssal circulation, pp. 80–82, Copyright (1958), with permission from Elsevier Science.)

Such basin-wide recirculation cells (Figure 7) are distinctly influenced by the local bottom topography. Their persistence remains widely unexplored. In summary, it is of considerable interest to quantify rates of sinking waters in high latitudes because the compensating abyssal upwelling is believed to drive the internal horizontal circulation of the oceans.

are interconnected by the Southern Ocean as the circum-Antarctic water ring is called. In this strongly simplified pattern of the global water mass circulation the abyss ranges from the lower third of the displayed water column down to the seafloor.

3

Sources of Abyssal Waters Figure 8 represents a refined global update of the highly schematized deep Atlantic circulation pattern in Figure 3. Antarctica lies in the center. The North Pole cap is split; it is situated towards the outer edges of the Pacific and Atlantic blocks. All three oceans

10 _ 400

3 3

100

kd ∂ T ∂z

1 °C

wT

4 Sv

1 _ 500

Vema Channel

Brazil Basin

Equator

Figure 4 Schematic representation of eqn [2], the balance of downward diffusivity and upwelling across the 11C isotherm level of the western South Atlantic. (After Hogg et al., 1994 ^ Springer Verlag.)

Figure 5 Some estimates of cross-isotherm diffusivity in cm2 s1 in the abyss of the world’s oceans. (After Hogg, 2001, ^ Academic Press.)



6000

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500

_36

0.1

_38

0 0.2

_34

0.3

_32

0.4

0.5

_30

Diffusivity

0.6 _ (10 4

_26

)

_ m2 s 1

0.7

Longitude

_28

0.8

_24

0.9

_22

2.0

5.0

_20

8.0

_18

22.0

_16

Figure 6 Directly observed distribution of diffusivity in the South Atlantic. The section covers the range between the continental rise off Brazil and the western flanks of the Middle Atlantic Ridge. Highest diffusivity values are correlated with the rough topography on the slopes of the ridge. (After Polzin et al., 1997 ^ Science.)

Water depth (m)

0

Brazil Basin

ABYSSAL CURRENTS 19

20

ABYSSAL CURRENTS 20°N

4000 4000 8

5

4000

00 20

10°

6

12

20

00

4000

0°

10°S 70°W

60°

40°

50°

30°

20°

10°

Figure 7 Example of abyssal recirculation cells in the tropical North Atlantic Ocean. Numbers indicate volume transports of lower North Atlantic Deep Water in 106 m3 s1. (After Friedrichs and Hall, 1993. Courtesy of the Journal of Marine Research.)

Regions of sinking waters are symbolized in Figure 8 by near-surface downward arrows. They are unevenly distributed in both hemispheres. In fact, they are limited to the Antarctic Circumpolar

Current System (ACCS) and to the high latitudes of the North Atlantic. In addition to bottom and deep water sources there are contributions from outflows of marginal seas (not shown in Figure 8). Although

ic

cif

Java Pa m Ind Fro ian Aus tr.

South Oceanern

Ind

one

sia

Pacific

ic

ant

Atl

Figure 8 The complex global circulation. Colors: red and purple, thermocline waters; green and blue, abyssal waters. For details see text. Abbreviations: SAMW, SubAntarctic Mode Water; LOIW, Lower Intermediate Water; SLW, Surface Layer Water; UPIW, Upper Intermediate Water; NIIW, North-west Indian Intermediate Water; BIW, Banda Intermediate Water; RSW, Red Sea Water; CDW, Circumpolar Deep Water; NADW, North Atlantic Deep Water; NPDW, North Pacific Deep Water; IODW, Indian Ocean Deep Water; AABW, Antarctic Bottom Water. (Reproduced with permission from the Woods Hole Oceanographic Institution, Schmitz 1996a, 1996b.)


ABYSSAL CURRENTS

80°

80°

80°

60°

40°

20°

80°

60°

60°

40°

80° 60°

80°

60° 40°

40°

60° 40°

40°

20°

20°

21

20°

0°

0° 20°

20°

40°

40° 60°

20°

60° 80°

80°

40°

40° 60°

< 0°C _ 0 1°C 1_ 2°C

60° 80°

80°

40° 60°

60° 80°

80°

> 2°C Inferred water movements

Figure 9 Global near-bottom potential temperature distribution and inferred flow paths of abyssal waters at 4000 m depth. (Reproduced with permission from an article by Charnock, The Atmosphere and the Ocean; in Summerhayes CP and Thorpe SA, Oceanography, An Illustrated Guide ^ 1996 Manson Publishing.)

70°

80° 60° 40° 100°

60° 20° 0° 20°

70°

60°

60°

2.9 0.5

5.2 5.1

0.5

3.5

50°

0.3

10.7

50°

13.3

3.9 2.4

40°

40°

60°

40°

0

Figure 10 Examples of deep entrainment from the subpolar North Atlantic. Numbers represent volume transports in 106 m3 s1. Note the observed significant increase of transport along the east Greenland side which is caused by Denmark Strait Overflow Water. The Denmark Strait between Iceland and Greenland is also called Greenland Strait. (Reproduced with permission from Dickson RR and Brown J, Journal of Geophysical Research 99, 12319–12341 (1994) ^ American Geophysical Union.)



_3600

_3400

_3200

_3000

_2800

_2600

_2400

_2200

_2000

_1800

_1600

_1400

_1200

_1000

_800

_600

_400

_200

0

1940

1940

1945

1945

1950

1950

1955

1955

1960

1960

1965

1965

Year

1970

1970

1975

1975

1980

1980

1985

1985

1990

1990

1995

1995

Figure 11 Demonstration of interannual variability (1938–97) of potential temperature in 1C (colors and white isotherms) and potential density d1.5 (kg m3) 1.5 (dashed black lines) referenced to 1500 dbar pressure level. (After I Yashayaev et al., www.mar.dfo-mpo.gc.ca/science/ocean/woce/labsea/labsea_poster.html.)

Pressure (db)

22 ABYSSAL CURRENTS

ABYSSAL CURRENTS

23

50° .1

.3 .4 .2

.8 40° nity

Sali

aly nom

.5

.6

.9

.7

a

1.0

30° N

Africa Meddy Observation

40°

50°

60°W

10°

20°

30°

Figure 12 Influence of the Mediterranean Water on the salinity distribution at 1100 m of the North Atlantic. Isohalines represent positive anomalies relative to 35.01. Circles depict observed high salinity eddies of Mediterranean Water (Meddies). (Reprinted from Progress in Oceanography 4, Richardson PL, Bower AS and Zenk W, A census of meddies tracked by floats, pp. 209–250 ^ (2000) with permission from Elsevier Science and the American Meteorological Society.)

0

34.3

35 34.58

34.6 500

34.54

Pressure (dbar)

NPIW 34.5 34.48

AAIW 34.52

1000

34.54

34.56 1500

4 ˚S

2 ˚S

0 ˚N

2 ˚N

4˚N Latitude

6˚N

8 ˚N

10˚N

12˚N

Figure 13 Salinity distribution (PSU) of Antarctic and North Pacific Intermediate Waters in the tropical western Pacific along 1501E. (After Holfort and Zenk, 1999.)


ABYSSAL CURRENTS

these flows reach only intermediate depth levels they affect characteristic water masses of the abyss beneath their own spreading level. Several discrete locations for water sinks have been identified around Antarctica. Completely homogenizing wintertime convection down to the bottom of this semi-enclosed basin was observed only in Bransfield Strait south of the South Shetland Islands. However, in all other locations around Antarctica cooled and freshly ventilated water sinks to the bottom where it behaves like a contour current on the rotating earth. These densest waters in thin layers on the bottom mix with surrounding slightly warmer water masses and finally are transformed into Antarctic Bottom Water (AABW) spreading equatorward in all oceans. Known sinking areas are the Weddell Sea, Ross Sea, Wilkes-Ade´lie Coast, and some locations off Enderby Land and the Prydz Bay. Because the generation of AABW is characterized by highest densities (due to the freezing temperatures

0m

37.0

36.0

36.0

36.0

35.0

1000

34.4

AAIW 34.3

34.4

Rio Grande Plateau

2000 34.9 34.94 NADW

3000

34.9

4000

45°

40°

Vema Ch.

34.7

35°

30°W

Hunter Ch.

Santos Plateau

s

500

1000

M4

M5

M6

M7

3 _3

1500

1500

2000 km

2500

_9

0

_6

3000 3500

3

4500 0

Figure 14 Salinity distribution (PSU) on a section across the Rio Grande Rise, South Atlantic, at nominally 301S between the slope off Brazil and the Middle Atlantic Ridge. The Vema and Hunter Channels allow an active equatorward flux of Antarctic Bottom Water (AABW). More saline North Atlantic Deep Water (NADW) is advected poleward above the bottom water. On top of the abyssal layers lower saline Antarctic Intermediate Water (AAIW) again spreads equatorward. (Reproduced with permission of the American Meteorological Society from Hogg et al., 1999.)

_3

2000

4000 0

M3

500 1000

34.8

5000

M2

M1 0

34.9

34.8 AABW

6000

in the source regions), AABW can be detected by low (potential) temperature signals close to the bottom everywhere in the Southern Hemisphere and in lower latitudes of the Northern Hemisphere. Figure 9 demonstrates the global deep flow at 4000 m inferred from temperature distribution in the world ocean. The main production regions of sinking waters in the Weddell Sea (and near Iceland in the Northern Hemisphere) are labeled by arrows. The northern counterpart of AABW is called North Atlantic Deep Water (NADW). Since its potential density is slightly different than that of AABW both water masses create an abyssal stratification where they encounter. Formation areas of NADW are located in the Nordic seas, i.e. the Norwegian and Greenland Seas. Convectively formed deep water from polar regions spills through outflow channels over sills to the north west and south east of Iceland. The outflow products are called Denmark Strait and Iceland Scotland Overflow Waters. On their way into the subpolar North Atlantic, the swift overflow plumes with speeds exceeding 50 cm s1 are subject to strong topographic control along the slopes of Greenland and the Mid-Atlantic Ridge, respectively. The high propagation speed of overflow waters favors further transformations by entrainment from intermediate waters increasing its transport rate in selected regions by a factor of 2 or more (Figure 10).

Depth (m)

24

50 100 150 200 250 300 350 400 450 500 Distance (km)

Figure 15 Eulerian long-term observation of the southwardflowing North Atlantic Deep Water at nominally 181S. Isoclines of current speeds in cm s1. The prime deep western boundary current of Antarctic Bottom Water is pressed against the continental slope off Brazil. Dashed lines denote the conventional upper and lower boundaries of NADW. (Reproduced with permission of the American Meteorological Society from Weatherley et al. 2000.)


ABYSSAL CURRENTS

Additional mixing occurs with seasonally generated Labrador Sea Water. The latter also is formed by deep wintertime convection; its seasonal generation rate fluctuates substantially on interannual scales (Figure 11). Local formation regions are not restricted to the western and central Labrador Sea itself. The western Irminger Sea seems to contribute to NADW formation as well. Formation and sinking rates due to deep convection of AABW around Antarctica and of NADW in the North Atlantic are about equal in volume. Each hemisphere contributes B10 106 m3 s1. In the Southern Ocean generation sites are less focused compared with the Nordic Seas of the Atlantic. In the Pacific and Indian Oceans northern sources of

25

abyssal waters are unknown (or at least are insignificant in the Pacific). Overlying waters from two additional sources influence the abyssal circulation. (1) The excess evaporation over precipitation in the Mediterranean and Red Seas increases salinities and densities of these marginal seas substantially in comparison with the adjacent oceans. These density differences force exchange currents in the connecting straits (Strait of Gibraltar and Bab-el-Mandeb). The outflows sink to their equilibrium levels where they form pronounced intermediate high salinity layers in the North Atlantic (Figure 12) and northern Indic, respectively. (2) In selected regions of the southern polar frontal zone notable amounts of low salinity Antarctic Intermediate Water (AAIW) downwell and become

50

40

30

_100

_80

20 _60

2

_40

_20

2 10

0

4

1 _10

5_

20

_30

7 Figure 16 Spreading paths of Antarctic Bottom Water in the Atlantic after miscellaneous observations. Numbers indicate volume fluxes in 106 m3 s1. Originally the AABW spreading path is restricted to the western side of the South Atlantic. It enters the subtropical region through the Vema and Hunter Channels at 301S. The eastern basin is only accessible via the Vema Fracture Zone (111S) and through the Romanche and Chain Fracture Zones near the equator. (Reproduced with permission of the American Meteorological Society from Stephens and Marshall, 2000.)


26

ABYSSAL CURRENTS

part of the intermediate circulation. AAIW is found northward of subpolar regions in all three oceans. In the Pacific there is also a northern mode of intermediate water which is called North Pacific Intermediate Water (Figure 13).

60°N 55°N

50°N

Deep Western Boundary Currents (DWBC) In contrast to the slow interior drift in ocean basins DWBC are much easier to access by direct observations. They can be revealed by property characteristics of parameters like temperature, salinity, nutrients or other chemical tracers, or be quantified by moored instrumentation. Following the schematic circulation diagram (Figure 8) we expect three stacked opposite DWBC cores at intermediate and abyssal levels in the South Atlantic: northward propagating AABW with mixing components of Circumpolar Deep Water and of the Weddell Sea Deep Water, southward-flowing NADW at least reaching the southern rim of the subtropics, and AAIW with a northward drift again. Water mass distributions and their dynamical imprints are exemplarily displayed in form of a hydrographic section across the Rio Grande Rise at B301S. This ridge separates the abyssal Argentine from the Brazil Basins (Figure 14). Long-term current observations with moored instruments farther north at 181S are reproduced in Figure 15. The core of the southward flowing NADW with averaged speeds of 410 cm s1 is centered at 2200 m about 200 km offshore. The integrated transport of the NADW plume amounts to 39 106 m3 s1. The deepest northward AABW component along the continental rise with long-slope speeds of o3 cm s1 is less developed. The most recent transport estimates within the AABW from the Atlantic are shown in Figure 16. The northern part of this graph is shown in more detail in a schematic diagram jointly with lower NADW invading the Sargasso Sea and adjacent regions from the northern end of the Americas (Figure 17). Mixing of both abyssal water masses is symbolized by stars in Figure 17. The NADW drift grows from B13 10 m6 s1 south of the southern tip of Greenland (Cape Farewell) to B40 106 m6 s1 off the Caribbean island arc. In contrast to the Atlantic the basin-wide spreading of abyssal waters in the Pacific is exclusively controlled by the Antarctic Circumpolar Current System (ACCS). The lowest stratum is filled with Circumpolar Deep Water (CDW), a mixture of converted AABW and transformed NADW, possibly homogenized by repeat circulation in the Southern

45°N 40°N 35°N 30°N 25°N 20°N 15°N 10°N 5°N 0° 80°W

70°W

60°W

50°W

NADW

30°W

Figure 17 The complex abyssal circulation in the western North Atlantic. Colors: red and brown, flow of North Atlantic Deep Water; blue, Antarctic Bottom Water. Dashed lines indicate conceivable recirculation branches. Stars denote regions of abyssal entrainment. (Reproduced with permission from McCartney and Curry, 2001.)

Ocean. Potential temperatures of CDW at 321S range from 0.6 to 1.21C. It enters the South Pacific as a DWBC and returns to the ACCS as Pacific Deep Water (PDW) after internal mixing and upwelling decades or centuries later. The region north east of northern New Zealand delineates the gateway for CDW into the southeastPacific Basin (Figure 18). Its boundary current system was found from 2-year long observations to be B700 km wide. The maximum mean velocity of the applied current meter array was 9.6 cm s1 on the eastern flank of the Tonga-Kermadec Ridge. The time-averaged transport amounts to 15.8 106 m6 s1 with a standard error of 9.2 106 m6 s1 for the focused northward advection of the CDW core. A striking aspect of the overall observations at 321S are the total transport (CDW plus PDW) fluctuations ranging from ( 17 to þ 51) 106 m6 s1, typically oscillating over periods of a few months with amplitudes of 1–2 times the mean. About 201 farther north the northward flow of CDW into the main basins of the Pacific is topographically controlled by the Samoan Passage. Recent observations with moored current meters have


ABYSSAL CURRENTS

27

30˚S

2500 m 35˚S 30˚S

4000 m 35˚S 30˚S

_1

5 cm s

Bottom 35˚S 180˚

175˚W

170˚

Figure 18 Long-term observations of the deep western boundary current system at 321300 S at three instrumented abyssal layers (2500, 4000 m, near-bottom) of the Pacific. Mooring locations at dots are shown by the origin of the presented mean velocity vectors. Ellipses inform about the current’s stability in form of the root mean square amplitudes. Small italics in the lower panel indicate depths in km. Bolder numbers are mooring identifiers. Note the clear bottom-intensified flow of Antarctic Bottom Water. The returning North Pacific Deep Water flow is less confined and highly variable. (Reprinted from Progress in Oceanography 43, Whitworth et al., on the deep western-boundary current in the Southwest Pacific Basin, 1–54 ^ (1999) Elsevier Science and the American Meteorological Society.)

yielded northward transports of 10.671.7 106 m6 s1 from an 18 month measuring program. In accordance with the Stommel-Arons concept further decreases of DWBC transports were estimated at 101N (9.6 106 m6 s1) and 241N ((4.9 and 9.1) 106 m6 s1) from snapshot hydrographic surveys (Figure 19).

To date estimates of the returning PDW transports are rare and quantitatively inconsistent. The abyssal conditions of the Indic resemble those of the Pacific: no deep convective sources in the north are available (Figure 8). Instead, the deep Indian Ocean is controlled by deep and bottom waters from the ACCS (CDW) with a particular influence from


28

ABYSSAL CURRENTS 55°N

40°N

4.9 _ 9.1 20°N 9.6

0° 10.6 ± 9.1

20°S 15.8 ± 1.4

40°S

55°S 160°E

135°E

175°E

150°E

125°E

100°E

Figure 19 Inflow of Antarctic Bottom Water into the western basins of the subtropical/tropical Pacific. Heavy arrows were inferred from long-term current meter observations, open arrows represent single realizations from hydrography. Numbers indicate volumes of the deep western boundary current transports in 106 m3 s1. (After Hogg, 2001 ^ Academic Press.)

0°

20°

40°

60°

80°

100° 120°

140° 20°

20°

0°

0°

20°

20° 40°

40° 60°

60°

Figure 20 The distribution of potential density s4 in kg m3 referenced to 4000 dbar is used as a tracer for the spreading of Antarctic Bottom Water in the Indian Ocean. Note the three distinct tongues of dense water invading the Indian Ocean from the Southern Ocean. (Adapted from Mantyla AW and Reid JL, Journal of Geophysical Research 100, pp. 2417–2439 (1995) ^ by the American Geophysical Union and from Schmitz, 1996b.)

the southern South Atlantic (NADW). The Indonesian passages are too shallow to affect the Indic abyss immediately. The topography of the Indic is different to that of the other oceans. Several meridionally aligned ridges divide this ocean into sub-basins. The pathways of deep flows are therefore more complex then assumed in the simple approach by the Stommel-Arons concept. Figure 20 displays the distribution of water density, referenced to 4000 dbar at the bottom of the Indic. The graph enables a view of large-scale bottom water spreading on the base of hydrographic observations. Three major inlets for deep water (CDW or modified NADW) are obvious at the northern tips of the Mozambique (381E) and the Crozet (601E) Basins. The spreading into the Australian Basin (1251E) is constrained by the Australian–Antarctic Discordance at 501S, 1231E. The near-Equator Somali and Australian Basins are only accessible for deep flows through the Amirante and the Diamartina


ABYSSAL CURRENTS

5

5

0

0

_5

_5 92 93 94 95 96

_15

_15

_20

_20

_25

_25

_30

_30

_35 _50

_45

(A)

_40

92 93 94 95 96

_10

Latitude

Latitude

_10

_30 _35 Longitude

_25

_20

_35 _50

_15

29

_45

(B)

_40

_35

_30 Longitude

_25

_20

_15

Figure 21 Displacement vectors from 800 days of the drift of North Atlantic Deep (A) and Antarctic Bottom Waters (B) in the Brazil Basin. Nominal observation levels were 2500 m (A) and 4000 m (B). Note the clear signal of boundary currents and the predominant zonal structure of the flow in the central basins. Numbers þ 1900 indicate the years in which the Lagrangian current observations by neutrally buoyant floats were started. (From Hogg, 2001 ^ Academic Press.)

Passages, respectively. Long-term transport estimates of CDW spreading are rare. They lie clearly below the 10 106 m6 s1 range.

Conclusions During the past experimental period of the World Ocean Circulation Experiment (1990–99) significant progress in understanding of the abyssal circulation has been achieved. Deep Western Boundary Currents have been quantified on virtually all rises along continents and parallel to meridionally aligned ocean ridges. Pathways and transports of Deep Western Boundary Currents (DWBCs) were identified and quantified mostly with moored current meter arrays. Today the concept of a quasi-steady mean deep circulation has been outdated. In view of the omnipresent eddy kinetic energy with surprisingly large fluctuations more observations on a wide scale of temporal and spatial scales from the whole water column are essential. The need for long-term timeseries in ‘ocean observatories’ has been recognized in international programs like the Climate Variability and Predictability Programme (CLIVAR) or the Global Ocean Observation System (GOOS). The Stommel-Arons concept seems to be confirmed in respect to the balance of abyssal upwelling and downward heat flux through the base of the

main thermocline. However, although the first Lagrangian vector time-series from the abyss of the Brazil Basin are now available, according to the theory, the slow poleward return drift remains inaccessible to direct observations, as was found some 30 years ago in the deep Sargasso Sea. In contrast to expectations, net flow displacements in the Brazil Basin over 800 days indicate preferred zonal advection of bottom water partly in opposing directions (Figure 21). Besides current meter moorings improved instrumental approaches with autonomous arrays of density recorders (‘moored geostrophy’) or acoustic ocean tomography at a basin-scale lie at the frontier of new technologies for future integral observations. Such methods attenuate eddy noise before data are recorded, and hence reveal motion characteristics that may be closer to robust circulation patterns. Progress has also been made with respect to the experimental determination of abyssal diffusivities. Heat transports through deep passages appears to be significantly higher and steady compared with unconstrained areas. This implies enhanced mixing above passages converting them into important stirring agents of the abyss. Finally, recirculation cells delineate a significant limb between the swift DWBC regime and the interior slow drift regions. These deep powerful current loops on a sub-basin-wide scale may explain the


30

ABYSSAL CURRENTS

enormously high speed of the DWBC cores in lower latitudes, particularly in the Atlantic. Their variability strongly affects the meridional overturning cell and questions the role of the conveyor belt (Figure 2) in climate variability.

See also Antarctic Circumpolar Current. Atlantic Ocean Equatorial Currents. Benguela Current; Brazil and Falklands (Malvinas) Currents. California and Alaska Currents. Canary and Portugal Currents. Current Systems in the Atlantic Ocean. Current Systems in the Southern Ocean. Deep Convection. Dispersion and Diffusion in the Deep Ocean. Double-Diffusive Convection. East Australian Current. Florida Current, Gulf Stream and Labrador Current. Heat Transport and Climate. Kuroshio and Oyashio Currents. Non-Rotating Gravity Currents. Open Ocean Convection. Overflows and Cascades. Pacific Ocean Equatorial Currents. Rotating Gravity Currents. Water Types and Water Masses

Further Reading Broecker W (1991) The great ocean conveyor. Oceanography 4: 79--89. Dickson RR and Brown J (1994) The production of North Atlantic Deep Water: Sources, rates and pathwalks. Journal of Geophysical Research 99: 12319--12341. Friedrichs MAM and Hall MM (1993) Deep circulation in the tropical North Atlantic. Journal of Marine Research 51: 697--736. Hogg N (2001) Deep circulation. In: Siedler G, Church J, and Gould JW (eds.) Ocean Circulation and Climate. London: Academic Press. Hogg N, Siedler G, and Zenk W (1999) Circulation and variability at the southern boundary of the Brazil Basin. Journal of Physical Oceanography 29: 145--157. McCartney M and Curry R (2001) Abyssal potential vorticity in the western North Atlantic and the formation

of Lower North Atlantic Deep Water. Progress in Oceanography (in preparation). Mantyla AW and Reid JL (1995) On the origin of deep and bottom waters of the Indian Ocean. Journal of Geophysical Research 100: 2417--2439. Pedlosky J (1979) Geophysical Fluid Dynamics. New York: Springer. Polzin KL, Toole JM, Ledwell JR, and Schmitt RW (1997) Spatial variability of turbulent mixing in the abyssal ocean. Science 276: 93--96. Richardson PL, Bower AS, and Zenk W (2000) A census of Meddies tracked by floats. Progress in Oceanography 4: 209--250. Schmitz WJ Jr (1996a) On the World Ocean Circulation: Some Global Features/North Atlantic Circulation. 1, Woods Hole Oceanographic Institution, Technical Report, WHOI-96-03. Schmitz WJ Jr (1996b) On the World Ocean Circulation: The Pacific and Indian Oceans/A Global Update. 2, Woods Hole Oceanographic Institution, Technical Report, WHOI-96-08. Siedler G, Church J, and Gould JW (eds.) (2001) Ocean Circulation and Climate. London: Academic Press. Stephens JC and Marshall DP (2000) Dynamical pathways of Antarctic Bottom Water in the Atlantic. Journal of Physical Oceanography 30: 622--640. Stommel H (1958) The abyssal circulation. Deep-Sea Research 5: 80--82. Summerhayes CP and Thorpe SA (1996) Oceanography. An Illustrated Guide. London: Manson Publishing. Warren BA and Wunsch C (1981) Evolution in Physical Oceanography. Weatherly GL, Kim YY, and Kontar EA (2000) Eulerian measurements of the North Atlantic Deep Water Western Boundary Current at 181S. Journal of Physical Oceanography 30: 971--986. Wefer G, Berger WH, Siedler G, and Webb DJ (eds.) (1996) The South Atlantic: Present and Past Circulation. Berlin: Springer-Verlag. Whitworth T III, Warren BA, Nowlin WD Jr, et al. (1999) On the deep western-boundary current in the Southwest Pacific Basin. Progress in Oceanography 43: 1--54.


ACCRETIONARY PRISMS J. C. Moore, University of California at Santa Cruz, Santa Cruz, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 28–34, & 2001, Elsevier Ltd.

Introduction Subduction of oceanic lithosphere along a convergent plate boundary transfers sediments and rocks from the underthrust lithosphere to the overriding plate, producing an accretionary prism. Accretionary prisms develop beneath the inner slopes of the deep ocean trenches that typically mark convergent plate boundaries. The subduction process destabilizes the mantle after about 100 km of underthrusting beneath the upper plate to produce magmas of the volcanic arcs that virtually always occur along convergent plate boundaries (Figure 1). As accretionary prisms grow through addition of oceanic material, they become coastal mountain ranges. When a continent collides with a subduction zone, the intervening accretionary prism becomes incorporated into the resultant great mountain belts. Thus, rocks in accretionary prisms sometimes are the only record of ancient vanished ocean basins. Accretionary prisms typically form on the upper plate of subduction zone thrust faults, which host the world’s largest earthquakes. Because accretionary prisms incorporate soft sediments at high rates of deformation, they produce

Figure 1 The setting of an accretionary prism in a generalized cross-section of a convergent plate boundary. Although the accretionary prism builds up primarily by scraping off material riding on the oceanic crust, some portions are deeply underthrust and flow back to the surface, while other portions are deeply subducted and participate in the formation of the igneous rocks of the volcanic arc.

some of the world’s most complexly deformed rocks, commonly called melanges. Sediments offscraped to form accretionary prisms are like sponges that yield fluids as they are squeezed and deformed during prism growth. The fluids affect the mechanics of faults; chemically dissolve, transfer, and deposit material; and support chemosynthetic biological communities. The shape of the accretionary prism is mechanically controlled by the strength of the material comprising the accretionary prism and its internal fluid pressure. At slightly fewer than half of modern convergent plate boundaries, accretionary prisms are not currently forming and the incoming sediment and rock is deeply underthrust. Sometimes the accretionary process is reversed and the upper plates of subduction zones are mechanically abraded or eroded by the underthrust plate, causing subsidence and contraction of the overriding plate.

Origin and Variation of Materials Incorporated in Accretionary Prisms Accretionary prisms vary in composition depending on the type of material on the subducting oceanic plate. The ideal sequence incoming to a subduction zone consists of oceanic basaltic igneous rocks covered by oceanic or pelagic sediments that change upsection to more rapidly deposited, continentally derived sandstones, shales, and even conglomerates. At a subduction zone starved for sediments, material available for accretionary prism construction may be the igneous rocks of the oceanic plate with thin overlying sedimentary deposits. Alternatively, incoming plates may be sediment-dominated and covered with a kilometers-thick sequence of deposits that are available for accretion. The resulting accretionary prisms may consist of slices of oceanic igneous rocks with minor amounts of interspersed sediments to thick thrust sheets of continentally derived clastic rocks. The Marianas subduction zone of the western Pacific is an example of the former, whereas the Cascadia subduction zone off the northwestern United States and Canada is an example of the latter. Because the sediment-dominated accretionary prisms (Figure 2) are more voluminous, they tend to be well recognized in the stratigraphic record, for example, parts of the Franciscan Complex of California. Sediment-starved accretionary prisms are thinner and typically dominated by basaltic and ultramafic igneous rocks. They are harder


31

32

ACCRETIONARY PRISMS

Figure 2 Cross-section of an accretionary prism showing the principal structural elements. Paths of fluid flow from deep sources to surface utilize high-permeability conduits, whether along sedimentary layers or faults.

to recognize in the ancient stratigraphic record. Factors controlling the amount of sediment and type of sediment available for accretion include the age of the oceanic plate, the types and rates of sediments being deposited along the transport path of the oceanic plate to the subduction zone, and rate of travel or residence time of any plate in a particular sedimentary environment.

Solid Material Transfer in Accretionary Prisms Sediments and rocks incoming to a subduction zone may be (1) offscraped as a series of thrust sheets at the frontal edge of the accretionary prism; (2) underplated or emplaced at depth along the base of the upper lithospheric plate; or (3) underthrust to great depths to participate in production of volcanic arc magmas or ultimately to be carried into the earth’s mantle (Figures 1 and 2). In the zone of offscraping, incoming materials are typically accreted as a series of imbricate thrust sheets extending from the surface to a basal detachment fault or decollement, beneath which all other material is underthrust. The decollement forms in a layer weaker than the adjacent sediment. With continued underthrusting, the weak layer may become stronger because of mineralogical changes or decreasing fluid pressure and step or migrate down through the underthrust plate. This down-stepping process underplates fault-bounded rock packages to the overlying plate (Figure 2).

The surfaces of some accretionary prisms and nonaccretionary convergent margins (see below) are marked by volcanoes of fluidized mud or serpentine (Figure 2). The fluidized mud and serpentine rise through the upper plate of the convergent margin because these materials are of lower density than the surrounding sediments and rocks. Deep underthrusting of mud and the associated production of natural gas or oil produces mud volcanoes. Serpentine, a low-density rock, is formed by addition of water from underthrust sediments or rocks to mantle rocks. In both cases the low-density rock occurs beneath higher-density rock and buoyancy forces drive the low-density material to the surface. Incoming sediment that is not offscraped at the front of or underplated beneath the main part of the accretionary prism continues to be underthrust. This underthrust sediment may be underplated beneath volcanic arc basement rocks at any point until it reaches the melting zone (Figure 1). Short-lived radioactive isotopes and other chemical tracers in volcanic rocks indicate that sediments less than several million years old are underthrust to the depths of melting (75–100 km) beneath the volcanic arc. Residual material from the melting zone may be deeply underthrust into the earth’s mantle. Continuing sediment accumulation occurs on top of the growing accretionary prism, forming an apron of slope deposits. Locally erosion may remove material from the accretionary prism, redepositing it on the oceanic plate for recycling back into the accretionary prism. In addition to the thrust faults and the


ACCRETIONARY PRISMS

decollement associated with accretionary processes, accretionary prisms are cut by thrust, normal, and strike-slip faults that form as the prism adjusts its shape in response to its continuing growth (see Mechanics below). Accretionary prisms include some of the most complexly deformed and puzzling rocks on the earth. Melanges or ‘mixed rocks’ and stratally disrupted rocks are included in this category. These rocks are marked by not only stratal discontinuity (Figure 3A) but by mixing of incompatible sedimentary and metamorphic environments (Figure 3B). These intricately deformed rocks form from hard igneous rocks and sediments that are partially consolidated and lithified in a high-strain and high-strain-rate environment. The extreme variation in strength between the hard rocks and soft sediments and the high

(A)

33

strain deformation results in a heterogeneously deformed rock mass. Various return flow processes at depth (Figure 1), faulting, and erosion and redeposition of previously accreted rocks contribute to the mixing of rocks derived from differing sedimentary and metamorphic environments (Figure 3B). Accretionary prisms include metamorphic rocks formed under high-pressure, low-temperature conditions. These rocks, called blueschists (Figure 4), are diagnostic of this subduction zone metamorphic environment. The high-pressure–low-temperature condition is caused by the rapid underthrusting of old, cold oceanic plates. Burial rates for underthrust rocks at subduction zones exceed 20 km My 1. The material is buried and returned to shallow depths more rapidly than it can be warmed by conduction from adjacent warmer parts of the earth. Thus, the low-temperature–high-pressure conditions of the subduction system are imprinted on the rocks and preserved by rapid uplift. The lowest average geothermal gradients through accretionary prisms are less than 101C km 1, or about a third of the typical gradient through continents. Thermal gradients beneath accretionary prisms may be much higher where young, hot oceanic crust is being subducted. In these cases blueschists are not formed but are replaced by metamorphic rocks characteristic of higher temperature regimes (greenschist and amphibolite). Overall, the material transfer in accretionary prisms is similar to that in thrust belts in overall form, fault geometry, and mechanics. Thrust belts on land tend to deform more consolidated and lithified sedimentary rocks at slower rates than occur at oceanic convergence zones. Therefore, in thrust belts, both the structural complexity and the effects

(B) Figure 3 Melanges. (A) Dismembered layers of light-colored sandstone in shale matrix. This type of deformation is typically developed along thrust faults with substantial displacement or along the decollement. (B) Melange consisting of block of basalt with overlying pelagic or open oceanic limestone included in shale matrix. Limestone and shale are incompatible depositional environments that are mixed together in the deformational environment of the accretionary prism. Both photographs are from the 60–65 My old accretionary prism of the Kodiak Islands, Alaska.

Figure 4 Blueschist metamorphic rock from an accretionary prism, indicating high pressures and low temperatures and showing ductile or plastic deformation (fold). This rock was metamorphosed about 200 My ago and occurs in the accretionary prism of the Kodiak Islands of Alaska.


34

ACCRETIONARY PRISMS

due to fluid expulsion are less pronounced than in accretionary prisms. A good example of a thrust belt is the Canadian Rocky Mountains just west of Calgary Alberta.

Fluids In and Out of Accretionary Prisms Accretionary prisms are like sponges, saturated as they begin to form but squeezed virtually dry as their rocks reach maximum depths of burial. About 40% of the sediment section entering the world’s subduction zones is composed of water in pore spaces. Additional water resides in pores in the igneous rocks of the crust and is bound to minerals both in the sediments and the oceanic crust. The rapid burial of incoming sediments and rocks due to incorporation into and underthrusting beneath the accretionary prism increases stress on the rock framework and raises fluid pressure (the sponge is squeezed). Burial also increases temperature, which releases water bound in minerals and converts sedimentary organic matter to oil and natural gas. Sediment microrganisms may also produce natural gas from organic matter. By 1501C, the minerals are substantially dehydrated and most of the organic matter has been converted to hydrocarbons. By about 4 km of burial, the pore volume of sedimentary rocks in the accretionary prism is reduced by about 90%. The high fluid pressure drives fluids out of the accretionary prism through sediment pores and fractures and along faults (Figure 2). Fluid migration out of the accretionary prism affects everything from surface biology, to large-scale structural features, to the fabric of prism rocks. The high fluid pressures that result from the fluid generation process weaken the rocks by reducing the contact stresses on mineral grains and fault surfaces, thereby reducing friction. This pressure-related weakening facilitates long-distance lateral transport on thrust faults (Figure 2), such as the decollement, and also controls the shape of the accretionary prism (see Mechanics). At temperatures from 1001C and up, rocks dissolve and re-precipitate (undergo pressure solution) with formation of preferred orientations of minerals and solution seams (fabrics) that record stress orientations. Rocks such as slates commonly form through this process. The constituents of the dissolved carbonate and silicate minerals may be locally precipitated or transported with the fluids and precipitated elsewhere as veins and cements that are common in prism rocks. The fluidmediated processes of dissolution, transport, and precipitation significantly change the physical

properties of the accretionary prism and ultimately affect how it deforms. Fluids expulsed from the surface of accretionary prisms contain dissolved methane and hydrogen sulfide, which are utilized by chemosynthetic organisms that are the basis for cold seep biological communities on the seafloor. In the sub-surface, microorganisms both produce and consume various fluid-borne chemical constituents, modifying the physical properties and composition of the host sediments or rocks. Because fluids alter the physical properties of sediment and rock, they affect the seismic reflection images of the prism interior. Fluid-enriched zones along faults reduce velocity and density and produce strong reflections in the seismic images. Methane near the surface of the accretionary prism freezes to form methane hydrate. Progressively deeper in the prism, the methane hydrate is unstable and is transformed back to free methane. Minor accumulations of free methane gas below the hydrate produce a large change in rock physical properties that is seen as a prominent bottomsimulating reflection in seismic images.

Seismogenesis and Accretionary Prisms Subduction zone thrust faults produce the largest earthquakes on the Earth because the plate-boundary thrust is in a zone of high strain rate and is inclined shallowly. The shallow inclination provides a large surface area, or seismogenic zone, subject to brittle failure and the production of earthquakes (Figure 1). In contrast, more steeply inclined faults, such as the San Andreas Fault, transition down-dip from the region of brittle or seismogenic deformation into the realm of ductile deformation over a shorter distance, therefore limiting the area capable of producing an earthquake. The low thermal gradients characteristic of many subduction zones also extends the brittle–ductile transition to greater depths than in other plate boundary settings and increases the area subject to catastrophic seismogenic failure. Commonly, the seismogenic zone earthquakes occur beneath accretionary prisms, such as the great Alaskan earthquake of 1964 (magnitude 9.2). The ability of accretionary prisms to build up enough elastic strain to be released in a sudden earthquake event testifies to strength developed during the evolution from soft sediments to hard rocks. In addition to becoming strong and rigid, the materials of the accretionary prism must evolve to fail in a ‘velocity weakening’ manner, such that there is an acceleration of slip along the fault surface. This accelerating slip produces a discrete seismic event, as opposed to a


ACCRETIONARY PRISMS

decelerating creep event that would not produce an earthquake.

Mechanics In 1983 Davis, Dahlen, and Suppe articulated and formalized the mechanics of accretionary prisms in the widely accepted ‘critical Coulomb wedge theory’ (Figure 5). Virtually all accretionary prisms approximate a wedge in shape, being thinner on the oceanic side and thicker toward the associated volcanic arc. According to Davis et al. accretionary wedges resemble piles of dirt (prism materials) being pushed forward by bulldozer (the volcanic arc basement). The stresses driving the prism seaward are that of the arc basement pushing the wedge from the rear and a seaward-directed lateral stress due to unequal gravitational stress resulting from the wedge shape. The latter is similar to the stress causing a pile of sand to fail if it is oversteepened. These driving stresses are resisted by a shear stress along the base of the wedge controlled by the frictional strength of the material, which varies with overburden stress, the fluid pressure along the base, and the material properties. Additionally, the component of the overburden stress acting parallel to the decollement must be overcome to, in essence, lift the prism up the decollement. The wedge is just as thick as it can be in order to move forward; that is, it is at its threshold of failure determined by its frictional strength and internal fluid pressure. If the wedge grows at its leading edge, the area of the decollement resisting motion increases and the prism must thicken at the rear in order to increase the area to which the driving stresses in the arc basement can be applied. Wedge theory has been largely successful in explaining the shape of accretionary prisms and the observations of fluid pressure, though the latter are not numerous. Because fluid pressure can counteract normal stresses along the decollement, it is a prime

35

parameter controlling the mechanics of accretionary prisms. The generally high fluid pressure along the decollement sharply decreases the frictional resistance there, allowing narrowly tapered prisms to be mechanically stable. The necessity to thicken the landward portion of the prism to keep a stable wedge taper can explain much of the faulting observed in the landward parts of prisms (Figure 2). Other mechanical conceptualizations of wedges utilize differing material properties; however, the Coulomb wedge theory, dependent on basic frictional behavior, is most successful at depths up to 10–20 km.

Non-accretionary Convergent Plate Boundaries According to a compilation by Von Huene and Scholl, somewhat more than half of the world’s convergent plate boundaries are forming accretionary prisms now. The remainder of convergent plate boundaries have inner trench slopes underlain by older accretionary prisms or igneous and metamorphic rocks of the continental crustal or volcanic arc origin (Figure 6). Some are underlain by igneous and metamorphic rocks of uncertain origin that may represent accreted pieces of seamounts or normal oceanic crust. Recent Ocean Drilling Program results off Costa Rica unequivocally demonstrate that virtually all of the sediment riding on the oceanic plate (Cocos Plate) is underthrust beneath the upper plate (Caribbean Plate) of the subduction zone. This process must also occur at many other convergent margins without accretionary prisms. At a number of other convergent margins, the presence of continental or volcanic arc rocks close to the trench suggests that portions of the forearc may have been tectonically eroded by underthrusting. Moreover, Ocean Drilling Program holes show that rocks currently located in deep water on inner slopes of trenches have subsided from much shallower depths.

Figure 5 Diagram showing stresses that control motion of accretionary prisms (and thrust belts). These stresses, integrated over the areas where they act, form a force balance. Forces driving the prism seaward (a push from the rear and internal gravitational forces) are offset by resisting forces (primarily frictional forces and gravitational forces acting along the base of the decollement).


36

ACCRETIONARY PRISMS

Figure 6 A convergent margin showing a range of features seen worldwide supporting nonaccretion or tectonic erosion. The absence of offscraping indicates no accretion at the front of the margin. At deeper levels, sediment may be being underplated. Locally, sediments and contained fluids are underthrust to great depths and participate in the melting process beneath the volcanic arc (see Figure 1). A previous accretionary prism much older than the age of incoming sediment may be being underthrust. Arc basement may be anomalously close to the trench, suggesting erosion of the upper plate. Slope deposits may be of much shallower water origin. These shallow-water deposits suggest substantial subsidence of the margin, which is commonly explained by tectonic erosion of the upper plate.

Presumably this subsidence is due to tectonic erosion of the trench inner slope by the underthrusting plate. So, although there are accretionary prisms forming at many convergent margins, the offscraping and underplating at the front of the margin is not a constant process. Accretionary prism formation may be episodic even with continuous subduction. Tectonic erosion may occur. Gaps in the record of accretion are to be expected. What determines whether convergent plate boundaries form accretionary prisms or tectonically erode the upper plate? The primary control is sediment supply. Where thick sequences of sediment enter the subduction zones, accretionary prisms form. Where the sedimentary sequences are thin, there is less material available to accrete and irregularities in the underthrusting oceanic plate may interact with the overthrusting plate to cause erosion. Protruding fault blocks or seamounts may tectonically abrade the lower surface of the upper plate. Also, as these high areas on the lower plate are underthrust, they may oversteepen and otherwise disturb the upper plate, causing it to fail at the surface by landsliding. The landslides accumulate in the trench, where they are underthrust. Accordingly, the front of the convergent margin is decimated and cycled to deeper levels in the earth.

Conclusions Accretionary prisms form at the leading edge of convergent plate boundaries by skimming-off sediments and rocks of the lower plate. In detail, the accretion process involves offscraping of rocks and

sediments at the front of the prism or underplating (emplacement beneath the prism). This deformational process stacks the sediments into thick vertical piles and shortens them horizontally. Consequently, fluids are expulsed and the sediments are progressively transformed to rocks. Because the deformation in accretionary prisms is large and fast, rocks emplaced therein are often severely disrupted and mixed, forming melanges. The rapid rate of underthrusting of the lower plate may carry rocks to great depths before they can heat up, forming a characteristic type of metamorphic rock called a blueschist. The fault surface bounding the base of the accretionary prism, the decollement, is the plate boundary thrust. Because it is shallowly inclined, this fault has a large area undergoing brittle deformation and produces the largest earthquakes on the planet. Mechanically, accretionary prisms resemble a pile of dirt being pushed by a bulldozer. Prisms are pushed from the rear by the volcanic arc basement, with forward motion being resisted by frictional and gravitational forces. Accretion does not occur at the leading edge of all convergent plate boundaries. In these cases, sediments and rocks are underthrust beneath the crustal framework of the upper plate. In some cases accreted rocks have been removed from the upper plate, apparently by tectonic erosion by the lower plate. Thus, an accretionary prism may be a very discontinuous recorder of the incoming sediments and rocks at a convergent plate boundary. Sediments, rocks, and fluids not emplaced in the accretionary prism are carried to great depths and either catalyze subduction zone volcanism or are transported into the mantle.


ACCRETIONARY PRISMS

Acknowledgments I thank the National Science Foundation for grants supporting my studies of accretionary prisms since 1974 (most recently grant # OCE 9802264). The Ocean Drilling Program provided the opportunity and support to participate in many cruises investigating accretionary prisms. Eli Silver’s insight into convergent margin tectonics substantially improved this review. Hilde Schwartz thoughtfully reviewed the final manuscript.

See also Mid-Ocean Ridge Geochemistry and Petrology. Mid-Ocean Ridge Seismic Structure. Mid-Ocean Ridge Tectonics, Volcanism and Geomorphology. Seismic Structure.

Further Reading Bebout GE, Scholl DW, Kirby SH, and Platt JP (1996) Subduction Top to Bottom, American Geophysical Union Monograph 96. Washington, DC: American Geophysical Union. Davis DJ, Suppe J, and Dahlen FA (1983) Mechanics of fold-and-thrust belts and accretionary wedges. Journal of Geophysical Research 88: 1153--1172. Fisher DM (1996) Fabrics and veins in the forearc: a record of cyclic fluid flow at depths of o15 km. In: Bebout GE, Scholl DW, Kirby SH, and Platt JP (eds.) Subduction Top to Bottom, American Geophysical Union Monograph 96, pp. 75--89. Washington, DC: American Geophysical Union. Fryer P, Mottl M, Johnson LE, et al. (1995) Serpentine bodies in the forearcs of Western Pacific convergent margins. In: Taylor B and Natland J (eds.) Active Margins and Marginal Basins of the Western Pacific,

37

American Geophysical Union Monograph 88, pp. 259--279. Washington, DC: American Geophysical Union. Hyndman RD (1997) Seismogenic zone of subduction thrust faults. The Island Arc 6: 244--260. Kastner M, Elderfield H, and Martin JB (1991) Fluids in convergent margins: what do we know about their composition, origin, role in diagenesis and importance for oceanic chemical fluxes. Philosphical Transactions of the Royal Society of London, Series A 335: 243--259. Meschede M, Zweigel P, and Kiefer E (1999) Subsidence and extension at a convergent plate margin: evidence for subduction erosion off Costa Rica. Terra Nova 11: 112--117. Moore JC and Vrolijk P (1992) Fluids in accretionary prisms. Reviews in Geophysics 30: 113--135. Moores EM and Twiss RJ (1995) Tectonics. New York: WH Freeman. Morris JD, Leeman WP, and Tera F (1990) The subducted component in island arc lavas; constraints from B-Be isotopes and Be systematics. Nature 344: 31--36. Silver EA (2000) Leg 170 Synthesis of fluid-structural relationships of the Pacific Margin of Costa Rica. In: Silver EA, Kimura G, Blum P and Shipley TH (eds) Proceedings of the Ocean Drilling Program, Scientific Results 170 [Online at http://www.odp.tamu.edu/ publications/170_SR/VOLUME/CHAPTERS/ SR170_04.PDF] Taira A, Byrne T, and Ashi J (1992) Photographic Atlas of an Accretionary: Geologic Structures of the Shimanto Belt. Japan, Tokyo: University of Tokyo Press. Tarney J, Pickering KT, Knipe RJ, and Dewey JF (1991) The Behaviour and Influence of Fluid in Subduction Zones. Philosophical Transactions of the The Royal Society of London, Series AE 335: 225--418. von Huene R and Scholl DW (1991) Observations at convergent margins concerning sediment subduction, subduction erosion, and the growth of continental crust. Reviews in Geophysics 29: 279--316.


ACOUSTIC MEASUREMENT OF NEAR-BED SEDIMENT TRANSPORT PROCESSES P. D. Thorne and P. S. Bell, Proudman Oceanographic Laboratory, Liverpool, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction: Sediments and Why Sound Is Used Marine sediment systems are complex, frequently comprising mixtures of different particles with noncohesive (sands) and cohesive (clays and muds) properties. The movement of sediments in coastal waters impacts on many marine processes. Through the actions of accretion, erosion, and transport, sediments define most of our coastline. Their deposition and resuspension by waves and tidal currents in estuarine and nearshore environments control seabed morphology. Fine sediments, which act as reservoirs for nutrients and contaminants and as regulators of light transmission through the water column, have significant impact on water chemistry and on primary production. Therefore, an improved understanding of sediment dynamics in coastal waters has relevance to a broad spectrum of marine science ranging from physical and chemical processes, to the complex biological and ecological structures supported by sedimentary environments. However, it is commonly acknowledged that our capability to describe the coupled system of the bed, the hydrodynamics, and the sediments themselves is still relatively primitive and often based on empiricism. Recent advances in observational technologies now allow sediment processes to be investigated with greater detail and precision than has previously been the case. The combined use of acoustics, laser and radar, both at large and small scales, is facilitating exciting measurement opportunities. The enhancement of computing capabilities also allows us to make use of more complex coupled sediment–hydrodynamic models, which, linked with the emerging observations, provide new openings for model development. It is readily acknowledged by sedimentologists that the presently available commercial instrumentation does not satisfy a number of requirements for nearbed sediment transport processes studies. Here we focus on the development of acoustics to fulfill some of these needs. The question could be asked: Why use acoustics for such studies? Sediment transport can be

38

thought of as dynamic interactions between: (1) the seabed morphology, (2) the sediment field, and (3) the hydrodynamics. These three components interrelate with each other in complex ways, being mutually interactive and interdependent as illustrated schematically as a triad in Figure 1. The objective therefore is to measure this interacting triad with sufficient resolution to study the dominate mechanisms. Acoustics uniquely offers the prospect of being able to nonintrusively provide profiles of the flow, the suspension field, and the bed topography. This exclusive combination of being able to measure all three components of the sediment triad, co-located and simultaneously, has been and is the driving force for applying acoustics to sediment transport processes. The idea of using sound to study fundamental sediment processes in the marine environment is attractive, and, in concept, straightforward. A pulse of high-frequency sound, typically in the range 0.5–5 MHz in frequency, and centimetric in length, is transmitted downward from a directional sound source usually mounted a meter or two above the bed. As the pulse propagates down toward the bed, sediment in suspension backscatters a proportion of the sound and the bed generally returns a strong echo. The amplitude of the signal backscattered from the suspended sediments can be used to obtain vertical profiles of the suspended concentration and particle size. Utilizing the rate of change of phase of the backscattered signal provides profiles of the three orthogonal components of flow. The strong echo from the bed can be used to measure the bed forms. Sediments Bed load−suspended load

Sediment triad

Flow Waves−current

Bed forms Flat−rippled

Figure 1 Illustration of the sediment processes triad and their interactions.


ACOUSTIC MEASUREMENT OF NEAR-BED SEDIMENT TRANSPORT PROCESSES

Acoustics therefore has the potentiality to provide profile measurements of near-bed sediment processes, with sufficient spatial and temporal resolution to allow turbulence and intrawave processes to be probed; this coupled with the bedform morphology observations provides sedimentologists and coastal engineers with an extremely powerful tool to advance understanding of sediment entrainment and transport. All of this is delivered with almost no influence on the processes being measured, because sound is the instrument of measurement.

Some Historical Background For over two decades the vision of a number of people involved in studying small-scale sediment processes in the coastal zone has been to attempt to utilize the potential of acoustics to simultaneously and nonintrusively measure seabed morphology, suspended sediment particle size and concentration profiles, and profiles of the three components of flow, with the required resolution to observe the perceived dominant sediment processes. The capabilities to measure the three components have developed at different rates, and it is only in the past few years that the potentiality of an integrated acoustic approach for measuring the triad has become realizable. A schematic of the vision is shown in Figure 2. The figure shows a visualization of the application of acoustics to sediment transport processes. ‘A’ is a

A

B

C

39

multifrequency acoustic backscatter system (ABS), consisting, in this case, of three downward-looking narrow beam transceivers. The differential scattering characteristic of the suspended particles with frequency is used to obtain profiles of suspended sediment particle size and concentration profiles. ‘B’ is a three-axis coherent Doppler velocity profiler (CDVP) for measuring co-located profiles of the three orthogonal components of flow velocity; two horizontal and one vertical. It consists of an active narrow beam transceiver pointing vertically downward and two passive receivers having a wide beam width in the vertical and a narrow beam width in the horizontal. This system uses the rate of change of phase from the backscattered signal to obtain the three velocity components. ‘C’ is a pencil beam transceiver which rotates about a horizontal axis and functions as an acoustic ripple profiler (ARP). This is used to extract the bed echo and provide profiles of the bed morphology along a transect. These measurements are used to obtain, for example, ripple height and wavelength, and assess bed roughness. ‘D’ is a highresolution acoustic bed (ripple) sector scanner (ARS) for imaging the local bed features. Although the ARS does not provide quantitative measurements of bed form height, it does provide the spatial distribution and this can be very useful when used in conjunction with the ARP. ‘E’ is a rapid backscatter ripple profiler system (BSARP) for measuring the instantaneous relationship between bed forms and the suspended sediments above it.

D

E

B

Waves

B Current Suspension

Bed forms

Figure 2 A vision of the application of acoustics to sediment transport processes. A, multifrequency acoustic backscatter for measuring suspended sediment particle size and concentration profiles. B, coherent Doppler velocity profiler for measuring the three orthogonal components of flow velocity. C, bed ripple profiler for measuring the bed morphology along a transect. D, high-resolution sector scanner for imaging the local bed features. E, backscatter scanning system for measuring the relationship between bed form morphology and suspended sediments.


40


Acoustic ripple scanners, ARSs, are based on sector-scanning technology, which has been specifically adapted for high-resolution images of bed form morphology. They typically have a frequency of around 1–2 MHz with beam widths of about 11 in the horizontal and 301 in the vertical. As the pulse is backscattered from the bed, the envelope of the signal is measured and usually displayed as image intensity. An example of the data collected by an ARS is shown in Figure 4. As can be seen, this provides an aerial image of the bed, clearly showing the main bed features. The advantage of the ARS is the area coverage that is obtained, as opposed to a single line profile with the ARP; however, direct information on the height of features within the image cannot readily be extracted. Ideally one would like to combine the two instruments and recently such systems have become available. This is essentially an ARP which also rotates horizontally through 1801 and therefore allows a three-dimensional (3-D) measurements of the bed. The sonar gathers a single swath of data in the vertical plane and then rotates the transducer around the vertical axis and repeats the process until a circular area underneath the sonar has been scanned in a sequence of radial spokes. An example of data collected by a 3-D ARP operating at a frequency of 1.1 MHz is shown in Figure 5.

What Can Be Measured? The Bed

Whether the bed is rippled or flat has a profound influence on the mechanism of sediment entrainment into the water column. Steep ripples are associated with vortices lifting sediment well away from the bed, while for flat beds sediments primarily move in a confined thin layer within several grain diameters of the bed. Therefore knowing the form of the bed is a central component in understanding sediment transport processes. The development of the ARP and the ARS has had a significant impact on how we interpret sediment transport observations. These specifically designed systems typically either provide quantitative measurements of the evolution of a bed profile with time, the ARP, or generate an image of the local bed features over an area, the ARS. Figure 3 shows data collected with a 2-MHz narrow pencil beam ARP in a marine setting. The figure shows the variability of a bed form profile, over nominally a 3-m transect, covering a 24-h period. Over this period the bed was subject to both tidal currents and waves and the figure shows the complex evolution of the bed with periods of ripples and less regular bed forms. The figure clearly shows the detailed quantitative measurements of the bed that can be obtained with the ARP.

Height (m)

0.2

0

−0.2 1.5 1

25 0.5 Di

20

sta

0

nc

e(

m)

15 −0.5

10 −1

5 −1.5

)

e (h

Tim

0

Figure 3 ARP measurements of the temporal evolution of a 3-m profile on a rippled bed, covering a 24-h period.


41


2

y (m)

1

0

−1

−2

−2

−1

0

1

2

x (m)

z (m)

Figure 4 Image of ripples on the bed in a large-scale flume facility, the Delta Flume, collected using an acoustic ripple sector scanner, ARS. The 0.5-m-diameter circle at the center right of the image is one of the feet of the instrumented tripod used to collect the measurements, while the dark area at the center of the image is a blind spot directly beneath the sonar.

−0.25 −0.5 −2

−1

0 )

m

x(

2 1

1 0 2

−1

y (m)

−2

Figure 5 Three-dimensional measurements of a rippled bed collected in the marine environment using a 3-D ripple scanner, 3-D ARP. The artifacts in the image are associated with reflections from the instrumented tripod used to collect the measurements.

The data shown in the figure were collected during a recent marine deployment. The bed surface shown is based on a 3-D scan, with a 100 vertical swaths at 1.81 intervals, each of which comprised 200 acoustic

samples at 0.91 intervals and spanning 1801. The plot clearly shows a rippled bed. There are one or two artifacts in the plot; these are associated with reflection from the rig on which the 3-D ARP was


42


mounted. However, the figure plainly contains information on both the horizontal and vertical dimensions of the ripples and can therefore be used to precisely define quantitatively the features of the bed over an area. The 3-D ARP is a substantial advance on both the ARP and the ARS.

The Flow

The success of acoustic Doppler current profilers (ADCPs), which typically provide mean current profiles with decimeter spatial resolution, and more recently the acoustic Doppler velocimeter (ADV), which measures, subcentimetric, subsecond, three velocity components at a single height, has stimulated interest in using acoustics to measure near-bed velocity profiles. The objective is to use the same backscattered signal as used by the ABSs, but process the rate of change of phase of the signal (rather than the amplitude as used by ABSs) to obtain velocity profiles with comparable spatial and temporal resolution to ABSs. The phase technique is utilized in CDVPs and the phase approach has been the preferred method for obtaining high spatial and temporal resolution velocity profiles. An illustration of a three-axis CDVP is shown in Figure 6. A narrowbeam, downwardly pointing transceiver, Tz, transmits a pulse of sound. The scattered signal is picked up by Tz, and two passive receivers Rx and Ry which are orthogonal to each other and have a wide beam in the vertical and narrow in the horizontal. Receiver Rx

Receiver Ry z Transmit transceiver

Tz

To examine the capability of such a system, measurements from a three-axis CDVP have been compared with a commercially available ADV. The system had a spatial resolution of 0.04 m, operated over a range of 1.28 m, and provided 16-Hz velocity measurements of the vertical and two horizontal components of the flow. Figure 7 shows a typical example of CDVP and ADV time series, power spectral density, and probability density function plots for u, the streamwise flow. The velocities presented in Figure 7 show CDVP results that compare very favorably with the ADV measurements; having time series, spectra, and probability distributions in general agreement. There are differences in the spectrum; the CDVP spectra begins to depart from the ADV above about 4 Hz, with the CDVP measuring larger spectral components at the higher frequencies. This trend was common to all the records and is a limitation of the CDVP system used to obtain the data, rather than an intrinsic limitation to the technique. Figure 8 illustrates the capability of the CDVP for flow visualization in the marine environment. The figure shows mean zeroed velocity vectors, u–w, v–w, and u–v, plotted over a 5-s time period, between 0.05 and 0.7 m above the bed. The length of the velocity vectors is indicated in the figure. A single-point measurement instrument such as an ADV can provide the time-varying velocity vectors at a single height above the bed; however, the spatial profiling which is achievable with the threeaxis CDVP provides a capability to visualize structures in the flow. The structures seen in Figure 8 are associated with combined turbulent and wave flows. This type of plot exemplifies the value of developing a three-axis CDVP with co-located measurement volumes, since it clearly illustrates the fine-scale temporal and spatial flow structures which can be measured in the near-bed flow regime. Linking such measurements with ABS profiles of particle size and concentration will provide a very powerful tool for studying near-bed fluxes and sediment transport processes.

y

The Suspended Sediments

Range bin

x

Figure 6 Schematic of the transducer arrangement for a threeaxis CDVP.

Multifrequency acoustic backscattering (ABS) can be used to obtain profiles of mean particle size and concentration. The ABS is the only system available that profiles both parameters rapidly and simultaneously. Also the bed echo references the profile to the local bed position. This is important because all sediment transport formulas use the bed as the reference point and predict profiles of suspension parameters relative to the bed location. Examples of the results that can be obtained are shown



u (m s−1)

(a)

43

1.5 1 0.5 0 −0.5 0

Time (s)

(b)

100

(c) 3 2.5 2 10−2

PDFu

PSDu (m2 s−2 Hz−1)

10−1

10−3

1.5 1 0.5

10−4 10−1

100

0 −1

101

−0.5

0

0.5

1

Velocity (m s−1)

Frequency (Hz)

Figure 7 Comparison of the streamwise flow, u, measured by an ADV (red) and a CDVP (black). (a) The velocities measured at 16 Hz over 100 s. (b) The power spectra of the zero-mean velocities. (c) The probability density functions of the zero-mean velocities for the ADV (open circles) and the CDVP (crosses).

Height above bed (m)

(a) 0.8 0.6 0.4 0.2 0 155

155.5

156

156.5

157

155.5

156

156.5

157

Height above bed (m)

(b)

158

158.5

159

159.5

157.5

158

158.5

159

159.5

160

0.8 0.6 0.4 0.2 0 155

Time (s)

(c) Height above bed (m)

157.5 Time (s)

160 −1

1ms

0.8 0.6 0.4 0.2 0 155

155.5

156

156.5

157

157.5 Time (s)

158

158.5

159

159.5

160

Figure 8 Demonstration of the capability of the triple-axis CDVP to provide visualizations of intrawave and turbulent flow. Plots in (a)–(c) show a time series over a 5 s period of the zero-mean velocities displayed as vectors u–w, v–w, and u–v, respectively.


44


(b)

0.8

0.8

0.7

0.7

0.6

0.6

0.5

0.5

z (m)

z (m)

(a)

0.4

0.4

0.3

0.3

0.2

0.2

0.1

0.1

0

0 0

100

200

300

10−2

d (µm)

10−1 C (kg

100

m−3)

Figure 9 Profiles of suspended sediments: (a) particle size and (b) concentration.

in Figures 9 and 10. These observations were collected off Santa Cruz Pier, California, as a seasonal storm passed through the area over the period of a couple of days. Figure 9 shows profiles of the suspended sediment particle size and concentration. The profiles of particle size are seen to be relatively consistent with a mean diameter of c. 180 mm near the bed and reducing to c. 120 mm at 0.7 m above the bed. The variability in size is relatively limited and due to changing bed and hydrodynamic conditions as the storm passed by. In Figure 9(b) the suspended concentrations are comparable in their form, although they have absolute values that vary by greater than an order of magnitude. This variation in suspended concentration is associated with the changing conditions as the storm passed through the observational area. It is interesting to note that over the period even though there is a large variation in concentration, the particle size remains nominally consistent. Figure 10 shows the temporal variation in particle size and

concentration with height above the bed. It can be clearly seen that some of the periods of increased particle size are associated with substantial suspended sediment events, as one might expect, however, there are one or two events where the correlation is not as clear. Figures 9 and 10 clearly illustrate the capability of ABS to simultaneously measure profiles of concentration and particle size and the combination of both significantly adds to the assessment and development of sediment transport formulas.

A Case Study of Waves over a Rippled Bed Here the use of acoustics is illustrated by application to a specific experimental study. Over large areas of the continental shelf outside the surf zone, sandy seabeds are covered with wave-formed ripples. If the ripples are steep, the entrainment of sediments into the water column, due to the waves, is considered



45

(a)

0.6

180

µm

z (m)

160 0.4 140 0.2 120

0

14.8

15

15.2

15.4

15.6

15.8

16

16.2

100

(b)

0.6

z (m)

−1.5 0.4 −2

log(kg m−3)

−1

−2.5

0.2

−3 0 14.8

15

15.2

15.4

15.6

15.8

16

16.2

Date Mar. 2003 Figure 10 Measurements of the temporal variations with height above the bed, of: (a) particle size with the color bar scaled in microns and (b) logarithmic concentration with the color bar scaled relative to 1.0 kg m 3.

(a)

(b)

(c)

(d)

(e) v1

v1

v1

v1

(f) v1

v1 v2

v2

v2

Figure 11 A schematic of vortex sediment entrainment by waves over a steeply rippled bed. The arrows show the direction and relative magnitude of the near-bed wave velocity. v1 and v2 are the lee slope-generated vortices.

to be primarily associated with the generation of vortices. This process is illustrated in Figure 11. As shown in Figures 11(a) and 11(b), a spinning parcel of sediment-laden water, v1, is formed on the leeside of the ripple at the peak positive velocity in the wave

cycle. This sediment-rich vortex is then thrown up into the water column at around flow reversal (Figures 11(c) and 11(d)), carrying sediment well away from the bed and allowing it to be transported by the flow. At the same time, a sediment-rich vortex,


46


v2, is being formed on the opposite side of the ripple due to the reversed flow. As shown in Figures 11(d)– 11(f), v2 grows, entrains sediment, becomes detached, and moves over the crest at the next flow reversal carrying sediments into suspension. The main feature of the vortex mechanism is that sediment is carried up into the water column twice per wave cycle at flow reversal. This mechanism is completely different to the flat bed case where maximum near-bed concentration is at about the time of maximum flow velocity. To study this fundamental process of sediment entrainment, experiments were conducted in one of the world’s largest man-made channels, specifically constructed for such sediment transport studies, the Delta Flume; this is located at the De Voorst Laboratory of the Delft Hydraulics in the north of the Netherlands. The flume is shown in Figure 12(a); it is 230 m in length, 5 m in width, and 7 m in depth, and it allows waves and sediment transport to be studied at full scale. A large wave generator at one end of the flume produced waves that propagated along the flume, over a sandy bed, and dissipated on a beach at the opposite end. The bed was comprised of coarse sand, with a mean grain diameter by weight of

(a)

(b)

330 mm, and this was located approximately halfway along the flume in a layer of thickness 0.5 m and length 30 m. In order to make the acoustic and other auxiliary measurements, an instrumented tripod platform was developed and is shown in Figure 12(b). The tripod, STABLE II (Sediment Transport and Boundary Layer Equipment II), used an ABS to measure profiles of particle size and concentration, a pencil beam ARP to measure the bed forms, and, in this case, electromagnetic current meters (ECMs) to measure the horizontal and vertical flow components. Figure 12(c) shows a wave propagating along the flume with STABLE II submerged in water of depth 4.5 m, typical of coastal zone conditions. To investigate and then model the vortex entrainment process it was necessary to establish at the outset whether or not the surface waves were generating ripples on the bed in the Delta Flume. Using an ARP a 3-m transect of the bed was measured over time. The results of the observations over a 90-min recording period are shown in Figure 13. Clearly ripples were formed on the bed and the ripples were mobile. To obtain flow separation and hence vortex formation requires a ripple steepness (ripple height/

(c)

Figure 12 (a) Photograph of the Delta Flume showing the sand bed at approximately midway along the flume and the wave generator at the far end of the flume. (b) The instrumented tripod platform, STABLE II, used to make the measurements. (c) A surface wave propagating along the flume.



47

−0.4

Height (m)

−0.6 −0.8 −1 −1.2 0

20

e

m Ti

40

)

in

(m

60 −0.5

80

0.5 1 1.5

−1

−1.5

0

ce (m)

Distan

Figure 13 Acoustic ripple profiler measurements of a transect of the bed, over time, in the Delta Flume.

ripple wavelength) of the order of 0.1 or greater; an analysis of the observations showed that this was indeed the case. Using the ABS, some of the most detailed full scale measurements of sediment transport over a rippled bed under waves were captured. These measurements from the Delta Flume are shown in Figure 14. The images shown were constructed over a 20-min period as a ripple passed beneath the ABS. The suspended concentrations over a ripple, at the same velocity instants during the wave cycle, were combined to generate a sequence of images of the concentration over the ripple with the phase of the wave. Four images from the sequence have been shown to illustrate the measured vortex entrainment. The length and direction of the arrows in the figure give the magnitude and direction of the wave velocity, respectively. Comparison of Figure 14 with Figure 11 shows substantial similarities. In Figure 14(a), there can be observed the development of a high-concentration event at high flow velocity above the lee slope of the ripple, v1. In Figure 14(b), as the flow reduced in strength, the near-bed sediment-laden parcel of fluid travels up the leeside of the ripple toward the crest. As the flow reverses, this sediment-laden fluid parcel, v1, travels over the crest and expands. As the reverse flow increases in strength (Figure 14(d)), the parcel v1 begins to lift

away from the bed and a new sediment-laden lee vortex, v2, is initiated on the lee slope of the ripple. In order to capture the essential features of these data within a relatively simple, and hence practical, 1-DV (one-dimensional in the vertical) model, the data has first been horizontally averaged over one ripple wavelength at each phase instant during the wave cycle. The resulting pattern of sediment suspension contours is shown in the central panel of Figure 15, while the upper panel shows the oscillating velocity field measured at a height of 0.3 m above the bed. The concentration contours shown here are relative to the ripple crest level, the mean (undisturbed) bed level being at height z ¼ 0. The measured concentration contours presented in Figure 15 show two high concentration peaks near the bed that propagate rapidly upward through a layer of thickness corresponding to several ripple heights. The first, and the strongest, of these peaks occurs slightly ahead of flow reversal, while the second, weaker and more dispersed peak, is centered on flow reversal. The difference in the strengths of the two peaks reflects the greater positive velocity that can be seen to occur beneath the wave crest (time ¼ 0 s) than beneath the wave trough (time ¼ 2.5 s). Between the two concentration peaks the sediment settles rapidly to the bed. Maybe rather unexpectedly this settling effect occurs at the times of strong forward and backward velocity at measurement


48


Concentration (kg m−3)

1

6 cm

v1 v1 0 42 cm

v1

v2

Figure 14 Acoustic imaging of suspended sand entrainment over a rippled bed due to waves, at four phases of the wave velocity. The length of the white arrow in each plot gives the magnitude and direction of the near-bed wave velocity.

levels well above the bed. The underlying mechanism of sediment entrainment by vortices shed at or near flow reversal is clearly evident in the spatially averaged measurements shown in Figure 15. Any conventional ‘flat rough bed’ model that attempts to represent the above sequence of events in the suspension layer runs into immediate and severe difficulties, since such models predict maximum near-bed concentration at about the time of maximum flow velocity, and not at flow reversal. Here therefore, for the first time in a 1-DV model, it has been attempted to capture these effects realistically through the use of a strongly time-varying eddy viscosity that represents the timing and strength of the upward mixing events due to vortex shedding. The model initially predicts the size of the wave-induced ripples and the size of the grains found in suspension, and then goes on to solve numerically the equations governing the upward diffusion and downward settling of the suspended sediment. The resulting concentration contours in the present case are shown on the lower panel of Figure 15. The essential twopeak structure of the eddy shedding process can be seen to be represented rather well, with the initial

concentration peak being dominant. The decay rate of the concentration peaks as they go upward is also represented quite well, though a phase lag develops with height that is not seen to the same extent in the data. Essentially, the detailed acoustic observations of sediment entrainment under waves over ripples of moderate steepness have begun to establish a new type of 1-DV modeling, thereby allowing the model to go on to be used for practical prediction purposes in the rippled regime, which is the bed form regime of most importance over wide offshore areas in the coastal seas.

Discussion and Conclusions The aim of this article has been to illustrate the application of acoustics in the study of near-bed sediment processes. It was not to detail the theoretical aspects of the work, which can be found elsewhere. To this end, measurements of bed forms, the hydrodynamics, and the movement of sediments have been used. These results show that acoustics is progressively approaching the stage where it



Velocity (m s−1)

0.5

0

−0.5 Observed

0.25

z (m)

0.2

0.15

0.1

0.05

Modeled 0.25

z (m)

0.2

0.15

0.1

0.05 0

1

2

3

4

5

Time (s)

−0.5 0.25 log10[Concentration] (re: kg m−3)

1

Figure 15 Measurement and modeling of suspended sediments with height z above a rippled bed under a 5-s period wave.


49

50


can measure nonintrusively, co-located and simultaneously, with high temporal and spatial resolution, all three components of the interacting sediment triad. Figure 16 shows the instrumentation used in a further recent deployment in the Delta Flume. This shows the convergence of the instrumentation and also its use in conjunctions with instruments such as laser in situ scattering transmissometry (LISST). Although substantial advances have been made in the past two decades, the application of acoustics to sediment transport processes is still in an ongoing developmental phase and there are limitations and shortcomings that need to be overcome, and further applications explored. Although there have been few reports to date on data collected using 3-D ARP, such systems are now becoming available. The 3-D ARP is a substantial development over the single-line ARP and the ARS, and should make a considerable contribution to the measurement and understanding of the formation and development of bed forms. There have been a number of reports on single-axis CDVP; however, it is the three-axis CDVP which is the way ahead. Again, these instruments are coming online, though they are still very much a research tool. However, the concept has now essentially been

proven and its use with a vengeance in sediment studies will begin to make an important impact in the next couple of years. It was with the concept of using acoustics to measure suspended sediment concentration that its application to the other components of the small-scale sedimentation processes triad followed. To date the use of sound to measure suspended sediment concentration and particle size has been successful when systems have been deployed over nominally homogeneous sandy beds. However, all who use acoustics recognize that the marine sedimentary environment is frequently much more complicated, and suspensions of cohesive sediments and combined cohesive and noncohesive sediments are common. To employ acoustics quantitatively in mixed and cohesive environments requires the development of a description of the scattering properties of suspensions of cohesive sediments and sediment mixtures. This would be interesting and very valuable work, and should significantly extend the deployment regime over which acoustic backscattering can be employed quantitatively. In conclusion, the objective of this article has been to describe the role of acoustics, in near-bed sediment transport studies. It is clearly acknowledged

ARP

LISST

ARS

PT

PS

ADV ABS

Three-axis CDVP

Figure 16 The instrumentation used in a recent Delta Flume experiment. LISST, laser in situ scattering transmissometry; PS, pumped samples; PT, pressure transducer.



that acoustics is one of a number of technologies advancing our capabilities to probe sediment processes. However, its nonintrusive profiling ability, coupled with its capability to measure all three components of the sediment dynamics triad, make it a unique and very powerful tool for studying the fundamental mechanisms of sediment transport.

See also Acoustic Scattering by Marine Organisms. Acoustics in Marine Sediments. Breaking Waves and Near-Surface Turbulence. Estuarine Circulation. Offshore Sand and Gravel Mining.

51

Hay AE and Mudge T (2005) Principal bed states during SandyDuck97: Occurrence, spectral anisotropy, and the bed storm cycle. Journal of Geophysical Research 110: C03013 (doi:10.1029/2004JC002451). Thorne PD and Hanes DM (2002) A review of acoustic measurements of small-scale sediment processes. Continental Shelf Research 22: 603--632. Vincent CE and Hanes DM (2002) The accumulation and decay of near-bed suspended sand concentration due to waves and wave groups. Continental Shelf Research 22: 1987--2000. Zedel L and Hay AE (1999) A coherent Doppler profiler for high resolution particle velocimetry in the ocean: Laboratory measurements of turbulence and particle flux. Journal of Atmosphere and Ocean Technology 16: 1102--1117.

Further Reading Crawford AM and Hay AE (1993) Determining suspended sand size and concentration from multifrequency acoustic backscatter. Journal of the Acoustical Society of America 94(6): 3312--3324. Davies AG and Thorne PD (2005) Modelling and measurement of sediment transport by waves in the vortex ripple regime. Journal of Geophysical Research 110: C05017 (doi:1029/2004JC002468).

Relevant Websites http://www.aquatecgroup.com – Aquatec Group Ltd. http://www.marine-electronics.co.uk – Marine Electronics Ltd. http://www.pol.ac.uk – POL Research, Proudman Oceanographic Laboratory.


ACOUSTIC NOISE I. Dyer, Marblehead, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 35–44, & 2001, Elsevier Ltd.

Introduction Some ocean scientists consider ambient noise to be a fairly simple and well-behaved property of the ocean. Ambient noise, after all, is often reported and summarized in highly averaged form, its naturally large variance mostly unstated. Other ocean scientists consider the variational complexity of ambient noise a richly colored portrait carrying images of basic ocean processes, including the physics of various noise sources and the acoustics of multiple noise propagation paths. Space and incomplete knowledge precludes a description here that can fully satisfy all ocean scientists or technologists. Instead, the objective is to summarize those aspects of ocean ambient noise that convey the more important recent research results and the more significant remaining research questions. A 1962 summary of ambient noise measurements in the ocean (see Figure 1) is still useful today, at least to classify the various noise sources and their average levels and smooth frequency spectra. Prevailing noises (those observed almost always) are caused by wave–wave interactions at the sea surface, by distributed seismic activity in the earth, by atmospheric or oceanic turbulence, by distant shipping, by windinduced sea surface agitation, and by thermally induced molecular agitation. According to Wenz, wave–wave interaction effects, seismic background, and/or turbulence dominates the noise at VLF (very low frequency band: 1ofo20 Hz), with power spectral density of the pressure field Sðf Þpf 4 . Distant shipping noise dominates at LF (low frequency band: 20ofo200 Hz), has a broad spectral peak around 50 Hz, and falls off sharply for f > 200 Hz as f 6 . At MF (midfrequency band: 200 Hzo fo50 kHz), noise caused by sea surface agitation typically dominates, with a broad peak within 200 Hzofo2 kHz and, beyond f E2 kHz, with Sðf Þpf 1:7 . Finally, molecular agitation typically dominates the noise at HF (high frequency band: f4100 kHz), with Sðf Þpf 2 . Other noise sources are classified as temporally intermittent or spatially discrete, rather than prevailing, and can often dominate. These include

52

sounds from marine earthquakes, from marine animals, from nearby ships or other nearby commercial activities in the ocean, from rain/hail/snow striking the sea surface, and from fractures of ice in the north or south polar oceans. With such a large number of prevailing and other noise sources, the band designations given in the previous paragraph are unlikely to be associated unequivocally with just one noise source or, for that matter, adopted fully by most ambient noise researchers or practitioners. They are of use, however, to help present the material to follow. The spectral summaries used in this Introduction are based on Wenz, and although still useful, modifications and additions are needed in the light of new knowledge. Urick published an excellent summary of ambient noise data acquired in various measurement programs through about 1980. Practitioners commonly use these data, plus the Wenz results, for prediction. Nevertheless, basic understanding of many ambient noise mechanisms through about 1980 was meager and, indeed, some suggested mechanisms were considered speculative. Fortunately, mechanisms for prevailing ambient noises have received considerable research attention since then, particularly from 1985 or so. The other noises have also been researched, in general to a lesser degree. Two volumes edited by Kerman and one by Buckingham and Potter are conference proceedings of recent ambient noise research, and are extraordinary seminal contributions to the understanding of ambient noise mechanisms in the ocean. The continuing flow of research results in archival journals and books, and the aforementioned volumes, provide important modifications and additions to the classical summary of ambient noise by Wenz. In what follows the more important new knowledge, or lack thereof, is summarized.

ULF Band: Wave–Wave Interaction Noise Measurements within 0.1ofo2 Hz, which has come to be called the ultralow frequency band (ULF), extended the Wenzian picture one decade lower in frequency1, and showed ULF noise to be a function of wind speed. The data (Figure 2), have a strong peak f0 located between about 0.2 and 0.7 Hz, with

1 To accommodate the ULF band, the band scheme described in the Introduction is redefined to be VLF2–20 Hz.


ACOUSTIC NOISE

53

120 Intermittent and local effects Earthquakes and explosions

Biologics

Precipitation Ships, industrial activity Sea ice

26

KEY Limits of prevailing noise Wind-dependent bubble and spray noise Low-frequency very-shallow-water wind dependence Heavy precipitation Heavy traffic noise Usual traffic noise _ shallow water Usual traffic noise _ deep water Thermal noise General pattern of noise from earthquakes and explosions Extrapolations

_

Ln ′ Spectrum level, dB re 0.0002 dyn cm 2 and 1 Hz

80

60

6

_14

Wind force (Beaufort) _ 34

_

40

Ln ′ dB re 1 dyn cm 2 and 1Hz (add 100 dB for ref. of 1 μPa and 1 Hz)

100

8 5

_ 54

20 3 2

Prevailing noises Turbulent-pressure fluctuations

0

1

_ 74

Oceanic traffic Bubbles and spray (Surface agitation)

Surface waves _ Second-order pressure effects (Seismic background)

_ 20 1

10

10

2

Molecular agitation 3

10

4

10

5

10

Frequency (Hz) Figure 1 Ambient noise spectra summarized by Wenz. (The ordinates are Ln ¼ 10 log10 Sðf Þ, with respect to the reference value. Add 100 dB to the right-hand scale to obtain Ln in dB re 1 mPa and 1 Hz.) The Beaufort Force translates to wind speeds, in ms 1, as follows: 1, 0.5–1.5; 2, 2–3; 3, 3.5–5; 5, 8.5–10.5; 8, 17–20. (Reproduced from Wenz, 1962.)

Sðf Þ at higher f proportional to about f 3 2f 5 , dependent upon wind speed. A long history of measurements, as well as theoretical surface wave interaction studies, presaged this result. The appearance of systematic data such as in Figure 2 apparently sparked even more research efforts that ultimately confirmed the basic aspects of ULF noise.

Pressure Spectral Density

Wave–wave interaction noise is caused by opposing wind-driven surface gravity waves, each at frequency fw , that to second-order create a pressure field in the water at f ¼ 2fw . (Orders higher than the second have been shown to be negligible.) In its simplest


54

ACOUSTIC NOISE

The noise spectral density Sðf Þ is proportional to [Sw ðf =2Þ2 , the surface elevation density squared and shifted in frequency. Since the frequency dependence of the last three terms in eqn [1] is relatively weak around the peak frequency fwo of Sw , the peak of S is essentially fo E2fwo. Both fwo and Sw are functions of wind speed U, as affected by other sea conditions (fetch, sea age, etc.), and similarly lead to the dependence of fo and S on U.

160

Ambient noise spectrum level (dB re μPa2/Hz)

150

140

Microseism Spectral Density 130

120

110 _

Time (z)

100

90 0.05

Wind speed (m s 1)

17 OCTOBER (0340) 17 OCTOBER (1940)

30 25


15 7.5


5 Uc, with the crossover speed Uc E8 m s1. What noise mechanism could account for all the foregoing observations? An extrapolation of wave– wave interaction noise to the VLF band, from data such as in Figure 2, suggests that Spf 4 U1=2 , give or take one integer in the exponent of f , and one-half integer in the exponent of U. However, this is unacceptably far from the data. The Crouch/Burt data suggest f 1 U0 and f 3 U4 for low and high U, respectively. The overall Nichols result is f 5 U5 from 2 to 5 Hz, and f 0 U5 from 5 to 20 Hz. Other data give f 0:8 U1:3 , f 3:2 U3:4 or an indefinite form, and f 0 U0:5 and f 0 U3 for low and high U, respectively. Without significant modifications applicable to the VLF band, it seems that the wave–wave interaction possibility must be set aside. Next, consider the atmospheric turbulence model. It has evolved as most theories do, but is contentious. It predicts Spf 0 U4 . This, too, is mostly far from the

55

functional form of the foregoing data, but does come close to Nichols and others (at the higher wind speeds) for the 5–25 Hz range. It seems inappropriate, however, to choose among available data sets for the ones that confirm a model. The difference between the data sets may well be caused by some mechanism that we are collectively ignorant of. Finally, it is possible that available data are at least partially contaminated by hydrophone flow noise, whose functional form goes as f 4 . None of the data sets matches this form. Thus, it can be concluded that the hydrophone flow noise mechanism is an unlikely cause of VLF noise. The identification of the mechanism responsible for VLF noise can thus not be made with confidence. In searching for candidate VLF noise mechanisms, one is inclined to look toward appropriate extensions or modifications of mechanisms in the adjacent ULF and LF bands, mainly because wave–wave interactions and distant shipping, respectively, are well established. Nonprevailing mechanisms should also be considered. For example, whale vocalizations are observed for 15of o35 Hz and can affect the VLF band. Noise data sets beyond those referred to here, supported by environmental data as suggested by candidate mechanisms, may well be needed. The Crouch/Burt data set incorporated a plausible but convoluted data analysis path to extract the VLF noise. The reported database of Nichols is not large. The VLF data of other workers could have been affected, as the authors acknowledged, by distant shipping noise. Perhaps because many of these research efforts were aimed at other objectives, environmental data provided with the noise data are generally too fragmentary to aid the search for VLF mechanisms.

LF Band: Distant Shipping Noise Evolving technology has altered the view of distant shipping noise. Increasing use of large aperture acoustic arrays, with attendant high-resolution beamwidths and focused scanning in range and use of highresolution frequency filters, blurs the distinction once sharp between distant and local ships. That is, ambient noise at LF can be observed with high-resolution technology as a countable number of discrete ship noise sources, rather than as a sum of noise from a very large number of widely distributed ships. Frequency Spectra

Figure 3 shows the noise radiated by a contemporary cargo ship. The radiation is largely tonal, as has long


56

ACOUSTIC NOISE

0.5 Hz Bandwidth levels (dB re. 1 μPa at 1 m)

190 B 4

B 3 G F 4 2

180 B F B 1 1 2

170

160

B 5 F 3

G 5

G 6

G 2

140

B 8

B 7

B 9 F 6

B 10

B 11

F 5

G G 8 9

140 rpm G G 10 11

G 3

150

B 6 F 4

G 7

G G G 15 G 12 G 14 16 G G 13 17 18

G 1

G 20

G 30

22

58 rpm 32

25

G 19

27 21

23

24

26

31 28

33

29

130

20

40

60

120 80 100 Frequency (Hz)

140

160

180

200

Figure 3 On-axis source level spectra of a cargo ship at 8 and 16 knots (4 and 8 m s1) measured directly below the ship. Noise levels at distances beyond 1 m may be obtained by subtracting the transmission loss. Levels for bandwidths other than 0.5 Hz can not be determined from this figure because the bandwidths of the tones are not given. (B, F, and G in the figure identify, respectively, the harmonics of the (propeller) blade rate, the (diesel engine) firing rate, and the (ship’s service) generator rate.) (Reproduced from Arveson and Vendittis, 2000.)

been known. However, this is not obvious from the spectra shown in Figure 1, because they entail sums over many ships. The tonal envelopes in Figure 3 maximize between about 20 and 80 Hz, in good agreement with the summation spectra shown in Figure 1. Acoustic propagation losses in the ocean change the shape of the source spectrum shown in Figure 3; above about 80 Hz, the spectrum observed distantly is increasingly reduced with increasing f and with increasing distance from the ship. Directional Spectra

Noise radiated by a ship is a function of azimuth fs , and vertical angle ys , in a cylindrical coordinate system attached to the ship. The azimuthal spectral shape Sf ðf; f ; zÞ observed for a single ship at longer ranges is close to that measured near the ship. However, propagation of the noise to large ranges fundamentally affects the vertical directional spectrum Sy ðy; f ; zÞ. In these spectra, f and y are in a coordinate system attached to the observer (y ¼ 0 is the local horizontal plane). Figure 4 shows Sy as measured in deep water by a vertical line array. It sums over the directional spectrum in azimuth Sf , and therefore over the areal distribution of ships. A prominent feature of Sy is a pedestal of high noise around the horizontal, which is weakly dependent on f and z, and varies in half-width yw from about 715 to 201. These values are consistent with

cosyw Ecz =cb , where cz and cb are the sound speed at the observation depth and at the bottom, respectively. Distant shipping noise in deep water thus arrives mostly from source radiation at the surface near 7ys 7 ¼ 0, and then propagates to the observer via refraction and surface reflection paths in the water. Bottom reflection or transmission losses are relatively high, so that these paths are less important. Other effects influence the pedestal including, but not limited to, surface waves that modulate the source amplitude, scattering rather than specular reflection from the rough sea surface, and scattering from a seamount or continental margin. Because of these oceanographic and topographical complexities, the shape of the noise pedestal in Figure 4 is not general but instead suggestive of the main features of Sy in the deep ocean. In shallow waters, if cz =cb > 1 (downward refracting profile), then Sy is governed by path or mode losses, including those attributable to the bottom. If cz =cb o1 (upward refracting profile), then Sy would have a pedestal, but with yw typically an order of magnitude smaller than that observed in the deep ocean. Summation Issues

With several thousand ships underway in each of the heavily traveled oceans, taking account of all at the same time to determine distant shipping noise is


ACOUSTIC NOISE

Wind speed High Mid Low Low

70 60 50

Power (dB / μPa/√Hz deg)

40 (A)

57

from limits that may be imposed by the focusing, filtering, or beam-forming processor, the spectral valleys are set by the ship’s radiation of continuous rather than tonal noise, and the spatial valleys by the sum of noise from more distant or less powerful ships. In addition, the valleys can be influenced by nonprevailing sources, the most prominent of which at LF is whale vocalization from about 15 to 35 Hz.

High 70

VLF Implications

Mid

60

With respect to the VLF band, the tonal envelope of a ship for f o20 Hz, at all of its higher speeds is about f 0 . Because very low frequency sounds can be detected at transoceanic distances, distant ships could cause the measured result to be Spf 0 U0 found in some VLF experiments, in the apparent but not real absence of distant ships.

Low

50 40 (B)

High 70

Mid

60

40 (C)

MF Band: Wind-driven Sea Surface Noise

Low

50 _

80

_

60

_

40 _ 20 0 20 Angle (deg)

40

60

80

Figure 4 Vertical angle directional spectrum SY summed in azimuth in deep water for f ¼ 75 Hz and U ¼ 3; 7, and 11 m s1. Positive angles are upward looking. (A) Shallow, (B) mid, and (C) deep depth refers to the vertical line array (VLA) centered at z ¼ 850, 1750, and 2650 m. The half-power beamwidth of the VLA at this frequency is about 1.31 when steered to y ¼ 0. (Reproduced from Sotrin and Hodgkiss, 1990.)

neither feasible nor necessary. At one extreme, lowresolution data in the LF band (20–200 Hz) are relatively insensitive to the detailed noise source characteristics of individual ships. Because summation from a large number of ships merges the details, only the broad and slowly evolving trends in shipping lane location, in shipping density, and in shipping composition will affect the level and horizontal directionality of the noise. At the other extreme, high-resolution data resolve the frequency and spatial spectra associated with distant ships, thus giving the experimenter the in situ noise field in relevant detail. The experimenter can avoid such noise in the spectral valleys between tones or the spatial valleys between high-noise beams2. Aside

2

Many ocean processes broaden tonal bandwidths and spatial beamwidths as the sound propagates from source to receiver. Such broadening is typically large enough to be a candidate ocean monitoring tool, but not so much that it completely fills the valleys.

Bubbles created by wind-driven surface waves have long been thought to be the dominant source of prevailing noise in the MF band (0.2–50 kHz). Many basic physical details, however, have only recently become better understood, and some relevant additional questions only recently posed. Various wavebreaking processes of wind-driven surface waves entrain air in the upper part of the ocean. Air-filled bubbles in the water are pinched off from the entrained air, which in turn oscillate and radiate noise as acoustic monopoles. Such noise thus entails wavebreaking and bubble hydrodynamics, both of which are addressed elsewhere in this encyclopedia. The acoustical aspects are addressed here. Vertical Directional Spectrum

In its simplest form, the theory for noise generated by a uniform distribution of sources on the surface is, from ray acoustics, 1 Sþ y ðy; f ; zÞ ¼ ðcs =cz ÞSDfsinys ð1 Rb Rs Þg yw ryrp=2

½2

þ S y ðy; f ; zÞ ¼ Rb Sy ; p=2ryr yw

½3

where cosyw ¼ cz =cs or yw ¼ 0


for for

cz =cs o1; cz =cs 41

½4

ACOUSTIC NOISE

and where the directional spectrum in vertical angle Sy , per unit solid angle, is a function of f (at least through S, the pressure spectral density of the source per unit surface area) and of observation depth z (through sound speed cz at z). The superscripts þ and – refer to positive (upward looking) and negative (downward looking) y obtained from cos cosy ¼ ðcz =cs Þcos ys (the grazing angle ys and sound speed cs pertain to z ¼ 0). Dðys ; f ) is the directivity of the elemental bubble noise sources distributed just below the surface. For present purposes, the surface source distribution S, and the acoustic waveguide, is taken as uniform and independent of range and azimuth, contributions from propagation in the bottom are neglected, and volumetric absorption in the seawater is neglected. (These simplifications are adopted to keep the main ideas clear but, for more precise needs, can readily be replaced by assumptions that are more realistic.) For a downward refracting sound speed profile to depth zðcz =cs o1), the ray theory of [3][4] predicts a refractive shadow zone or notch of width 2yw around y ¼ 0, the horizontal plane. Wave theory must be used to properly predict the field in the notch, which also can be partially filled by scattering of the noise from midwater depths by fish schools and by ocean inhomogeneities. For an upward refracting profile (cz =cs 41), the field around y ¼ 0 is directly due to the surface-generated noise, plus possible scattering contributions. In eqns [2] and [3], Rb and Rs are, respectively, the coefficients of bottom and surface specular reflection. Terms involving these parameters can be important in the directional spectrum (but since perfect reflection is not likely for an acoustic waveguide in the ocean, they do not lead to singularities as eqns [2]–[4] might appear to suggest). For example, consider that Rb and Rs approach unity (but do not reach it) as the grazing angles at the bottom and the surface, respectively, approach zero. Then, for y within about 7p=4; Sy can be increased in typical situations by about 10 dB. In addition, the bottom propagation paths neglected here can actually contribute, especially at the lower MF frequencies. Thus details of the acoustic waveguide affect Sþ y and Sy and, along with the sound speed profile cðzÞ, could account for the plethora of somewhat dissimilar measured MF vertical directional spectra in the literature. Eqn [2] contains the bubble source directivity Dðys ; f ) that, unfortunately, is not known with confidence. At least two models for directive radiation from aggregated bubbles have been considered. One assumes an exponential decrease of uncorrelated monopoles below a horizontal perfectly reflecting

surface, and the other assumes a similarly situated monopole distribution concentrated on a submerged plane. Then, respectively, D ¼ 2f1 sincð2ks dsinys Þg

½5

2

½6

D ¼ 4sin ðks dsinys Þ

where ks is the acoustic wavenumber at the surface, d is the effective depth (the e-folding depth and the d-function depth, respectively), and sinc (x)(sin xÞ=x. In the limit ks d sin ys 51, these functions have the same shape, and close to the same magnitude (E1 dB different). Data, however, show that the two are distinct. For eqn [5], the data suggest ks dEp, whereas for eqn [6] ks dEp In either case, the idealized states assumed in eqn [6] might not represent the relevant complexity of the radiating bubbles beneath a breaking wave. For example, the exponential decay of bubble density with depth may well be a good model for horizontally isotropic bubbles quasistatically present as a result of previous wave breaking events, but a poor model for radiating bubbles immediately caused by a new event. Integration of eqns [2]–[4] over y to obtain the noise spectral density Sðf ; z) also depends sensitively on Rb and Rs (and on possible bottom propagation paths). This emphasizes the need to compare experimentally derived values of Sðf ; z) with appropriate knowledge of the acoustic waveguide. Alternatively, with use of eqns [2]–[4],S may be extracted from vertical line array (VLA) data. When a VLA is steered to y ¼ p=2, the specular reflection and the bottom propagation paths will contribute at most

65

Noise level (dB// 1 μPa2/sr-Hz)

58

60

55

50

45

5

10 25 15 20 20 Log (Wind speed) (kn)

30

Figure 5 Level of the source spectral density S, in dB re mPa2 per m2 Hz, for f ¼ 110 Hz. The 10 m wind speed U is in kn (1 knE1/2 m s1). (Reproduced from Chapman and Cornish, 1993.)


ACOUSTIC NOISE

Source Spectral Density

Chapman and Cornish measured S in deep water with an upward-looking VLA. They apparently assumed eqn [6] for D, with ks dEp=2 . Their data at f ¼ 110 Hz, and for the wind speed interval 2oUo15 m s1, are reproduced in Figure 5, and show a crossover wind speed Uc E4.3 m s1. The frequency interval for their measurements is 13of o300 Hz, within which they found that Uc is about 4.3 m s1 for f r110 Hz, and that Uc is somewhat smaller (E3.5 m s1) for f 4110 Hz. Chapman/Cornish attribute the crossover speed to a transition in source mechanism physics. Furthermore, by regression analyses, their data show that for UoUc ; Spf 2:1 U0:6 , and for U > Uc ; Spf 2:1 U2:7. These results hold on average within the speed and frequency intervals measured. Kewley, Browning, and Carey reviewed and compared several data sets, mostly deep water VLA measurements, to extract S. They also used eqn [6] and ks dEp=2, and concluded that for 30of o 1300 Hz and 1oUo15 m s1, Uc E6 m s1 with SpU1 for UoUc and Sp U3 for U4Uc. The Kewley et al. wind speed exponents of 1 and 3 are not too different from those of Chapman/Cornish. When one considers that the former tilted their exponent choices somewhat to agree with extant physical models proposed for the below and above Uc regimes, the agreement can be considered quite satisfactory.3 What is more relevant, however, is that a universal spectral shape is not evident for either regime in the Kewley et al. comparisons. More likely than not S, f , and D need to be scaled by hydrodynamic parameters other than or additional to U, as shown below. Basic Wave-Breaking Correlates

Research results on hydrodynamically based scaling of noise from breaking waves have been reported. Kerman has proposed that at u =uc E1, where u is the friction velocity and uc is the minimum phase speed of gravity/capillary surface waves, the wavebreaking process transitions from one that has an aerodynamically smooth sea surface to another that

3

When compared at the same U and f , S is about 3 dB higher in the Chapman/Cornish data set than in the data reviewed by Kewley et al.

is rough. Kennedy analyzed VLA data in a deep, acoustically isolated bay (40of o4000 Hz, 2oUo15 m s1), with unlimited wind fetch but limited wave fetch. It was found that u =uc 0.9 defined a rough surface regime. (It may therefore be presumed that the crossover speed discussed in the foregoing section is Uc E0:9uc ) Figure 6 shows that the spectral data for u =uc 40.9 aggregate to an almost universal scalable spectrum. What garners the caveat of ‘almost’ is that frequency is scaled by fp, the observed peak frequency. Both Kerman and Kennedy point out that fp does not vary strongly. It ranges from about 300 to 800 Hz in the Kennedy data, and is not unlike that sketched by Wenz (Figure 1). But experimental interest does not always include measurement of fp , in which case a user of Figure 6 must slide the frequency scale without benefit of Kennedy’s judgement. Neither, however, can properly be accused of intellectual sloth. Kerman provides a model for fp, which contains wave-breaking parameters that unfortunately are poorly known. Kennedy’s collapsed spectral spread although acceptably small, is large enough, and the frequency dependence for 1/3of =fp o10 is weak enough, to forego fine attention to fp . Although apparently not used, fp is related semi-empirically to breaking-wave whitecap size fp (in Hz)E1400/(OLW), where L and W are, respectively, the whitecap along-crest length and crosscrest width, both in meters]. The source spectral density S in Figure 6 is obtained by Kennedy from a dipole directivity model. In effect, eqn [6] was used with ks d51, in this limit known as a compact dipole. With this assumption, the integral of Dðf ; ys Þ over a hemispherical surface _7

10

u* > 0.9 uc _8

10

ADSD (f)B pawcu *3

weakly. Such a measurement is thus dominated by local surface sources, so that S may be compared among measurements with less concern for waveguide properties.

59

_9

10

_ 10

10

_ 11

10

0.01

1

0.10

10

_f fp

Figure 6 Source spectral density S versus f , both nondimensional as described in the text. The source directivity model used is presumably eqn [6] with ks d 51 (the compact dipole model). Data are for the aerodynamically rough regime (u 4uc ). (Reproduced from Kennedy, 1992.)


60

ACOUSTIC NOISE

yields AD ¼ 2p=3 and this appears as one of the S scaling terms in Figure 6.4 Another term is B ¼ 1000 Hz, the nominally observed bandwidth of the noise; its role is simply to create an integral measure SB for dimensional clarity. In the term E ¼ 5ra u3 ; ra is air density, and E is the major scaling variable, the average rate of energy dissipated per unit surface area by the breaking waves. Finally, AD SB=ðrw cs ), where rw is water density, is the average rate of acoustic energy radiated per unit surface area. Virtually simultaneously and independently, others have researched in greater depth some concepts that are related to the Kerman/Kennedy result. Noise spectral density Sðf ; zÞ has been correlated with dissipation E, in deep water under steady wind and wave conditions, for the intervals 4.3of o14 kHz and 2oUo12 m s1. The data on average show SpSpf 0:4 E0:74 , with the exponent of E varying from 0.86 to 0.67 from the low to the high end of the frequency interval. At constant E, the frequency dependence agrees reasonably with an extrapolation of Figure 6. But the dissipation dependence can not be compared without scaling the peak frequency fp, which was not observed. Thus, for a range of E one can seek the range of fp to satisfy linear scaling in E. The peak frequency fp would then need to decrease about a factor of 4 from low to high E, a factor so large as to suggest that a major change in noise physics occurs at these higher frequencies. Does the quasistatic bubble layer below the sea surface increasingly attenuate the noise, or increasingly inhibit its generation, at these frequencies? Bubbles are known to attenuate sound as a function of frequency and size distribution, but data analyses do not consider this. With use of the Fresnel field of an array of hydrophones, sound radiated by individual breaking waves has been measured in deep water (0.35of o4 kHz, 4oUo15 m s1). The on-axis source levels of individual breaking events, were obtained and modeled as spatially and temporally discrete compact dipoles eqn ([6] with ks d51). The individual source levels were correlated with U and cb , the latter being the speed of a breaking wave event, a measure closely connected to breaking wave dissipation E. The correlation with cb was found to

4 A spherical surface for the integral would seem more appropriate, since the only way a monopole can become a dipole is by including the negative image above the free surface, in which case AD ¼ 4p=3. Had a noncompact dipole been assumed with ks d ¼ p=2 (eqn [6]), then AD ¼ 2p. There is as much as 5 dB difference in these values compared to the one used by Kennedy.

be significantly better than that with U and, via physical arguments it was concluded that the source levels are well correlated with E. This measurement technique is also important as it determined the probability density of the dipole source levels, and the spatial density of discrete breaking wave events. It was also concluded, again via physical arguments, that the source spectral density for the frequencies measured are on average pE0.83, which in view of the lower frequencies observed might be taken as reasonably consistent with the E0.74 obtained by other workers. Thus the question remains on a possible frequency-dependent bubble layer effect. The foregoing results clearly have not answered all questions on MF noise caused by breaking waves. They do, however, provide more general predictive tools than those previously available, and identify at least some of the more important physical attributes of noise from breaking waves.

HF Band: Molecular Noise Molecules impinging on the surface of a pressure sensor cause noise, as estimated from physical principles and as plotted in Figure 1. Molecular motion, and thus momentum reversal on the sensor (i. e., force per unit area) is a function of molecular kinetic energy, and thus seawater temperature. On an absolute temperature scale, all oceans may be considered at a constant temperature. Hence, one line in Figure 1 is sufficient to estimate the noise.

See also Acoustics, Arctic. Acoustics, Acoustics, Shallow Water. Ships.

Deep

Ocean.

Further Reading Arveson PT and Vendittis DJ (2000) Radiated noise characteristics of a modern cargo ship. Journal of the Acoustical Society of America 107: 118--129. Buckingham MJ and Potter JR (eds.) (1995) Sea Surface Sound ’94, vol. 3. Singapore: World Scientific, 494pp. Chapman NR and Cornish JW (1993) Wind dependence of deep ocean ambient noise at low frequencies. Journal of the Acoustical Society of America 93: 782--789. Crouch WW and Burt PJ (1972) The logarithmic dependence of surface-generated ambient-sea-noise spectrum level on wind speed. Journal of the Acoustical Society of America 51: 1066--1072. Kennedy RM (1992) Sea surface sound source dependence on wave-breaking variables. Journal of the Acoustical Society of America 91: 1974--1982.


ACOUSTIC NOISE

Kerman BR (1984) Underwater sound generation by breaking wind waves. Journal of the Acoustical Society of America 75: 149--165. Kerman BR (ed.) (1988) Sea Surface Sound, vol. 1. Dordrecht: Kluwer Academic Publishers, 639pp. Kerman BR (ed.) (1993) Sea Surface Sound, vol. 2. Dordrecht: Kluwer Academic Publishers, 750pp. Kewley DJ, Browning DG, and Carey WM (1990) Lowfrequency wind-generated ambient noise source levels. Journal of the Acoustical Society of America 88: 1894--1902. Kibblewhite AC and Evans KC (1985) Wave–wave interactions, microseisms, and infrasonic ambient noise

61

in the ocean. Journal of the Acoustical Society of America 78: 981--994. Nichols RH (1981) Infrasonic ambient ocean noise measurements: Eleuthera. Journal of the Acoustical Society of America 69: 974--981. Sotrin BJ and Hodgkiss WS (1990) Fine-scale measurements of the vertical ambient noise field. Journal of the Acoustical Society of America 87: 2052--2063. Urick RJ (1986) Ambient Noise in the Sea. Los Altos, CA, Peninsula Publishing. Wenz GM (1962) Acoustic ambient noise in the ocean: spectra and sources. Journal of the Acoustical Society of America 34: 1936--1955.


ACOUSTIC SCATTERING BY MARINE ORGANISMS K. G. Foote, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 44–53, & 2001, Elsevier Ltd.

Historical Overview Development of underwater sonar as a tool for navigation and military operations, following sinking of the Titanic in 1912, led inevitably to applications to marine organisms. By the 1930s, echoes from fish schools had been detected. In the 1940s, the deep sound-scattering layer was observed. Its biological origin in mesopelagic fish was identified in the 1950s. At the same time, applications to commercial fish were pursued with vigor, and both scientific echo sounders and fishery echo sounders began to be manufactured. Steady improvements in transduction enabled individual fish of certain species and sizes to be detected at ranges of hundreds of meters. The ultrasonic frequency of 38 kHz was becoming a standard at this time; it was subsequently shown to be near the optimum for achieving detection of commercially important fish in the presence of attenuation due to spherical spreading and absorption. Parallel to studies of single-fish scattering at ultrasonic frequencies were studies of scattering at sonic frequencies, especially to determine the resonance frequency in swimbladder-bearing fish, which is a measure of size. Echo integration was introduced in 1965 as a tool for quantifying fish aggregations at essentially arbitrary conditions of numerical density. This was rapidly developed, and it has been used routinely in surveys of fish stock abundance since about 1975. Introduction of standard-target calibration in the early 1980s served the cause of quantification by providing a rapid, high-accuracy method of enabling the results of echo integration to be expressed in absolute physical units. With few exceptions, standard-target calibration has become the method of choice. Sonar, with one or more obliquely oriented or steerable beams, began to find common application in the 1970s for counting fish schools that might be missed by a vertical echo sounder beam. This was a significant development for acknowledging the narrowness of the sampling volume of vertically

62

oriented directional echo sounder beams and the possibility of fish avoidance reactions to the transducer platform, typically a research vessel. In another parallel development, the Doppler principle was exploited to measure the rate of approach or recession of fish targets. Both horizontally oriented echo sounder beams and sonar beams were used. Early applications determined the swimming speeds of schools of small pelagic fish and individual salmon in rivers. Applications of acoustics to fish in the 1970s were accompanied by notable applications to zooplankton, if pursued less intensively owing to differences in commercial importance. Because of the enormous diversity of zooplankton species in size, shape, and composition, it was recognized early that insonification over a band of frequencies is required, even for routine observation. This has usually been achieved by the use of multiple resonant transducers, but genuinely broadband sonars are also proving successful in yielding spectra of individual euphausiids and copepods. Recognition of the importance of bandwidth in scattering by zooplankton was accompanied by appreciation of the role of interpretive models. Acoustic scattering models have been developed and applied to fish since the 1950s and to zooplankton since the 1970s. The transition from analog to digital technologies in the 1970s facilitated processing of echo data. This has become steadily more automated and sophisticated, but always with operator control of important decisions through the man–machine interface. Other developments in technology since the 1970s have extended the range of applications of acoustic scattering by marine organisms. Multiple-element transducers have been used to determine the threedimensional locations and movement, as well as the target strength, of individual animals. Compact, high-frequency sonars have been mounted on fish capture gear to observe the behavior of fish during catching operations. Steerable high-frequency sonars have been used to track fish schools during capture and to map their three-dimensional shapes.

Physical Basis for Scattering Acoustic scattering by a marine organism is, in principle, no different from that of any other kind of scattering. Differences in the physical properties of the causative bodies with respect to the surrounding


ACOUSTIC SCATTERING BY MARINE ORGANISMS

medium are accompanied by reflection and refraction, or more generally diffraction, of incident waves. Organisms, with contrasts in mass density or elasticity relative to sea water, are thus sources of scattering. The processes of reflection, refraction, and diffraction occur at surfaces, both external and internal, marking discrete changes in physical properties and throughout the volume or inside embedded inhomogeneities, as characterized by continuous changes in properties. The net result of the individual processes is a redistribution in space of the incident energy field. Changes in direction and amplitude characterize the scattering.

Classification of Marine Organisms as Scatterers Marine organisms are conveniently divided into groups based on considerations of taxonomy and anatomy. Two major groups are those of fish and zooplankton, but others are also treated. Fish may be distinguished as cartilaginous or bony. Bony fish may be acoustically distinguished because the fish possesses or lacks a gas-filled swimbladder. Swimbladders may be closed, with gas exchange effected by the rete mirabile, or open, with gas exchange effected by gulping air at the surface or by releasing a sphincter muscle on a duct leading to the exterior. The respective swimbladder types are called physoclists and physostomes. They are illustrated by cod (Gadus morhua) and herring (Clupea harengus), respectively. Some mesopelagic fish possess gas-filled swimbladders, including a number of myctophid species. Some other myctophids, as well as the deepwater fish orange roughy (Hoplostethus atlanticus), possess swimbladders that are invested with wax esters. The whiptail (Coryphaenoides subserrulatus), a macrurid, possesses a swimbladder that contains gas in a spongy matrix of tissue. Swimbladderless fish are illustrated by mackerel (Scomber scombrus). Cartilaginous fish lack a swimbladder, but their liver is large and presents a marked density contrast with the surrounding fish flesh. Zooplankton come in many shapes and sizes, but acoustically their variable physical composition admits of a severe reduction. Three prominent classes have been identified: the liquidlike, the hard-shelled, and the gas-bearing. These are illustrated by, respectively, euphausiids, pteropods, and siphonophores. Other marine organisms have also been detected by scattering. These include squid, gelatinous zooplankton, algae, benthos, marine mammals, and

63

even diving birds. The first five groups are considered in a separate section in the following.

Dependences of Scattering In general, scattering by marine organisms is affected by a number of factors. Some are listed here.

Intrinsic factors. Intrinsic to the scatterers are size, shape, internal composition, and condition. Condition may be affected by the stage of development, presence of reproductive products, and degree of stomach filling. Behavior is another intrinsic factor, if often directly affected or determined by the external environment. It is typically quantified through the attitude, or orientation, of the organism and its velocity of movement. Extrinsic factors. Scattering is affected by the insonification signal, hence by its spectral composition. For impulsive signals, the spectrum may be broadly continuous. For a typical pulsed sinusoid containing many wavelengths, the spectrum will be narrow, and the signal can be characterized by the center frequency, pulse duration, and amplitude. Depth and history of depth excursion may also influence the scattering, as in the case of rapid depth changes for physoclists. For swimbladdered fish lacking rete mirabile, depth excursions will necessarily affect the swimbladder form, with the volume changing in accordance with Boyle’s law, thus inversely with the ambient pressure.

Quantification of Scattering Nomenclature

Scattering properties of organisms are distinguished as belonging to individual organisms or to aggregations of organisms. The fundamental scattering property of a single organism is the scattering amplitude. This is described through the idealization of a plane harmonic wave incident on a finite scattering body. At a great distance r from the body, the scattered pressure field or amplitude psc is related to the incident pressure amplitude pinc by eqn [1]. psc ¼ pinc f expðikrÞ=r

½1

In eqn [1] f is the far-field scattering amplitude, r is the distance from the scatterer, k is the wavenumber 2p/l, and l is the acoustic wavelength. The scattering


64


amplitude f describes the angular characteristics of the scattered field. The differential or bistatic scattering cross-section is jf j2 . In the backscattered direction f ¼ fb, and the backscattering cross-section is given by eqn [2], where the dual convention of using both sbs and s is shown. sbs ¼ jfb j2 ¼

s 4p

½2

The target strength TS is a logarithmic measure (eqn [3]; where r0 is the reference distance, typically 1 m). TS ¼ 10 log

sbs r20

½3

When many scatterers are concentrated in a volume in which individual scatterers cannot be distinguished by their echoes, a collective standard measure of scattering is used. This is the volume scattering coefficient. In the backscattered direction, the volume backscattering coefficient sv is given by eqn [4], where fb,i is the backscattering amplitude for the ith scatterer of N, and V is the volume. sv ¼ V 1

N X fb;i 2

½4

i¼1

The volume backscattering strength is given by eqn [5]. Sv ¼ 10 logðr0 sv Þ

½5

A quantity useful in echo integration is the area or column backscattering coefficient sa, (eqn [6]), where the integration is performed over the range interval [r1, r2].

sa ¼

Zr2

sv dr

½6

r1

In scattering by fish, a numerically more convenient measure of sa is eqn [7], which refers the backscattering to the reference area of one square nautical mile. sA ¼ 4p18522 sa

characteristic or mean backscattering cross section. sA ¼ r A s

Another measure of scattering is the extinction cross-section. This measures the relative loss of energy due to scattering and internal absorption. It may be defined for an individual scatterer, but is generally applied to aggregations of organisms if they are sufficiently numerous. With few exceptions, the issue of calibration must be addressed when making measurements. Standard methods are available for this, the aim being to define the system characteristics so that the result of a measurement, a voltage signal for instance, can be expressed as a pressure-wave amplitude in the water medium. Measurement

There are dozens of techniques for measuring the scattering properties of individual organisms and aggregations of organisms. These are commonly distinguished as being in situ, without constraint in the natural environment of the organisms, or ex situ, hence constrained in some way, wherever this might be. Target strength is a key quantity in many investigations. It may be determined with a single-beam echo sounder; for example, by repeated measurement of similar organisms that are acoustically resolved and by appropriate statistical reduction of these measurements. Alternatively, it may be measured directly with a dual- or split-beam echo sounder, in which the beam pattern can be determined in the direction of the organism, enabling the backscattering cross-section to be extracted from each individual echo. Similar measurements can be performed on single organisms ex situ with greater control and hence knowledge of their state during measurement. Measurements on tethered organisms, constrained to maintain a given orientation during insonification, are popular. Aggregations of organisms are frequently quantified acoustically through the volume backscattering coefficient. If the number and occupied volume of the organisms are known, then the characteristic target strength can be inferred through eqn [9].

½7

This form is particularly useful, for the fundamental equation of echo integration is simply eqn [8], where rA is the numerical density of fish referred to the same area of one square nautical mile, and s is the

½8

Sv ¼ 10 logn þ TS

½9

Here n is the numerical density of organisms, and TS is the so-called mean target strength corresponding to a single organism, but derived as the logarithmic measure of the mean backscattering cross-section.



Cages are often employed to confine a known or knowable number of organisms to a fixed volume. Measurement of Sv can then yield a value for TS. Modeling

The importance of target strength in many studies involving scattering by marine organisms is so great that recourse is frequently made to theoretical models. On the basis of assumptions about the shape and internal composition of subject organisms, mathematical expressions may be derived that can be evaluated for particular conditions of concentration or frequency that might not be realistically explored through measurement. Ultimately, measurements may be used to refine models, and models to interpret measurements.

where g is the ratio of specific heats at constant pressure and volume, P is the ambient pressure at depth, r is the mass density of fish flesh, and a is the equivalent spherical radius. For elongated bubbles or swimbladder shapes, the resonance frequency is modified. The backscattering cross-section s at frequency v is given by eqn [11]. s¼

4pa2 h i2 ½n0 =ðnH Þ2 þ ðn0 =nÞ2 1

Fish as Scatterers The swimbladder shape varies with species and with condition of the individual specimen. An example of a swimbladder in corpus is shown in Figure 1. Here the swimbladder of an Atlantic herring (Clupea harengus) has been exposed by careful dissection. Low frequencies At low frequencies, with acoustic wavelengths much greater than characteristic swimbladder dimensions, the effect of a pressure wave on the swimbladder is essentially that of uniform compression and rarefaction. Consequently, a spherical model can be used. In fact, some swimbladder-bearing fishes have been modelled successfully as a spherical gas bubble surrounded by a finite layer of fish flesh that acts as a viscous fluid medium supporting surface tension on the interface between the shell and fish flesh. The volume of a bubble of radius a is equivalent to that of the swimbladder. Equation [10] gives the resonance frequency v0 of an immersed spherical gas bubble, 1 3gP 1=2 n0 ¼ 2p ra2

2pan x þ 2 ; n0 c pa n0 r

Figure 1 Drawing of a specimen of Atlantic herring (Clupea harengus), female, 36.0 cm long, 453 g, with exposed swimbladder. (Drawing by H. T. Kinacigil, used with permission.)

½12

Some numerical values for the various parameters are r ¼ 1050 kg m3 and x ¼ 50 Pa s. The speed of sound in sea water varies over the range 1450– 1550 m s1, depending on temperature, salinity, and pressure. For gadoids and clupeoids in the size range 8– 30 cm, n0 varies over 2.2–0.3 kHz. Given the inverse relationship of resonance frequency and size in eqn [10], smaller fish will have higher resonance frequencies. Thus, mesopelagic fish with partially waxinvested swimbladders may have resonance frequencies in the low ultrasonic range. Very large swimbladdered fish, say with a total fish length exceeding 1 m, will have resonance frequencies of the order of hundreds of hertz. The corresponding backscattering cross-section, hence target strength, can be computed from eqs [10]–[12]. It is important to note that the quality factor of the resonance condition, eqn [13], where Dn describes the range in frequency over which s decreases to one-half its maximum value, may be of the order of 1.5–3. Q ¼ n0 =Dn

½10

½11

H is the damping factor given by eqn [12], where c is the speed of sound in water, and x is the viscosity of fish flesh. H 1 ¼

Swimbladder-bearing Fish

65

½13

Implicit in the low-frequency condition of the model is that s is independent of orientation. Averages of s with respect to arbitrary orientation distributions will be identical to s itself. When computing average values of s for aggregations of swimbladdered fish of varying size, s must be averaged with respect to the size distribution. The characteristic target strength is determined from the definition in eqn [3]. Intermediate frequencies As the acoustic wavelength decreases toward characteristic


66


swimbladder dimensions, the scattering becomes markedly directional, and the backscattering begins to depend sensitively on the orientation of the fish. From measurements made both in situ and ex situ, the empirical relationship of eqn [14] between mean target strength TS at 38 kHz and total fish length l in centimeters has been derived for a number of gadoids. TS ¼ 20 log l 67:5

½14

Equation [15] applies for clupeoids TS ¼ 20 log l 71:9

½15

The average backscattering cross-section s may be determined immediately from eqn [3]. For a cod of length l ¼ 50 cm, TS ¼ 33.5 dB and s ¼ 56 cm2. For a herring of length l ¼ 30 cm, TS ¼ 42.4 dB and s ¼ 7.2 cm2. Blue whiting is an important commercial stock in both hemispheres, and it is routinely surveyed by acoustics. To convert measurements of acoustic density at 38 kHz to numerical density in accordance with the echo integration equation [8], eqn [16], where l is the fork length in centimeters. is used for the northern-hemisphere blue whiting (Micromesistius poutassou): TS ¼ 21:7 log l 72:8

½16

Equation [17] applies for the southern-hemisphere southern blue whiting (Micromesistius australis), where l is again the fork length in centimeters. TS ¼ 25:0 log l 81:4

½17

Coincidentally, perhaps, the target strength of yellowfin tuna (Thunnus albacares) at 38 kHz is nearly identical to that of Micromesistius australis and is given by eqn [18]. TS ¼ 25:3 log l 80:6

is the total fish length in centimeters). TS ¼ 20 log l 72:7

Some stocks of orange roughy (Hoplostethus atlanticus) are being surveyed about their seamount habitats. Determination of the target strength of this deepwater fish with fat-invested swimbladder is admittedly problematical. Some work suggests convergence of the mean target strength of a 35 cm long orange roughy at 38 kHz to about 48 dB. If the standard equation for mean target strength–length were used, namely eqn [21], TS ¼ 20 log l þ b

½21

the coefficient b would be 79 dB. For modeling scattering by swimbladdered fish at these frequencies, the Kirchhoff approximation model can be used. This assumes that the fish is represented by the swimbladder, which acts as a pressure-release surface where it is directly insonified, and as a surface without response otherwise. A more general scattering model is that of the boundary-element method. The swimbladder is represented by a mesh of points, called nodes, spanning the surface, illustrated in Figure 2. The harmonic wave equation is solved numerically, assuming continuity of pressure and normal component of velocity at each node. It is thus possible to model the effects of internal gas density and pressure. To convert modeled values for s as a function of orientation to an average value, an orientation distribution is required. Ideally, this is done on the basis of in situ observations, but often such data are lacking and an orientation distribution must be assumed. Some orientation distributions are described in the literature. In some special circumstances it has

½18

The target strength of bigeye tuna (Thunnus obesus) under similar conditions is given by eqn [19]. TS ¼ 24:3 log l 73:3

½20

z

½19 x

These relations were established from specimens in the approximate size range 50–130 cm and 3–50 kg. The whiptail (Coryphaenoides subserrulatus), with a swimbladder containing gas-filled spongy tissue, seems to have a mean in situ target strength at 38 kHz that is consistent with the equation developed for another macrurid, the blue grenadier or hoki (Macruronus novaezelandie) (eqn [20], where l

y Figure 2 Boundary element model of the swimbladder of a specimen of pollack (Pollachius pollachius), 34.5 cm in length, with anterior end to the lower right (y direction). (Model by D. T. I. Francis, used with permission.)



been possible to infer the orientation distribution by a combination of acoustic measurement and modeling. The relationship of maximum and average measures of s is given approximately by eqn [22]. smax E7save

½22

Alternatively, eqn [23] can be used. TSmax ETSave þ 5 dB

½23

Measures of the extinction cross-section are relatively rare, there being few occasions when it is necessary to compensate for scattering losses. However, measurement or inference suggest that the extinction cross-section is very roughly 1–3 times the backscattering cross-section at intermediate frequencies. Ultimately, the cross-sections and their ratio must depend on the behavior of the organism, as quantified through the orientation distribution. High frequencies When the acoustic wavelength becomes very small compared to the swimbladder size, scattering by other tissues may become important. The contributions of head structure, vertebrae, and even scales at very high frequencies have been established through ex situ measurement. Modeling of scattering by such structures can be computationally excessive, suggesting the advantages of stochastic modeling if direct measurement is not possible or convenient. Swimbladderless Fish

The mackerel is a prominent example of a swimbladderless fish. Its target strength must be attributed to the non-swimbladder structures and hence is intrinsically complicated at nearly all frequencies. At intermediate frequencies, the mean target strength is roughly 10 dB less than that of a gadoid of comparable size (eqn [24]). TSmackerel ETSgadoid 10 dB

½24

For cartilaginous fish, such as sharks, the liver may be very large. In pelagic sharks, this may be of the order of 7–23% by weight; in demersal sharks, 3–6%. The specific gravity of lipids is of the order of 0.87–0.92 in pelagic sharks and 0.93–0.94 in demersal sharks, further suggesting the role of the liver in buoyancy and its significance in acoustic scattering. At least for the pelagic sharks, the size and difference in mass density may explain much of the target strength. Were a model to be constructed,

67

a pelagic shark might be represented by a body with the size, shape, and physical properties of the liver.

Zooplankton as Scatterers Liquid-like Bodies

A number of prominent and abundant zooplankton can be classified as liquidlike in their acoustic properties. Extensive modeling and measurement have demonstrated that internal shear waves have negligible influence in scattering by such organisms. The animals are thus generally fluidlike in their properties. If the same animals lack sizable organs or other tissue presenting large contrasts in mass density or compressibility relative to the sea water immersion medium, then the acoustic properties of the organisms are more particularly liquidlike, and their acoustic scattering is consequently relatively weak. Two examples of zooplankton with liquidlike properties are euphausiids and copepods. These are also representative of homogeneous and inhomogeneous scatterers, respectively. Homogeneous liquidlike bodies The expectation of relatively weak scattering by euphausiids has been confirmed by measurement. For example, the target strength of Antarctic krill (Euphausia superba) of mean lengths 30–39 mm is in the range from 88 to 83 dB at 38 kHz and from 81 to 74 dB at 120 kHz. The respective acoustic wavelengths are 39 and 12.5 mm. For a scattering body that is relatively long compared to the wavelength, the scattering will be inherently directional. Laboratory measurement has demonstrated strong effects of orientation on scattering by euphausiids in the size range 30–42 mm at frequencies of 120 kHz and higher. In modeling scattering by homogeneous liquidlike zooplankton, there are just two significant material properties, the mass density and compressibility, or longitudinal-wave sound speed. A variety of models can be used to represent shape. At low frequencies, a single euphausiid can be represented by a finite circular cylinder or even a sphere, with volume equal to that of the animal. At higher frequencies, the same animal might be represented as a finite, bent, tapered cylinder or, better, by the actual shape of the exoskeleton. Scattering models for euphausiids have demonstrated the sensitive dependence of target strength on both the material properties and orientation of the organism. Given the rarity of measurements of material properties, their seasonal and individual


68


variability, and the generally unknown orientation, there has been little systematization of measured values of target strength. Theoretical understanding of scattering by euphausiids has succeeded in associating large lobes with the echo spectrum at rather short acoustic wavelengths. When these are combined with knowledge of the target strength to within about an order of magnitude, it is possible to classify euphausiids by their acoustic signature. Inhomogeneous liquidlike bodies Copepods, like euphausiids, also display relatively weak acoustic scattering. Unlike euphausiids, however, their internal structure is acoustically distinct, being composed of two dominant scatterers, a prosome and an embedded oil sac. Because of the low density of lipids in the oil sac, of the order of 900 kg m3, the prosome must be correspondingly more massive. Because the copepod body as a whole is close to neutral buoyancy in sea water, the target strength is due to the internal contrast in mass density and compressibility, or longitudinal-wave sound speed, between the prosome and oil sac. Measurement has shown that the target strength of a 2 mm long copepod, Calanus finmarchicus, is in the approximate range from 95 to 90 dB over the frequency range 1600–2400 kHz. Copepods have been modeled as composite twoliquid-body structures. Numerical values for the mass density and longitudinal-wave sound speed have been derived from measurements or have been assumed. The shapes of embedded oil sac and encompassing prosome, illustrated in Figure 3, have been determined from videomicroscopic cross-sections in dorsal and lateral views. Results of modeling of copepods have shown the expected weak dependence on orientation at low or moderate frequencies, and an overall mean target strength that is in line with measured values. Hard-shelled Bodies

An example of a hard-shelled zooplankton is the pteropod Limacina retroversa, a marine snail with a spiral shell, opercular opening, and wings that propel it through the mid-water column. The target strength of specimens of shell length 2 mm has been measured over the approximate frequency range from 350 to 750 kHz. The target strength varies between 80 and 60 dB, depending on both frequency and orientation. The pteropod has been modeled as a rough spherical shell with a circular opening. Predictions of scattering have been in reasonable agreement with

z

y x z

y x Figure 3 Boundary element models of the prosome and oil sac of a specimen of Calanus finmarchicus, stage 6 female, 2.74 mm in length, with anterior end to the lower left (x direction). (Models by D. T. I. Francis, used with permission.)

measurements at wavelengths roughly comparable to the maximum shell dimension. Gas-bearing Bodies

Siphonophores are representatives of gas-bearing zooplankton, with gas inclusions in the pneumatophores. These are generally small compared to overall dimensions of specimens, and the target strength varies widely over the frequency range 350– 750 kHz. In particular, the target strength varies over the range from 90 to 60 dB, but with no apparent systematic dependence on frequency. This wide range is suggestive of interference between echoes from the gas inclusions and the nongaseous tissue, the basis of an acoustic model.

Other Organisms as Scatterers Squid

A number of specimens of squid have been observed by acoustics. These include Todarodes pacifica, Loligo opalescens, and Loligo vulgaris reynaudii. In a survey of the second species, performed at



120 kHz, the target strength of specimens of mean dorsal mantle length 11.6 cm and mean mass 23.7 g was about 59 dB. Thus in the standard target strength–length equation [21], but with l representing the mean dorsal mantle length, b is about 80 dB. For Todarodes pacifica of mean dorsal mantle length 16 cm and mean mass 95 g, the target strength is about 51 dB at 28.5 kHz and 55 dB at 96.2 kHz, corresponding to values of b of 75 and 79 dB, respectively. For Todarodes pacifica of mean dorsal mantle length 23.7 cm and mean mass 340 g, the respective mean target strengths at 28.5, 50, 96.2, and 200 kHz are about 45.7, 46.5, 48.0, and 47.6 dB, with respective values of b of 75, 74, 76, and 76 dB. For Loligo vulgaris reynaudii, the target strength was measured at 38 kHz for sufficiently dispersed animals of mean mass 300 g. The target strength when referred to 1 kg was 42.5 dB. This compares favorably with the measurements on Loligo opalescens at 120 kHz and Todarodes pacifica at 28.5 kHz. When expressed relative to 1 kg, the respective target strengths are 42.3 and 41.1 dB.

69

and the masses 0.33–0.8 kg. Target strengths expressed relative to 1 kg of biomass vary from 35 to 28 dB at 50 kHz, from 33 to 24 dB at 70 kHz, and from 29 to 22 dB at 200 kHz. Smaller algae, the phytoplankton Prorocentrum micans, Peridinium triquetrum, Olistodiscus luteus, Dunaliella salina, Platimous viridis, and Phaeodactilum tricornutum, are also being studied by acoustics. Measurements of reverberation, in particular, are being used in attempts to quantify the volume of gas vacuoles. Clams

Both the razor clam (Tagelus dombeii) and the surf clam (Mesodesma donacium) have been surveyed by acoustics. Beds of the razor clam have been surveyed in shallow water over a flat bottom. Echograms that show the bottom–surface–bottom reflection in addition to the first bottom reflection show an enhanced registration above the so-called second bottom echo. Counting of its characteristic serrations provides a quantitative measure of clam density. Marine Mammals

Common Jellyfish

In anticipation of acoustic surveying of the ctenophore Mnemiopsis leidyi and other gelatinous zooplankton, namely Aurelia aurita and Pleurobrachia pileus, in the Black Sea, measurements have been made of the target strength of the common jellyfish Aurelia aurita. Functional regression equations have related the mean target strength in decibels to the disk diameter d in centimeters. At 120 kHz, the relation is eqn [25]. TS ¼ 14:7 log d 74:6

½25

At 200 kHz it is eqn [26]. TS ¼ 39:6 log d 104:4

½26

Thus for a specimen with mean diameter 10 cm, TS ¼ 59.9 and 64.8 dB at the respective frequencies. Algae

Algae, such as kelp, are being surveyed by acoustics. For purposes of quantification, the acoustic properties of the plants themselves are being studied, both by experiment and by theoretical modeling. Measurements have been performed on leaves of Laminaria saccarina and L. digitata at three ultrasonic frequencies. The lengths of these span the range 0.7– 2 m; the widths 0.4–0.9 m; the thicknesses 1–5 mm;

A few measurements have been reported on the target strength of the sperm whale (Physeter catodon) and the humpback whale (Megaptera novaeangliae) in situ. Measurements have been made of the Atlantic bottlenose dolphin (Tursiops truncatus) in captivity. Measurements made on a 2.2 m long 126 kg female dolphin in broadside aspect at the surface revealed a mean target strength that decreased from about 10 dB at the lowest measurement frequency of 23 kHz to about 24 dB at 45 kHz, rising to about 20 dB at 65 kHz, then falling to 25 dB at 80 kHz. The observed degree of variability about these nominal values due to repeated insonification was 4–11 dB to within the first standard deviation to either side.

Challenges For all of the instances and applications of acoustic scattering by marine organisms, there is an enormous demand for enhanced imaging capability and more quantitative understanding, including both improved measurement methods and models. In addition to refinement of current measurement methods, including those for quantifying concentrations of marine organisms, instruments are being developed or adapted for application. These include high-frequency sonars, multibeam sonars, and continuously broadband echo sounders, operating at both low and high frequencies.


70


In general, the addition of bandwidth to acoustic devices, whether achieved by multiple frequencies or a continuous spectrum, is a firm objective of many development efforts. Its usefulness in classification is appreciated from certain studies in zooplankton scattering, but it would aid studies of nekton scattering if successful. Recognition of the importance of understanding the acoustic properties of individual organisms is similarly influential in promoting developments and applications. Determining the properties of single organisms when found en masse remains a challenge, as does quantifying avoidance reactions or avoiding inducing them. While there are many techniques for determining target strength, their application requires ingenuity to elucidate some of the principal dependences. The general lack of information on the depth dependence of target strength for gas-bearing organisms is a particular, prominent example. Modeling of scattering by marine organisms offers much potential for resolving physically intractable problems, such as those involving separation of echoes from individual organisms in the midst of their social aggregations or inferring the acoustic properties of organisms that are very fragile or that occur in extreme environments. Both analytical and numerical models, however, require knowledge of the physical properties, shape, and behavioral characteristics, such as the orientation distribution, of the subject organisms. Acoustic inference of the in situ properties of organisms, by special measurement techniques and aided by models, appears very attractive if generally difficult. An enhanced imaging capability based on acoustic scattering is also valuable. If realized in a compact device, this could aid fishing practice, as in providing fishers with information on the species and size of organisms present in the water column or on the bottom without actually having to capture the organism to make the determination. For the researcher, being able to distinguish different organisms with overlapping distributions would be invaluable in aiding the study of relationships, ultimately to advance the goals of ecosystem analysis and understanding.

Acknowledgments This is Woods Hole Oceanographic Institution contribution number 10271.

See also Acoustics, Deep Ocean. Bioacoustics. Fish Locomotion. Mesopelagic Fishes. Pelagic Fishes. Plankton. Sonar Systems.

Further Reading Craig RE (ed.) (1984) Fisheries Acoustics A symposium held in Bergen, 21–24 June 1982, Rapports et ProcesVerbaux des Reunions, vol. 184. Copenhagen: International Council for the Exploration of the Sea. Foote KG (1997) Target strength of fish. In: Crocker MJ (ed.) Encyclopedia of Acoustics, vol. 1, pp. 493--500. New York: Wiley. Foote KG and Stanton TK (2000) Acoustical methods. In: Harris RP, et al. (ed.) ICES Zooplankton Methodology Manual, pp. 223--258. London: Academic Press. Freon P and Misund OA (1999) Dynamics of Pelagic Fish Distribution and Behaviour: Effects on Fisheries and Stock Assessment. Oxford: Fishing New Books. Karp WA (ed.) (1990) Developments in Fisheries Acoustics, A symposium held in Seattle, 22–26 June 1987, Rapports et Proces-Verbaux des Reunions, vol. 189. Copenhagen: International Council for the Exploration of the Sea. Margetts AR (ed.) (1977) Hydro-acoustics in Fisheries Research, A symposium held in Bergen, 19–22 June 1973, Rapports et Proces-Verbaux des Reunions, vol. 170. Copenhagen: International Council for the Exploration of the Sea. Medwin H and Clay CS (1998) Fundamentals of Acoustical Oceanography. San Diego, CA: Academic Press. Nakken O and Venema SC (eds.) (1983) Symposium on Fisheries Acoustics, Selected papers of the ICES/FAO Symposium on Fisheries Acoustics, Bergen, Norway, 21–24 June 1982, FAO Fisheries Report no. 300. Rome: Food and Agriculture Organization of the United Nations. Ona E (1990) Physiological factors causing natural variations in acoustic target strength of fish. Journal of the Marine Biological Association of the United Kingdom 70: 107--127. Physics of Sound in the Sea (1969) Reprint of the 1946 edition. Washington, DC: Department of the Navy. Progress in Fisheries Acoustics (1989) Proceedings of the Institute of Acoustics, vol. 11, no. 3. St. Albans: Institute of Acoustics. Simmonds EJ and MacLennan DN (eds.) (1996) Fisheries and Plankton Acoustics, Proceedings of an ICES international symposium held in Aberdeen, Scotland, 12–16 June 1995. ICES Journal of Marine Science, vol. 53, no. 2. London: Academic Press. Urick RJ (1983) Principles of Underwater Sound, 3rd edn. New York: McGraw-Hill.


ACOUSTIC SCINTILLATION THERMOGRAPHY P. A. Rona, Rutgers University, New Brunswick, NJ, USA C. D. Jones, University of Washington, Seattle, WA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: Mapping Diffuse Flow at Seafloor Hydrothermal Sites Acoustic scintillation thermography or AST is a method to detect and map diffuse flow at seafloor hydrothermal sites. Diffuse flow is the discharge of low-temperature hydrothermal solutions (up to tens of degrees Celsius) as seepage through areas of the seafloor. It is considered widespread at low intensity (temperature less than 1 1C; flow rate less than 1 cm s 1) in ocean basins and at higher intensity (temperature less than 100 1C; flow rate less than 1 m s 1) in seafloor hydrothermal fields. Quantitative assessment of diffuse flow is important because the cumulative thermal flux of diffuse flow through areas of the seafloor may equal or exceed that of focused flow from associated high-temperature (up to 405 1C) higher-flow-velocity (flow rate tens to hundreds of centimeters per second) black smoker vents. Chemical fluxes in diffuse flow may be selectively significant. Measuring diffuse flow is difficult. Occurrence is patchy and temperatures and flow velocities are low. Because it is clear, lacking the suspended particulate matter of black smoker plumes, diffuse flow cannot be detected by measuring attenuation or backscatter of light and sound.

Method The AST method uses the phase-coherent correlation of acoustic backscatter from consecutive sonar

scans of the seafloor to detect weak fluctuation in the index of refraction of the water near the seafloor. Random index of refraction changes result from temporal variations in the water temperature caused by turbulent mixing, which produce detectable changes in travel time of an acoustic ray as the ray propagates from a stationary acoustic transducer through the diffuse flow to the seafloor and is scattered back through the diffuse flow. These changes in travel time increase from the nearer to the farther portion of the sonar footprint, causing the associated echo waveform to change in shape with each transmission and resulting in decorrelation of successive sound pulses. In essence, the scintillation of the acoustic wave as it passes through a turbulent flow field and scatters off the underlying seafloor makes the seafloor appear to shimmer, much as the eye would detect hot water or hot air shimmering against a stationary backdrop. For certain types of seafloor hydrothermal flow (such as the case with diffuse flow where buoyant turbulent microplumes are concentrated near the seafloor), the AST method can provide a means of remotely detecting areas of flow and a potential method of measuring the scales of temperature and velocity fluctuations in the near-bottom boundary layer. Consider the scattering geometry illustrated in Figure 1, where a turbulent boundary layer overlays a rough seafloor and the incident acoustic field propagates forward and back between the transducer and the backscattering seafloor through the boundary layer. Assume that the seafloor is not changing in time and there is no motion of the acoustic transducer, but the turbulent flow is evolving temporally. Between consecutive and rapid sonar returns from the same spot on the seafloor, turbulent mixing will cause slight spatial and temporal fluctuations of the index of refraction of the water near the seafloor. These fluctuations will cause weak random forward scattering of the acoustic field as it propagates

z Stationary source receiver

Vertical beam pattern

Incident field Turbulent flow x

Seafloor Figure1 Propagation of an acoustic pulse through a turbulent boundary layer near the seafloor produced by diffuse hydrothermal flow and backscatter of the pulse from the seafloor.


71

72

ACOUSTIC SCINTILLATION THERMOGRAPHY

through the turbulent flow, resulting in temporal fluctuations in the amplitude and phase of the acoustic field measured at the receiver. The temporal changes in backscatter can be measured by estimating the phase-coherent temporal decorrelation of backscatter between successive echoes, whereas the backscatter intensity from the turbulence itself is too weak to measure. In this case, the rate and magnitude of temporal decorrelation can be used to detect and potentially measure scales of mixing in the boundary layer.

Application The method was first applied to detect and map diffuse flow in seafloor hydrothermal fields using a human occupied vehicle (HOV). Several later experiments have shown that deep ocean diffuse flow can be detected over relatively large areas of the seafloor using a remotely operated vehicle (ROV) platform. In these experiments, backscatter was recorded using a 200 kHz multibeam sonar mounted on an ROV as it hovered above the seafloor at intervals along a track line. Figure 2 shows an area of detected diffuse flow (as verified by video) and the corresponding backscattered intensity image. The two images are derived from the same sonar scans of the seafloor taken at 0.1 s intervals. Future applications will extend the AST method to seafloor observatories to record changes in area of diffuse flow at a seafloor site on timescales up to years. An informative representation is to drape areas of diffuse flow detected by the AST method over a corresponding bathymetric map to show the relation between diffuse flow and seafloor morphology. Backscatter intensity (dB)

Volume and heat fluxes from areas of diffuse flow can be estimated when in situ measurements of water temperature and vertical flow velocity are made in the areas of diffuse flow. For example, in situ measurements in diffuse flow of temperature using thermistors and of vertical flow velocity using video to record rise rate against a calibrated rod in an acoustically imaged area of diffuse flow and black smokers in the Main Endeavour Field indicated that heat flux of the diffuse flow was a multiple of that of the associated black smokers. Figure 3 shows areas of decorrelation intensity (draped onto bathymetry) for a relatively large area of seafloor surveyed at the Clam Bed hydrothermal field on the Juan de Fuca Ridge (near the Main Endeavour Field site). The AST method was applied using a sonar mounted on a hovering ROV to map the diffuse flow over a 3500 m 900 m area of the axial valley of the Endeavour segment of the Juan de Fuca Ridge. Using in situ sensors to simultaneously measure temperature and flow velocity, a diffuse heat flux of 150 MW was integrated over the areas of the AST decorrelation anomalies, indicating that the heat transferred by diffuse flow is a significant component of total heat flux in the study area. The temporal correlation of the backscattered field is found by correlating collocated returns from the seafloor as a function of transmission time. Consider the monostatic geometry for backscatter as illustrated in Figure 1, where a source/receiver system is fixed in space and measures backscatter at discrete pulse transmission times tn. The temporal correlation of backscatter signals between a transmission at time tm and a later transmission at tn is defined as Cðr; tm ; tn Þ ¼ huðr; tm Þu ðr; tn Þi

50

50

40

40

30

30

20

20

10

10

0

20

0 Meters

20

40

0

½1

Decorrelation intensity (dB)

20

0 Meters

20

40

Figure 2 An AST image of diffuse flow near Hulk vent in the Main Endeavour Field on the northern Juan de Fuca Ridge is shown in the right panel. The areas of diffuse flow (light patches) are detected as an increase in the decorrelation intensity. A conventional sonar image of the same area is shown in the left panel. Colors indicate level of backscatter intensity (red is highest intensity) related to bottom roughness.



73

30 20 10

50 100 150

Me

150

ter

s

200 100 250 50 300

Figure 3 Areas of diffuse hydrothermal flow from the seafloor (Clam Bed site on the northern Juan de Fuca Ridge) detected by backscatter decorrelation (area in meters; Jones et al., 2000). Colors indicate level of decorrelation intensity (yellow is higher intensity) corresponding to intensity of diffuse flow.

where u(r; tn) ¼ A(r; tn) exp[if(r; tn)] is the complex envelope of the backscattered field as a function of range r along the seafloor. Range is determined by round-trip acoustic time-of-flight as is usual in sonar. The complex envelope, also known as the baseband signal, incorporates the amplitude (A) and phase (f) of the echo signal, and the asterisk in eqn [1] denotes complex conjugation. For m ¼ n, the correlation function is real and proportional to the signal intensity. With man, the correlation function will be complex due to signal fluctuations in amplitude and phase, with phase dominating. If the area of seafloor observed has not changed because of motion of the source/receiver, fluctuations in the signal will be due to changes in the water above the seafloor as a function of time tn. With m ¼ 0 as the reference time and n ¼ 1, 2, 3,y increasing monotonically, eqn [1] is the cumulative temporal correlation function. For a single realization and finite number of discrete data points, an estimate of the ideal correlation function is found by windowing the backscattered signals in range and summing over range bins, M X ˆ i ; t0 ; t n Þ ¼ 1 uðri ; t0 Þu ðri ; tn Þ Cðr M i¼1

½2

where ri is a set of M discrete range values that correspond to a finite area (or patch) of the seafloor centered at range r.

The correlation function is a measure of the change in the scattering medium, but noise and motion will also cause signal decorrelation. To correct for the effects of low signal-to-noise levels when the return from the seafloor is weak, it is convenient to define the decorrelation intensity as Ir ðr; t0 ; tn Þ ¼ Iðr; t0 ; tn Þ½1 rðr; ˆ t0 ; tn Þ

½3

where the normalized temporal correlation coefficient is .qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ˆ ˆ t n ; tn Þ ˆ t0 ; t0 Þ Cðr; t0 ; tn Þ rðr; Cðr; ˆ t0 ; tn Þ ¼ Cðr; ½4 The scalar intensity I is the geometric mean of the intensity in the range window, defined as

Iðr; t0 ; tn Þ ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ˆ t0 ; t0 Þ Cðr; ˆ tn ; t n Þ Cðr;

½5

The decorrelation intensity [3] is useful as a relative measure of change when the scattering strength over an area of the seafloor is nearly uniform and helps avoid the comparison of areas where the scattered signal level is relatively low. Motion of the


74


source/receiver is more difficult to compensate and will depend on the randomness in the seafloor. Angular motion of the source/receiver between recordings should be much smaller than the angular correlation scale of the scattering from the seafloor. If there is motion the rate of signal decorrelation due to motion must be less than the rate of signal decorrelation due to the diffuse flow. The size of the area of seafloor (or patch) used to estimate the correlation is defined by the length of the transmitted pulse and the number of samples needed to form an accurate estimate. For a bandwidth-limited Gaussian signal with bandwidth B, it is well known that the normalized mean-square error of the correlation estimate is given as h i h i ˆ t0 ; tn Þ ¼ 1 1 þ r2 ðr; t0 ; tn Þð1 þ N=SÞ2 e2 Cðr; 2BT ½6 where r is the desired true correlation coefficient and T is the length of the record used to form the estimate (corresponding to the number of points M or patch size). The signal-to-noise ratio (S/N) is assumed constant and uncorrelated between transmission times t0 and tn.

Scattering by Marine Organisms. Acoustics, Deep Ocean. Acoustics in Marine Sediments. Acoustics, Shallow Water. Bioacoustics. Dispersion from Hydrothermal Vents. Hydrothermal Vent Biota. Hydrothermal Vent Deposits. Hydrothermal Vent Ecology. Hydrothermal Vent Fluids, Chemistry of.

Further Reading Bendat JS and Piersol AG (2000) Random Data: Analysis and Measurement Procedures, 3rd edn., pp. 291–296. New York: Wiley-Interscience. Johnson HP, Hautala SL, Tivey MA, et al. (2002) Survey studies hydrothermal circulation on the northern Juan de Fuca Ridge. EOS, Transactions, American Geophysical Union 83(73): 78--79. Jones CD, Jackson DR, Rona PA, and Bemis KG (2000) Observations of hydrothermal flow (abstract). Journal of the Acoustical Society of America 108(5): 2544-2545. Rona PA, Jackson DR, Wen T, Jones CD, Mitsuzawa K, and Bemis KG (1997) Acoustic mapping of diffuse flow at a seafloor hydrothermal site: Monolith Vent, Juan de Fuca Ridge. Geophysical Research Letters 24(19): 2351--2354.

See also Acoustic Measurement of Near-Bed Sediment Transport Processes. Acoustic Noise. Acoustic


ACOUSTICS IN MARINE SEDIMENTS T. Akal, NATO SACLANT Undersea Research Centre, La Spezia, Italy Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 72–88, & 2001, Elsevier Ltd.

Introduction Because of the ease with which sound can be transmitted in sea water, acoustic techniques have provided a very powerful means for accumulating knowledge of the environment below the ocean surface. Consequently, the fields of underwater acoustics and marine seismology have both used sound (seismo-acoustic) waves for research purposes. The ocean and its boundaries form a composite medium, which support the propagation of acoustic energy. In the course of this propagation there is often interaction with ocean bottom. As the lower boundary, the ocean bottom is a multilayered structure composed of sediments, where acoustic energy can be reflected from the interface formed by the bottom and subbottom layers or transmitted and absorbed. At low grazing angles, wave guide phenomena become significant and the ocean bottom, covered with sediments of different physical characteristics, becomes effectively part of the wave guide. Depending on the frequency of the acoustic energy, there is a need to know the acoustically relevant physical properties of the sediments from a few centimeters to hundreds of meters below the water/ sediment interface. Underwater acousticians and civil engineers are continuously searching for practical and economical means of determining the physical parameters of the marine sediments for applications in environmental and geological research, engineering, and underwater acoustics. Over the past three decades much effort has been put into this field both theoretically and experimentally, to determine the physical properties of the marine sediments. Experimental and forward/ inverse modeling techniques indicate that the acoustic wave field in the water column and seismoacoustic wave field in the seafloor can be utilized for remote sensing of the physical characteristics of the marine sediments. Sediment Structure as an Acoustic medium

Much of the floor of the oceans is covered with a mixture of particles of sediments range in size from

boulder, gravels, coarse and fine sand to silt and clay, including materials deposited from chemical and biological products of the ocean, all being saturated with sea water. Marine sediments are generally a combination of several components, most of them coming from the particles eroded from the land and the biological and chemical processes taking place in sea water. Most of the mineral particles found in shallow and deep-water areas, have been transported by runoff, wind, and ice and subsequently distributed by waves and currents. After these particles have been formed, transported, and transferred, they are deposited to form the marine sediments where the physical factors such as currents, dimensions and shapes of particles and deposition rate influence the spatial arrangements and especially sediment layering. Particles settle to the ocean floor and remain in place when physical forces are not sufficiently strong to move them. In areas with strong physical forces (tidal and ocean currents, surf zones etc.) large particles dominate, whereas in low motion energy areas (ocean basins, enclosed bays) small particles dominate. During the sedimentation process these particles, based on the physical and chemical interparticle forces between them, form the sedimentary acoustic medium: larger particles (e.g., sands) by direct contact forces; small particles (e.g., clays and fine silts) by attractive electrochemical forces; and silts, remaining between sands and clays are formed by the combination of these two forces. The amount of the space between these particles is the result of different factors, mainly size, shape, mineral content and the packing of the particles determined by currents and the overburden pressure present on the ocean bottom. Figure 1 is an example of a core taken very carefully by divers, to ensure an undisturbed internal structure of the sediment sample. Sediment structures have been quantified by using X-ray computed tomography to obtain values of density with a millimeter resolution for the full three-dimensional volume. The image shown in Figure 1 is a false color 3D reconstruction of a core sample at a site where sediments consist of sandy silt (75% sand, 15% silt, and 10% clay) and shell pieces. The results of the X-ray tomography of the same core, can also be shown on an X-ray cross-section slice along the center of the core (Figure 2A) and the corresponding two-dimensional spectral density levels for that cross-section (Figure 2B). The complex


75

76

ACOUSTICS IN MARINE SEDIMENTS

300 250 200 150 100 110

50

100

350 300 250

60 200 150 100

10

50

90

80

50

40

30

70

20

Figure 1 Three-dimensional reconstruction of a sediment core sample containing 75% sand, 15% silt and 10% clay. Scale in mm.

_ 100

CT number 2250

X-ray CT Section

dB re m

_ 55

_ 80

_ 20

_ 60 2200

20 2150

40

60

_

0

Vertical frequency (m 1)

Vertical distance (mm)

2

Spectral density level

2100

80

_ 60

_ 40

_ 65

_ 20 _ 70

0 20

_ 75

40 _ 80

60 80

100 20

(A)

40

60

80

100

120

140

Horizontal distance (mm)

160

2050

_ 85

100 _ 100 _ 80 _ 60 _ 40 _ 20

(B)

0

20

40

60

_

80

100

Horizontal frequency (m 1)

Figure 2 (A) X-ray cross-section slice along the center of the core shown in Figure 1. (B) 2-D spectral density levels for the crosssection shown in (A).

structure of the sediments can be with seen strong heterogeneity (local density fluctuation) of the medium that controls the interaction of the seismoacoustic energy. In addition to the complex fine structure described above, the seafloor can also show complex layering (Figure 3A) or a simpler structure (Figure 3B). These structures result from the lowering of sea levels during the glaciation of the Pleistocene epoch during which sand was deposited over wide areas of the continental shelves. Unconsolidated

sediments subsequently covered the shelves as the sea level are in postglacial times.

Biot–Stoll Model Various theories have been developed to describe the geoacoustic response of marine sediments. The most comprehensive theory is based on the Biot model as elaborated by Stoll. This model takes into account various loss mechanisms that affect the response of


77


N 42 56.536 E10 13.836

19-JUL-1999 Time UTC 18:30

18:20

18:10

18:00

17:50

17:30

17:40

17:20

17:10

N 42 48.615 E10 13.899 17:00

75

0.120

90

0.140

105

0.160

120

0.180

135

0.200

150

3

2

5

4

6

Twtt (s)

N 42 36.054 E10 47.726 13:50 0.140

9

10

11

13

12

03-JUL-1999 Time UTC 13:40

13:30

13:20

N 42 34.525 E10 52.215 13:00

13:10

14

12:50 105

0.160

120

0.180

135

0.200

150

0.220

165

0.240

180

6.50 (B)

7 8 Range (km)

6.00

5.50

5.00

4.50

3.00 4.00 3.50 Range (km)

2.50

2.00

1.50

1.00

Figure 3 Seismic reflection profiles showing (A) complex and (B) simple structures of sediment layers.


0.50

0

Depth (m)

1 (A)

Depth (m @ 1500 m s

Twitt (s)

_1

)

0.100

78


porous sediments that are saturated with fluid. The Biot–Stoll theory shows that acoustic wave velocity and attenuation in porous, fluid-saturated sediments depend on a number of parameters including porosity, mean grain size, permeability, and the properties of the skeletal frame. According to the Biot–Stoll theory, in an unbounded, fluid-saturated porous medium, there are three types of body wave. Two of these are dilatational (compressional) and one is rotational (shear). One of the compressional waves (‘first kind’) and the shear wave are similar to body waves in an elastic medium. In a compressional wave of the ‘first kind’, the skeletal frame and the sea water filling the pore space move nearly in phase so that the attenuation due to viscous losses becomes relatively small. In contrast, due to out-of-phase movement of the frame and the pore fluid, the compressional wave of the ‘second kind’ becomes highly attenuated. The Biot theory and its extensions by Stoll have been used by many researchers for detailed description of the acoustic wave–sediment interaction when basic input

parameters such as those shown in Table 1 are available.

Seismo-acoustic Waves in the Vicinity of the Water–Sediment Interface As mentioned above, when acoustic energy interacts with the seafloor, the energy creates two basic types of deformation: translational (compressional) and rotational (shear). Solution of the equations of the wave motion shows that each of these types of deformation travels outward from the source with its own velocity. Wave type, velocity, and propagation direction vary in accordance with the physical properties and dimensions of the medium. The ability of seafloor sediments to support the seismo-acoustic energy depends on the elastic properties of the sediment, mainly the bulk modulus (incompressibility, K) and the shear modulus (rigidity, G). These parameters are related to the compressional and shear velocities CP and CS respectively, by: CP ¼ ½ðK þ 4G=3Þr1=2

Table 1

Basic input parameters to Biot–Stoll model

Frequency-independent Variables Porosity (%) P Mass density of grains rr Mass density of pore fluid rf Bulk modulus of sediment grains Kr Bulk modulus of pore fluid Kf Variables affecting global fluid motion Permeability k Viscosity of pore fluid Z Pore-size parameter a Structure factor a Variables controlling frequency-dependent response of frame Shear modulus of skeletal frame m¯ ¼ mr ðoÞ þ imi ðoÞ Bulk modulus of skeletal frame K¯ b ¼ Kbr ðoÞ þ iKbi ðoÞ

Table 2

CS ¼ ðG=rÞ1=2 where, r is the bulk density. Table 2 shows basic seismic–acoustic wave types and their velocities related to elastic parameters These two types of deformations (compressional and shear) belong to a group of waves (body waves) that propagate in an unbounded homogeneous medium. However, in nature the seafloor is bounded and stratified with layers of different physical properties. Under these conditions, propagating energy undergoes characteristic conversions every time it interacts with an interface: propagation velocity,

Basic seismo-acoustic wave types and elastic parameters

Basic wave type Body wave Ducted wave

Wave velocity Compressional

CP ¼ ½ðK þ 4G=3Þ=r1=2

Shear

CS ¼ ðG=rÞ1=2

Love

CL ¼ ðG=rÞ1=2

CSCH ¼ ðG=rÞ1=2 Elastic parameters in terms of wave velocities (C) and bulk density (q) Bulk modulus (incompressibility) K ¼ rðCP2 4CS2 =3Þ Compressibility b ¼ 1=K Young’s modulus E ¼ 2CS2 rð1 þ sÞ Poisson’s ratio (transv./long. strain) s ¼ ð3E rCP2 Þ=ð3E þ 2rCP2 Þ Shear modulus (rigidity) G ¼ rCS2 Lame’s constant l ¼ rðCP2 2CS2 Þ Surface wave

Scholte



energy content, and spectral structure and propagation direction changes. In addition, other types of waves, i.e., ducted waves and surface waves, may be generated. These basic types of waves and their characteristics together with their arrival structure as synthetic seismograms for orthogonal directions are illustrated in Figure 4.

Body Waves

These waves propagate within the body of the material, as opposed to surface waves. External forces can distort solids in two different ways. The first involves the compression of the material without changing its shape; the second implies a change in shape without changing its volume (distortion). From earthquake seismology, these compressional and distortion waves are called primus (P) and secoundus (S), respectively, for their arrival sequence on earthquake records. However, the distortional waves are very often called shear.

79

Shear waves Shear waves are those in which the particle motion is perpendicular to the direction of propagation. These waves can be generated at a layer interface by the incidence of compressional waves at other than normal incidence. Shear wave energy is polarized in the vertical or horizontal planes, resulting in vertically polarized shear waves (SV) and horizontally polarized shear waves (SH). However, if the interfaces are close (relative to a wavelength), one cannot distinguish between body and surface waves. Ducted Waves

Love waves Love waves are seismic surface waves associated with layering; they are characterized by horizontal motion perpendicular to the direction of propagation (SH wave). These waves can be considered as ducted shear waves traveling within the duct of the upper sedimentary layer where total reflection occurs at the boundaries; thus the waves represent energy traveling by multiple reflections. Interface Waves

Compressional waves Compressional waves involve compression of the material in such a manner that the particles move in the direction of propagation.

Propagation direction

Basic wave types z A. Body waves

Seismic interface waves travel along or near an interface. The existence of these waves demands the combined action of compressional and shear waves. Thus at least one of the media must be solid, whereas

Particle motion

y

x x

Compressional

Shear

Synthetic seismogram

Sv Sh

B. Ducted waves

Love C. Surface waves

Scholte

Figure 4 Basic seismo-acoustic waves in the vicinity of a water–sediment interface.


y

z

80


the other may be a solid (Stoneley wave), a liquid (Scholte wave), or a vacuum (Rayleigh wave). For two homogeneous half-spaces, interface waves are characterized by elliptical particle motion confined to the radial/vertical plane and by a constant velocity that is always smaller than the shear wave. When different types of seismic waves propagate and interact with the layered sediments they are partly converted into each other and their coupling may create mixed wave types in the vicinity of the interface. Figure 4 shows the basic seismo-acoustic waves in the vicinity of a water/sediment interface. Basic characteristics of these waves together with their particle motion and synthetic seismograms at three orthogonal directions (x, y and z) are also illustrated in the same Figure. Under realistic conditions, in which the seafloor cannot be considered to be homogeneous, isotropic, nor a half-space, some of these waves become highly attenuated or travel together, making identification very difficult. In fact, the interface waves shown in Figure 4 are for a layered seafloor, where the dispersion of the signal is evident (homogeneous half-space would not give any dispersion). Under realistic conditions, i.e., for an inhomogeneous, bounded and anisotropic seafloor, some of these waves convert from one to another. The different wave types may travel with different speeds or together, and they generally have different attenuation. As an example, Figure 5 shows signals from an explosive source (0.5 kg trinitrotoluene (TNT))

Hydrophone

a

c

b

Vertical

d

received by a hydrophone and three orthogonal geophones placed on the seafloor, at a distance of 1.5 km from the source. The broadband signal of a few milliseconds duration generated by the explosive source is dispersed over nearly 18 s demonstrating the arrival structure of different types of waves. The characteristics of these waves are indicated in the figure for four different sensors. They can be identified in order of their arrival time as: (a) head wave; (b) water arrival (compressional wave); (c) interface wave; (d) Love waves.

Seafloor Roughness The roughness of the water–sediment interface and layers below is another important parameter that needs to be considered in sediment acoustics. The seafloor contains a wide spectrum of topographic roughness, from features of the order of tens of kilometers, to those of the order of millimeters. The shape of the seafloor and its scattering effect on acoustic signals is be covered here.

Techniques to Measure Geoacoustic Parameters of Marine Sediments The geoacoustic properties of the seafloor defined by the compressional and shear wave velocities, their attenuation, together with the knowledge of the material bulk density, and their variation as a function of depth, are the main parameters needed to solve the acoustic wave equation. To be able to determine these properties of the seabed, different techniques have been developed using samples taken from the seafloor, instruments and divers conducting measurements in situ, and remote techniques measuring seismo-acoustic waves and inverting this information with realistic models into sediment properties. Some of the current methods of obtaining geoacoustic parameters of the marine sediments are briefly described here.

Transversal

Laboratory Measurements on Sediment Core Samples c Radial

0

2

4

6

8

10

12

14

16

18

20

22

Time (s) Figure 5 Signals from an explosive source of 0.5 kg trinitrotoluene received by a hydrophone and three orthogonal geophones at 1.5 km distance (a, Head wave; b, Compressional wave; c, Interface wave; d, Love wave).

Most of our knowledge of the physical properties of sediments is acquired through core sampling. A large number of measurements on marine sediments have been made in the past. In undisturbed sediment core samples, under laboratory conditions, density and compressional velocity can be measured with accuracy, and having measured values of density and compressional velocity, the bulk modulus can be selected as the third parameter, where it is can be calculated (Table 1).



There are several laboratory techniques available to measure some of the sediment properties. However, the reliability of such measurements can be degraded by sample disturbance and temperature and pressure changes. In particular, the acoustic properties are highly affected by the deterioration of the chemical and mechanical bindings caused by the differences in temperature and pressure between the sampling and the laboratory measurements. Controlling the relationships between various physical parameters can be used to check the accuracy of measured parameters of sediment properties. It has been shown that the density and porosity of sediments have a relationship with compressional velocity, and different empirical equations between them have been established. Over the past three decades, at the NATO Undersea Research Centre (SACLANTCEN), a large number of laboratory and in situ measurements have been made of the physical properties of the seafloor sediments. These measurements have been conducted on the samples with the same techniques as when the same laboratory methods were applied. This data set with a great consistency of the hardware and measurement technique, has been used to demonstrate the physical characteristics of the sediment that affect acoustic waves. The relationship between measured physical parameters From 300 available cores, 20 000 measured data samples for the density and porosity, and 10 000 samples for the compressional velocity were obtained. To be able to handle this large data set taken from different oceans at different water depths, all bulk density and compressional velocity data were converted into relative density ðrÞ and relative velocity ðCÞ with respect to the in situ water values:

than used here and shown to have a strong linear correlation. Theoretically this linearity only exists if the dry densities of the mineral particles are the same for all marine sediments. The density of the sediment would then be the same as the density of the solid material at zero porosity, and the same as the density of the water at 100% porosity. Figure 6 shows this relationship. Porosity and relative compressional wave velocity The relationship between porosity and compressional wave velocity has received much attention in the literature because porosity can be measured easily and accurately. Data from the SACLANTCEN sediment cores giving the relationship between porosity and compressional wave velocity are shown in Figure 7. As shown in Figure 6 due to the linear relation between density and porosity, the relationship between density and compressional wave velocity is similar to the porosity compressional wave velocity relation. In situ Techniques

There are several in situ techniques available that use instruments lowered on to the seafloor mainly by means of submersibles, remotely operated vehicles (ROVs), and autonomous underwater vehicles AUVs and divers. The first deep-water, in situ measurements of sediment properties were made from the bathyscaph Trieste in 1962. These measurements provided accurate results due to the minimum disturbance of temperature and pressure changes compared to bringing the sample to the surface and for 2.6

= 2.635_ 0.01675 P Number of data points = 13885

2.4

where, rS is sediment bulk density, rW is water density, CS is sediment compressional velocity and CW is water compressional velocity. The data are not only from the water–sediment interface but cover sedimentary layers of up to 10 m deep. Relative density and porosity The density and porosity of the marine sediment are least affected during coring and laboratory handling of the samples. Porosity is given by the percentage volume of the porous space and sediment bulk density by the weight of the sample per unit volume. The relationship between porosity and bulk density has been investigated by many authors with fewer data

Relative density ( = s /w)

r ¼ rS =rW C ¼ CS =CW

81

2.2 2.0 1.8 1.6 1.4 1.2 1.0 0

20

40 60 Porosity (%)

80

100

Figure 6 Relationship between relative density and porosity.


82


Figure 9 Examples of received compressional wave signals from fine sand sediment.

Figure 7 Relationship between relative compressional wave velocity and porosity.

Figure 10 Examples of signals recorded from two shear wave receivers in hard-packed fine sand sediments. Figure 8 In Situ Sediment Acoustic Measurement System (ISSAMS) for near-surface in situ geoacoustic measurements.

analysis in the laboratory. The most reliable direct geoacoustic measurement techniques for marine sediments are in situ techniques. Some of the approaches used in recent years are discussed below. Near-surface method A system has been developed to measure sediment geoacoustic parameters, including compressional and shear wave velocities and their attenuation at tens of centimeters below the sediment–water interface. Figure 8 shows the main features of the In situ Sediment Acoustic Measurement System (ISSAMS). Shear and compressional wave probes are attached to a triangular frame that uses weights to force the probes into the sediment. In very shallow water, divers can be used to insert the probes into the sediment, whereas in deeper water, a sleeve system (not shown) allows the ISSAMS to penetrate into the seafloor.

Using the compressional and shear wave transducers measurements are made with a continuous wave (cw) pulse technique where the ratio of measured transducer separation and pulse arrival time yields the wave velocity. Samples of compressional and shear wave data are shown in Figure 9 and 10 respectively. Table 3 gives a comparison of laboratory and in situ values of compressional and shear wave velocities from two different types of Adriatic Sea sediment. Cross-hole method Measurements as a function of depth in sediments can be made with boreholes, using either single or cross-hole techniques. Boreholes are made by divers using water–air jets to penetrate thin-walled plastic tubes for cross-hole measurements. Figure 11 shows the experimental set up for the cross-hole measurements. The source is in the form of an electromagnetic mallet securely coupled to the inner wall of one of the plastic tubes with a hydraulic clamping device.



Table 3

83

Comparison of laboratory and in situ measurements Wave velocity (ms1)

Sediment type

Porosity (%)

Mean grain size (f)

In situ (CP)

Laboratory (CP)

In situ (CS)

Laboratory (CS)

37 68

3.5 8.6

1557–1568 1467–1488

1580–1604 1468–1487

78–82 27–31

50 15

Sand Mud

Pump

Projector Clamp

Receiver

S

Geophones Clamp

Electromagnetic mallet

Figure 11 The experimental setup for cross-hole measurements.

0

4

Time

0

_4

Spectrum level (dB)

Relative velocity amplitude (V)

8

0.1 s

_ 20

_ 40

_ 60

_8

5 (B)

(A)

10

50

100

Frequency (Hz)

Figure 12 Cross-hole shear wave signal (A) and its spectrum (B) for a silty-clay bottom in the Ligurian Sea.

With a separation range of 2.9 m, moving-coil geophone receivers are also coupled to the inner wall of the second plastic tube with a hydraulic clamping device. The electromagnetic mallet generates a point source on the thin plastic tube wall, which results in a multipolarized transient signal to the sediment. Depending on the orientation of receiving sensors, vector components of the propagating compressional and shear waves are received and

analyzed for velocity and attenuation parameters. Examples of a time series and a frequency spectrum for a shear wave signal received by a geophone with a natural frequency of 4.5 Hz are shown in Figure 12. Figure 13 shows shear wave velocity as a function of depth obtained in the Ligurian Sea. It can be seen that the shear wave velocities are around 60 m1 s at the sediment interface and increase with depth.


84


Remote Sensing Techniques

Even though in situ techniques provide the most reliable data, they are usually more time consuming and expensive to make and they are limited to small areas. Remote sensing and inversion to obtain geoacoustic parameters can cover larger areas in less time and provide reliable information. These techniques are based on the use of a seismo-acoustic 0

signal received by sensors on the seafloor and/or in the water column. Figure 14 illustrates a characteristic shallow water signal from an explosion received by a hydrophone close to the sea bottom. Three different techniques to extract information relative to bottom parameters from these signals are described briefly below. Reflected waves The half-space seafloor. When the seafloor consists of soft unconsolidated sediments, due to its very low shear modulus it can be treated as fluid.

1

Depth (m)

2 w 3

r

Water

Cw

Sediment

s

t

Cs 4

5 0

75 100 50 _ Shear wave velocity (m s 1)

125

Figure 13 Shear wave velocity profile obtained from cross-hole measurements.

Figure 15 Geometry and notations for a simple half-space water–sediment interface.

Refracted wave

Interface wave

Waterborne wave 0.2

0.4

0.6 Time (ms)

0.8

1.0

Figure 14 A characteristic shallow-water signal.


1.2

1.4


Reflection may occur whenever an acoustic wave is incident on an interface between two media. The amount of energy reflected, and its phase relative to the incident wave, depend on the ratio between the physical properties on the opposite sides of the interface. It is possible to calculate frequency

independent reflection loss as the ratio between the amplitude of the shock pulse and the peak of the first reflection (after correcting for phase shift, absorption in the water column, and differences in spreading loss). If the relative density r ¼ rS =rW, and the relative compressional wave velocity

0°

0°

20°

20°

Grazing angle

angle Grazing

85

40°

40°

60°

60°

80°

80°

Time (ms)

Time (ms) 0

24

12 0°

Grazing angle

0

90°

24

12 0°

Grazing angle

90°

10

20

dB Loss Relative sound speed 1.00 SAND

POR.(%) 60 1.05

1.1

REL. DENSITY

1.9

4m

1.05

60 1.1

100

REL. DENSITY

1.9

1 2

2 3

1.00

CLAY

1

CLAY

POR.(%)

Relative sound speed

100

SAND

3 4 5m

(A)

(B)

Figure 16 Acoustic signals and reflection loss as a function of grazing angle and sediment core properties: (A) for critical angle; (B) for angle of incidence case.


86


However, the phase of the reflected wave is then shifted relative to the phase of the incidence wave by an angle varying from 01 to 1801 and is given as:

C ¼ CS =CW , are used to present the contrast between the two media, reflection coefficient for such a simple environmental condition can be written as:

p

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rsinY ð1=C2 cos2 YÞ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi R¼ rsinY ð1=C2 cos2 YÞ

F ¼ 2 arctan

ðcos2 Y 1=C2 Þ rsinY

Figure 16 shows measured and calculated reflection losses (20 log R) for this simple condition together with the basic physical properties of the core sample taken in the same area for explosive signals at different grazing angles.

Where Y is the grazing angle with respect to the interface as shown in Figure 15. For an incident path normal to a reflecting horizon, i.e., Y ¼ 901, the reflection coefficient is:

Angle of intromission case. Especially in deepwater sediments the sound velocity in the top layer of the bottom is generally less than in the water above (Co1). In such conditions there is an angle of incidence at which all of the incident energy is transmitted into the sedimentary layer and the reflection coefficient becomes zero:

R ¼ rC 1=rC þ 1 Critical angle case. When the velocity of compressional wave velocity is greater in the sediment layer (C>1), as the grazing angle is decreased, a unique value is reached at which the acoustic energy totally reflects back to the water column. This is known as the critical angle and is given by:

cosYi ¼

p r

2

1=C2 r2 1

The phase shift is 01 when the ray angle is greater than the intromission angle and 1801 when it is smaller. Thus, acoustical characteristics of the bottom, such as the critical angle, the angle of incidence, the phase shift,

Ycr ¼ arccosð1=CÞ When the grazing angle is less than this critical angle, all the incident acoustic energy is reflected. Direct signal Reflected signal D(t )

R(t )

F

F

D(f )

R(f ) R(f ) D(f ) H (f ) Transfer func.

F

Impulse response

_1

20 Time (ms)

0 A (f ) e Reflection coefficient A (f )

i(f )

Phase shift (f ) Reflection loss _ 20 log A dB

+180° 0° _ 180° 0 10 20 30 dB 0

1

2

3 kHz

Figure 17 Acoustic reflection data and analysis technique for a layered seafloor.


4

5

40

Loss dB

Grazing angle


Computed

87

Measured

0

36° 10 20 10 54° 20 30 10 72° 20 30

0

2 kHz

4

2 kHz

0

4

Impulse responses Measured

Computed 18° 36° 54° 72° 90° 0

10 0 Time (ms) For layered bottom

10 Time (ms)

Figure 18 Measured and calculated reflection losses for a layered seafloor.

and the reflection coefficient, are primarily influenced by the relative density and relative sound velocity of the environment as a function of the ray angle. The layered seafloor. Since the seafloor is generally layered, a simple peak amplitude approximation cannot be implemented because of the frequency dependence. In this case one calculates the transfer function (or reflection coefficient) from the convolution of the direct reference signal with the reflected signal. Examples of the phase shift and reflection loss as a function of frequency are shown in Figure 17. The reflectivity can also be described in the time domain by the impulse response, which is the inverse Fourier transform of the transfer function as shown in the same figure. This type of data can be utilized for inverse modeling to obtain the unknown parameters of the sediments. Figure 18 is an example of measured and calculated reflection losses and impulse responses for a layered seafloor.

ray

one ar

Geoph

Explosive source

Figure 19 Experimental setup to measure seismo-acoustic waves on the seafloor.

Refracted waves Techniques developed for remote sensing of the uppermost sediments (25–50 m below the seafloor) utilize broad-band sources (small


88


0 0

25

50

75

100

Offset (m) 125 150

175

200

225

250

500

Time (ms)

1000 1500 2000 2500 3000

8.39 8.50 4.87 8.84 7.22 8.94 8.40 10.26 10.00

6.81 6.96 3.64 7.26 7.40 7.54 7.68 7.82 7.95 8.08 8.21 4.24 8.46 8.59 8.71 8.83 8.95 9.07

4000

0.19 0.23 0.06 0.27 0.57 0.68 0.37 0.77 0.47 3.52 3.75 3.76 3.92 5.02 3.09 3.86 3.44 4.88 6.49

3500

Figure 20 Signals received by geophone array.

The slope of the travel-time curve (fitted parabola) gives the rate of change of the range with respect to time that is also the velocity of propagation of the diving compressional wave at the level of its deepest penetration (turning point) into the sediment. At each range D, the depth corresponding to the deepest penetration is then calculated using the following integral zðVÞ ¼ 1=p

Z

D

cosh1 ðVðdt=dxÞÞdx

0

where 1=V ¼ ðdt=dxÞx ¼ D

Offset (m) 0

80

160

240

320

15

05

m

50 Time (ms)

explosives) and an array of geophones deployed on the seafloor. To obtain estimates of the bottom properties as a function of depth, both refracted compressional and shear waves as well as interface waves are analyzed. Inversion of the data is carried out using modified versions of techniques developed by earthquake seismologists to study dispersed Rayleigh-waves and refracted waves. Figure 19 illustrates the basic experimental setup and Figure 20 the signals received by an array of 24 geophones that permit studies of both interface and refracted waves. These data are analyzed and inverted to obtain both compressional and shear wave velocities as a function of depth in the seafloor. Studies of attenuation and lateral variability are also possible using the same data set. Figure 21 shows the expanded early portion of the geophone data shown in Figure 20, where, the first arriving energy out to a range of about 250 m has been fitted with a curve showing that compressional waves refracted through the sediments just beneath the seafloor travel faster as they penetrate more deeply into the sediment. At zero offset, the slope indicates a velocity of 1505 m s1 whereas at a range offset at 250 m the slope corresponds to a velocity of 1573 m s1. At ranges over 250 m, a strong head wave becomes the first arrival and, the interpretation would be that there is an underlying rock layer with compressional wave velocity of about 4577 m s1. A compressional wave velocity–depth curve for the upper part of the seafloor can be derived from the first arrivals shown in Figure 21 using the classical Herglotz–Bateman–Wieckert integration method.

100 150

s

_ 1

157

3m

s _1

4577 m

s

_1

200 250

Figure 21 Expanded early portion of the data shown in Figure 21.



The result is the solid velocity–depth curve shown in Figure 22.

_

1500 0

Compressional wave velocity (m s 1) 1600

Data

Depth (m)

Interface waves In order to obtain a shear wave velocity–depth profile from the data, later arrivals corresponding to dispersed interface waves may be utilized (Figure 20). The portion of each individual signal corresponding to the interface-wave arrival can be processed using multiple filter analysis to create a group velocity dispersion diagram (Gabor diagram). The result of applying this technique to a dispersed signal is a filtered time signal whose envelope reaches a maximum at the group velocity arrival time for a selected frequency. The envelope is computed by taking the quadrature components of the inverse Fourier transform of the filtered signal. Filtering is carried out at many discrete frequencies over selected frequency bands. Once the arrival times are converted in to velocity, the envelopes are arranged in a matrix and contoured and dispersion curves are obtained by connecting the maximum values of the contour diagram (Figure 23). Having obtained the dispersion characteristics of the interface waves, the geoacoustic model, made of a stack of homogeneous layers with different compressional and shear wave velocities for each layer that predict the measured dispersion curve is determined. Figure 24 illustrates a number of examples from the Mediterranean sea covering data from soft clays to hard sands.

10

20 Predicted

Figure 22 Compressional wave velocity versus depth curves derived from data and predicted (dashed line) by the model.

120

_

Velocity (m s 1)

100

80

60

40 2

6

10

89

14

2

Frequency (Hz) Figure 23 Dispersion diagram and measured and predicted dispersion curves.


6 Frequency (Hz)

10

90


0

Depth (m)

10

20

Mediterranean Sea West Atlantic 30 0

400

200

600

_

Velocity (m s 1) Figure 24 Summary of shear wave velocity profiles from the Mediterranean Sea.

60 Compressional

Depth (m)

50

TL (dB)

Water

SD: 50 m RD: 50 m

60 Sand

Data Calculate

d

70

Shear

70

80 Predicted Sand model Clay model

Clay

90

80 0

2

4

8 6 Range (km)

10

12

0

500

1000

1500

2000

2500

_

Velocity (m s 1)

Figure 25 Comparison of transmission loss (TL) data and SAFARI prediction. Effects of changing bottom parameters from sand to clay is also shown together with the input profiles utilized for SAFARI predictions.

Transmission Loss Technique

The seafloor is known to be the controlling factor in low-frequency shallow water acoustic propagation. Forward modeling is performed with models giving exact solutions to the wave equation, i.e., SAFARI where, compressional and shear wave velocities, the attenuation factors associated with these waves, and the sediment density as a function of depth are the main input parameters. Acoustic energy propagating through a shallow water channel interacts with the seafloor causing partitioning of waterborne energy into different types of seismic and acoustic waves. The propagation and attenuation of these waves observed in such an environment are strongly dependent on the physical characteristics of the sea bottom. Transmission loss (TL), representing the

amount of energy lost along an acoustic propagation path, carries the information relative to the environment through which the wave is propagating. Figure 25 shows a comparison of TL data and model predictions together with the input parameters used at 400 Hz. The effects of changes in bottom parameters are also shown in the figure. This technique becomes extremely useful when seafloor information is sparse.

Conclusions Acoustic/seismic characteristics of the marine sediments have been of interest to a wide range of activities covering commercial operations involving trenching and cable lying, construction of offshore



foundations, studies of slope stability, dredging and military applications like mine and submarine detection. Experimental and theoretical work over the years has shown that it is possible to determine the geoacoustic properties of sediments by different techniques. The characteristics of the marine sediments and techniques to obtain information about these characteristics have been briefly described. The studies so far conducted indicate that direct and indirect methods developed over the last four decades may give sufficient information to deduce some of the fundamental characteristics of the marine sediments. It is evident that still more research needs to be done to develop these techniques for fast and reliable results.

See also Acoustics, Deep Ocean. Acoustics, Shallow Water. Benthic Boundary Layer Effects. Calcium Carbonates. Clay Mineralogy. Ocean Margin Sediments. Pore Water Chemistry. Seismic Structure.

Further Reading Akal T and Berkson JM (eds.) (1986) Ocean SeismoAcoustics. New York and London: Plenum press. Biot MA (1962) Generalized theory of acoustic propagation in porous dissipative media. Journal of Acoustical Society of America 34: 1254--1264.

91

Brekhovskikh LM (1980) Waves in Layered Media. New York: Academic Press. Cagniard L (1962) Reflection and Refraction of Progressive Seismic Waves. New York: McGraw-Hill. Grant FS and West GF (1965) Interpretation and Theory in Applied Geophysics. New York: McGraw Hill. Hampton L (ed.) (1974) Physics of Sound in Marine Sediments. New York and London: Plenum Press. Hovem JM, Richardson MD, and Stoll RD (eds.) (1992) Shear Waves in Marine Sediments. Dordrecht: Kluwer Academic. Jensen FB, Kuperman WA, Porter MB, and Schmidt H (1999) Computational Ocean Acoustics. New York: Springer-Verlag. Kuperman WA and Jensen FB (eds.) (1980) Bottom Interacting Ocean Acoustics. New York and London: Plenum Press. Lara-Saenz A, Ranz-Guerra C and Carbo-Fite C (eds) (1987) Acoustics and Ocean Bottom. II F.A.S.E. Specialized Conference. Inst. De Acustica, Madrid. Pace NG (ed.) (1983) Acoustics and the Sea-Bed. Bath: Bath University Press. Pouliquen E, Lyons AP, Pace NG et al. (2000) Backscattering from unconsolidated sediments above 100 kHz. In: Chevret P and Zakhario ME (ed.) Proceedings of the fifth European Conference on Under water Acoustics, ECUA 2000 Lyon, France. Stoll RD (1989) Lecture Notes in Earth Sciences. New York: Springer-Verlag.


ACOUSTICS, ARCTIC P. N. Mikhalevsky, Science Applications International Corporation, McLean, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 53–61, & 2001, Elsevier Ltd.

Introduction The Arctic Ocean is an isolated mediterranean basin with only limited communication with the world’s oceans, principally the Atlantic Ocean via the Fram Strait and the Barents Sea, and the Pacific Ocean via the Bering Strait. The ubiquitous feature of the Arctic Ocean is the sea ice that covers the entire Arctic basin during the winter months and only retreats off the shallow water shelf areas in the summer months, creating a permanent cap over most of the central Arctic basin (Figure 1). The presence of the yearround sea ice cover determines the unique character of acoustic propagation and ambient noise in the Arctic Ocean. The sea ice nsulates the Arctic Ocean from solar heating in the summer months, creating a yearround upward refracting sound speed profile with the sound speed minimum at the water–ice interface. Sound, therefore, is refracted upward and is continuously reflected from the ice as it propagates, causing attenuation by scattering, mode conversion, and absorption that increases rapidly with frequency. The lack of solar forcing and the Arctic Ocean’s restricted communication with the other oceans of the world creates a very stable acoustic channel with significantly reduced fluctuations of acoustic signals in comparison with the temperate oceans. In contrast to the central basin, acoustic propagation on the Arctic shelves and in the marginal ice zones (MIZs, those areas between the average ice minimum and maximum) (Figure 1), is quite complex and variable owing to the seasonal retreat of the sea ice, river run-off, and bottom interaction (see Acoustics, Shallow Water, Acoustics in Marine Sediments). Over the last half-century Arctic acoustics research and development has largely supported submarine operations. The importance of the Northern Sea Route to the Soviet Union, and the prospect of Soviet nuclear ballistic missile submarines exploiting the unique Arctic acoustic environment to remain undetected provided the need for this research. Since the end of the Cold War and the beginning of concern about ‘global warming’ there has been a new focus for Arctic acoustics on acoustic thermometry

92

and acoustic remote sensing (see Tomography). The Arctic Ocean is the world’s ‘air-conditioner’, maintaining the surface heat balance, and it provides fresh water to the world’s oceans, principally in the form of sea ice discharged from the Fram Strait. The latter regulates convective overturning in the Greenland and Norwegian Seas that in turn drives the global thermohaline circulation with significant impact on climate. Monitoring changes in the temperature and stratification of the Arctic Ocean and sea ice thickness using acoustics is an important capability that will improve our understanding of the Arctic Ocean and its role in global climate change.

History The first large program dedicated to acoustics research in the Arctic Ocean was undertaken by the US Navy Underwater Sound Laboratory, as it was then known, in 1958 in connection with the International Geophysical Year. Also in that year the US nuclear submarine Nautilus (SSN 571) made its historic voyage to the North Pole, marking the beginning of regular submarine operations in the Arctic Ocean. From 1958 to 1975 the USA conducted much of its acoustics research from manned ice islands (thick tabular sections of land-fast ice that occasionally break away from the Ellesmere ice shelf) that remained within the polar pack ice (see Cryosphere: Sea Ice). The most famous of these was Fletcher’s Ice Island, also known as T3, which was discovered in June 1950. T3, originally 14.5 km long, 6.4 km wide and 52 m thick, provided an ideal platform for Arctic research including acoustics. The first experiments in 1958 were aimed at investigating the feasibility of using the RAFOS (Ranging and Fixing of Sound) for submarine navigation in Arctic waters. Explosive signals were deployed from T3 and other ice stations while the USS Skate (SSN 578) attempted to receive the signals for acoustic crossfixing. While RAFOS was never operationally deployed, a significant amount was learned about the upward refracting sound speed profile, propagation, and resulting scattering losses, as well as ambient noise and reverberation. The Soviet Union operated at least two manned year-round ice stations in the Arctic continuously from 1937 through 1992. In 1971 they began POLEX (Polar Experiment), which was an intense set of oceanographic, ice, and atmospheric measurements over the entire Arctic Ocean basin that


ACOUSTICS, ARCTIC

93

Figure 1 Map of the Arctic Ocean showing the average minimum and maximum sea ice extent, as well as the major bathymetric features, and other significant geographic locations.

continued for a decade. From 1971 through 1994 the US Office of Naval Research sponsored or co-sponsored many international acoustic-related research science programs based from seasonal ice camps in the Arctic, staged out of facilities in Alaska, Canada, Greenland, and Svalbard. These included the Arctic Ice Dynamics Experiments (AIDJEX), 1971–76; the FRAM and MIZEX series in the central Arctic basin and marginal ice zones respectively, 1979–87; the Coordinated East Arctic Experiment (CEAREX), 1988–89; the Sea Ice Mechanics Initiative (SIMI), 1993–94; and the Transarctic Acoustic Propagation Experiment (TAP) in 1994. These programs

quantified sea ice properties and ice scattering processes, bottom and surface reverberation, Arctic plate tectonics with seismic reflection and refraction, identified the mechanisms of ice generated ambient noise, discovered the exceptional phase stability of low frequency propagation (B20–30 Hz), and demonstrated the feasibility of basin scale acoustic thermometry in the Arctic Ocean. The US, UK, and Soviet Union (now Russia) have operated submarines in the Arctic Ocean and conducted a significant amount of acoustics research. In particular, the US Navy’s Submarine Ice Exercises (SUBICEXs) were conducted over almost three


94

ACOUSTICS, ARCTIC

decades by the Arctic Submarine Laboratory (ASL) under the leadership of Dr Waldo Lyon, often in conjunction with seasonal ice camps. Dr Lyon was one of the first to investigate propagation in polar stratified waters, beginning shortly after the end of World War II, working with Canadian researchers. From the Naval Electronics Laboratory (NEL) in San Diego, Dr Lyon led the development of upwardlooking ice profiling sonar, and the forward-looking ice avoidance sonar, that proved critical for safe under-ice submarine operations. SUBICEXs were designed to test all aspects of submarine operations including surfacing through ice, and weapons effectiveness (see Under Ice in Further Reading). In 1993 the US Navy supported a dedicated 60 day science cruise in the Arctic with the USS Pargo (SSN 650). This started the Submarine Science Ice Expeditions (SCICEX) that were conducted each year from 1995 to 1999. These multidisciplinary cruises have confirmed the large changes in the Arctic Ocean thermohaline structure, including the warming of the Arctic Intermediate Water (AIW, see below) that was measured acoustically during the TAP experiment in 1994. Other data and modeling have shown strong evidence of decadal Arctic oscillations, but longer observational time series are needed. A trans-basin acoustic section was started in October 1998 from a source located in the Franz Victoria Strait to a receiving array in the Lincoln Sea. This 3-year US/ Russian collaborative effort is expected to be followed by more permanent long-term monitoring of the Arctic Ocean using acoustics.

Arctic Ocean Sound Speed Structure The structure of the typical sound speed profile in the Arctic consists of components that correspond closely to the primary stratified water masses of the Arctic Ocean (Figure 2). The uppermost layer is Polar Water (PW) defined by temperatures o01C and salinities o34.5 ppt and extends from the surface to depths of 100–200 m. There is often a mixed layer of Polar Water 30–60 m thick just below the ice with a nearly constant temperature near the freezing point and nearly constant salinity at around 33.6 ppt. Below the mixed layer the temperature and salinity increase uniformly, attaining 01C and 34.5 ppt respectively, at the base of the Polar Water layer. Arctic Intermediate Water (AIW) – also known as the Atlantic Layer reflecting its origin (and sometimes referred to as Atlantic Intermediate Water) – is defined by temperatures 401C and salinities 434.5 ppt and extends from the base of the Polar Water to a depth of 1000 m. Temperature increases with depth through the AIW layer to a maximum at approximately 200–500 m depth and then decreases with depth to 01C at about 1000 m. AIW enters the Arctic Ocean via the West Spitzbergen Current from the Fram Strait and via the Barents Sea east of Franz Josef Land (see Ocean Currents: Arctic Basin Circulation). The shallower depths (200 m) of the temperature maximum of the AIW occur in the eastern Arctic and approaches 21C. This temperature maximum deepens to 500 m the western Arctic in the Canada Basin and is approximately 0.41C. Salinity increases with depth in the AIW layer

_

Mode functions

Sound speed profile (m s 1)

Ra y trace

1440

0

1460

1480

Polar water

Mode 1 9.3˚

Atlantic Intermediate Water

500 Depth (m)

Mode 2

9.6˚

Sound speed

10.0˚

1000

Mode 3

10.5˚

Temperature

Mode 4

Deep Arctic Water

1500

Salinity

12.8˚

2000 0

20

40 Range (km)

60

80 33.5

34.5

35.5 _ 2

Salinity (ppt)

0

2

Temperature (˚C)

Figure 2 The Arctic sound speed profile shown on the right was computed from the measured temperature and salinity, also shown, from an ice camp in the eastern Arctic Ocean in April 1994. A ray trace for this sound speed profile is plotted for a source at a depth of 100 m. The mode shapes computed at 20 Hz using this profile are shown on the left. The major Arctic water masses (see text) are indicated.


ACOUSTICS, ARCTIC

from 34.5 ppt to 34.9 ppt at about the same depth of the temperature maximum and remains constant at 34.9 ppt below this depth. Deep Arctic Water (DW) is a relatively homogeneous water mass from a depth of 1000 m to the bottom, with temperatures o01C and a nearly constant salinity of 34.9 ppt. The resulting sound speed profile (Figure 2) for these typical deep Arctic basin conditions has a minimum at the ice–water interface of 1435–1440 m s1 and increases with depth to the bottom. In the nearsurface mixed layer of the PW the sound speed gradient is þ 0.016 s1, due entirely to the increase of pressure with depth, as temperature and salinity are constant. Below the mixed layer the sound speed increases rapidly with the increasing temperature and salinity (temperature having the far dominant effect) to the depth of the maximum temperature in the AIW layer with gradients of þ 0.1 s1 or more, reaching a sound speed between 1455 and 1465 m s1. Below the depth of the AIW temperature maximum the sound speed continues to increase, but with reduced gradients of þ 0.01 s1 or less as the temperature decreases, but the pressure increases and the salinity is constant. Below 1000 m in the DW both temperature and salinity are nearly constant and the sound speed continues to increase with the þ 0.016 s1 gradient to the bottom associated with the increasing pressure. In general the Arctic or polar profile can be characterized (and is often approximated) roughly as a bilinear upward refracting profile with the large positive gradient at the near surface creating a strong near-surface duct and a smaller positive gradient below this ‘knee’ to the bottom. This structure is important to understanding the resulting propagation effects described below. There are both regional and temporal variations of this typical sound speed structure associated with the corresponding variations in the water masses described above. The strong positive gradient associated with the upper AIW layer weakens as the ‘knee’ in the sound speed profile (Figure 2) deepens from 200 m to about 500 m as one moves from the eastern Arctic Ocean near Svalbard (where the profile in Figure 2 was measured) into the western Arctic Ocean and the Beaufort and Chukchi Seas. In the Beaufort and north Chukchi Seas the influx of warmer Bering Sea water can create a sound speed maximum just below the mixed layer of the PW at depths between 50 and 80 m, creating a double duct. Because of the year-round ice cover, seasonal, interannual, and recently identified decadal time scales dominate the temporal variations of the sound speed profile. The seasonal and interannual variability is confined almost entirely to the PW layer. The decadal variability is seen in the PW and AIW layers. Mesoscale and internal wave

95

Figure 3 A photograph of the pack ice in the central Arctic Ocean taken at the FRAM II Ice Camp in May 1980. A weathered pressure ridge is shown on the left bounding a relatively flat multi-year ice floe on the right that is typical of the central arctic. The blocks of ice in the foreground are 3–4 m high. (Photo by P. Mikhalevsky.).

dynamics are approximately one order of magnitude less energetic in the Arctic Ocean than in the temperate seas and so have a small influence on the sound speed profile and this results in the exceptional signal stability at low frequencies (discussed below).

Propagation Acoustic propagation in the Arctic is dominated by the repeated reflection of the sound from the sea ice as illustrated in Figure 2. The ray trace shown was computed using the sound speed profile of Figure 2. The morphology of sea ice is quite complicated in general. Figure 3 is a photograph taken in the pack ice at an ice camp in the eastern Arctic. It shows a typical pressure ridge on the left that forms between two ice floes when winds and currents compress the pack. Pressure ridges can extend as much as several tens of meters below the ice, and consist of unconsolidated blocks in newly formed ridges as well as refrozen consolidated structures in older ridges. In general the sea ice in the Arctic consists of relatively smooth floes (the right part of Figure 3) laced with pressure ridges and/or open leads as the pack works between convergent and divergent conditions. Average ice roughness and thickness measurements derived from upward-looking submarine sonars have historically (1958–76) been in the range of 1–5 m rms and 2–4 m respectively.1 Recent analysis of SCICEX data (1993–97), however, has shown that

1 LeSchack LA (1980) Arctic Ocean Sea-ice Statistics Derived from Upward-Looking Sonar Data Recorded during Five Nuclear Submarine Cruises. Technical Report to ONR under Contract N0001476-C-0757/NR 307-374. Maryland, LeSchack Associates, Ltd.


96

ACOUSTICS, ARCTIC

_ 50

20 Hz 50 Hz 100 Hz +200 Hz

_ 60

Transmission loss, dB

_ 70 _ 80 _ 90 _ 100 _ 110 _ 120 _ 130 _ 140 _ 150 10

0

1

10

2

10 Range, km

10

3

Figure 4 The transmission loss plotted as a function of range for 20, 50, 100, and 200 Hz. The source and receiver were located at 60 m depth. The ice was modeled as a rough elastic plate including statistical parameters for the upper and lower side of the ice representative of the ice floe and ridge structure typical of the central Arctic.

since the late 1970s the ice has thinned by as much as 40%,2 which would also imply a reduction in ice roughness as well. The thicker ice and correspondingly greater roughness is typically found in the eastern Arctic and just north of the eastern Canadian Archipelago and Greenland, where the transpolar drift pushes the ice. The acoustic energy is reflected and scattered from the rough ice and converted to both shear waves and compressional waves within the ice, resulting in significant frequency-dependent attenuation loss. Because of this, trans-basin propagation in the Arctic is limited to frequencies typically below 30 Hz or wavelengths exceeding about 50 m. Figure 4 shows the propagation loss plotted as a function of range for several frequencies. The curves in Figure 4 were generated using a model that includes a rough elastic plate representing the ice cover, with a two scale roughness spectrum that models pressure ridges and the smoother intervening floes.3 The model includes the statistics for both the underside and surface of the ice since at lower frequencies as the wavelength becomes much greater than the ice thickness the reflection occurs at the ice–air interface. Submarine ice draft statistics from the eastern Arctic over the

2 Rothrock DA, Yu Y and Maykut GA (1999) Thinning of the arctic sea-ice cover. Geophysical Research Letters 26(23): 3469– 3472. 3 Kudryashov VM (1996) Calculation of the acoustic field in an arctic waveguide. Physical Acoustics. 42: 386–389.

Nansen Basin (Figure 1) were used for Figure 4. As Figure 4 shows the Arctic acoustic waveguide is a low pass filter. So much so in fact that in comparison with propagation in the temperate oceans which scatter from surface waves that have a similar roughness spectrum, at a given range, equivalent propagation loss in the Arctic requires transmitting at a frequency as much a factor of ten less. On the left side of Figure 2 the acoustic mode shapes for modes 1–4 at 20 Hz are plotted for the profile shown. The acoustic modes can be thought of as the interference pattern of upward and downward going acoustic rays (Figure 2) whose turning depths (the depth where the rays are horizontal) correspond with the depth of the deepest peak of the mode amplitude. At 20 Hz, mode 1 consists of those paths that are trapped in the strong near-surface duct created by the thermocline in the transition from the PW to AIW. The higher modes correspond to rays with greater launch angles that turn at successively deeper depths in the Arctic Ocean. Rough surface scattering theory tells us that the loss per bounce or reflection increases with grazing angle and frequency (i.e., the steeper rays, or higher modes have greater loss per bounce). However, even though the lower modes (lower grazing angles) have lower per bounce loss, those that are trapped in the upper duct experience many more interactions with the ice, as clearly seen in Figure 2, and consequently their net loss per kilometer is higher. At shorter ranges for a source and receiver in the upper duct (Figure 4) the propagation is dominated by the lower modes (or equivalently the lower grazing angle rays that are trapped), but these get stripped away as the range increases and at longer ranges (>100–200 km) the higher modes (or equivalently the higher grazing angle rays that are not trapped) dominate. Because the source and receiver are shallow (at 60 m depth) for the model run in Figure 4, there is a less efficient excitation of mode 1 at 20 Hz (i.e., destructive interference with the surface reflected energy) than at the higher frequencies. Thus less of the energy is trapped in the near-surface duct. This results in lower energy being received on the shallow receiver at closer ranges and hence the higher initial propagation loss at 20 Hz. As the frequency increases the mode shapes tend to compress toward the surface (because the wavelength is getting smaller). For a shallow source more of the energy is trapped in this upper duct at higher frequencies (see discussion of frequency dispersion and Figure 6 below). Given the higher per bounce losses at the higher frequencies, and the greater number of bounces experienced by these trapped modes, it can be seen why attenuation in the Arctic increases rapidly with frequency. For


ACOUSTICS, ARCTIC

Time(s)

0

2

5

3

1

10 0

10

20

30

40 50 60 70 Frequency (Hz)

80

90

100

Figure 6 A composite of various measurements of Arctic ambient noise. (1) Central Arctic pack ice Nansen Basin, firstand multi-year floes, 2–3 m thick, with ridging, April 1982. (Reproduced with permission from Dyer, 1984.) (2) Beaufort Sea, summer and fall conditions: (a) highest levels observed, (b) typical cold weather situation following rapid temperature drop and thermal-induced cracking, (c) quietest conditions during warmer stable temperature periods with low winds September– October, 1961 and May–September 1962. (Reproduced with permission from DiNapoli et al., 1978, in Von Winkle, 1984.) (3) Shore-fast ice in the Canadian Archipelago, winter conditions, thermal cracking, February 1963. (Reproduced with permission from Milne AR and Ganton JH (1964) Ambient noise under ArcticSea ice. Journal of the Acoustic Society of America 36(5): 855– 863.) (4) Shore-fast ice in the Canadian Archipelago, spring conditions: (a) noisiest conditions observed during diurnal cooling, (b) quietest conditions observed during diurnal warming, April 1961 (Milne and Ganton 1964). (5) Chukchi Sea, spring conditions, cold stable temperatures, low winds, April 1999 (Mikhalevsky, APLIS Ice Camp).

deeper sources the near-surface duct is not as important and higher grazing angle (higher mode) propagation and loss are dominant at all ranges; however, the higher loss per bounce at higher frequencies still results in more rapid attenuation as the frequency increases. There is a significant body of research on the scattering and reverberation of acoustic energy from the sea ice. The earliest models relied on rough free surface scattering theory with empirical fits to important parameters such as rms ice roughness and average ridge spacing. The free surface scattering models that included ridge-like morphology performed reasonably well at frequencies from 250 Hz to 2000 Hz. At lower frequencies they tended to underpredict the loss. When elastic coupling and scattering into the ice (particularly conversion to shear waves and the attenuation in the ice) are included, better agreement at the low frequencies has been achieved. In particular

97

the ice keels and their spacing are important in the conversion of acoustic compressional waves into flexural modes in the ice. Le Page and Schmidt (see Further Reading) have shown that the intensity of acoustic scattering into the flexural modes of a rough ice plate strongly depends on the width of the roughness spatial spectra. In order to achieve consistent agreement with data the propagation loss models have become more complex, demanding additional input data, particularly about the structure of the sea ice and its constitutive properties. Bottom interaction must also be included, for those paths which cross major features like the Lomonosov Ridge, and propagation on the arctic shelves (Figure 1). The Lomonosov Ridge can rise to 1500 m below the surface. Figure 2 shows that modes 5 and higher (rays 411–121) will start to interact with this feature and will begin to get stripped away for longrange propagation. As the frequency decreases below 20 Hz the modes tend to expand away from the surface and finally all modes will begin to interact with the bottom except in the very deepest parts of the Arctic basin. This leads to a lower frequency bound at 5–10 Hz for efficient long-range propagation. However, for high source levels at these frequencies there is significant backscatter reverberation from these bathymetric features that can be detected and mapped. Excellent correlation with known Arctic bathymetry as well as the discovery of uncharted features at basin scale ranges (1000 km) has been achieved using large explosive sources deployed from the ice. Over the shelves the shallow water propagation dictates bottom and surface/ice interaction at all frequencies; however, there are optimal propagation frequencies similar to deep water that depend upon the depth and boundary properties. The stable upward refracting Arctic sound speed profile causes a very predictable modal and frequency dispersion. As can be seen in Figure 2, successively higher modes are propagating at higher sound speeds corresponding to higher group velocities. As a consequence at a given range, mode 1 arrives last, preceded in order by modes 2, 3, 4, etc. This can be observed by transmitting a wideband waveform such as an impulsive or explosive source and plotting the signal intensity as a function of frequency and time. Figure 5 shows the arrival time from the computed group velocities of modes 1–3 for an explosive shot at a range of approximately 580 km in the Beaufort Sea. Although not plotted, this calculation agreed very closely with the measured data. Note also from Figure 5 that as the frequency increases (as described above), the modes compress towards the surface and first mode 1, then mode 2 and finally mode 3 become trapped in the slower near-surface duct, as evidenced


98

ACOUSTICS, ARCTIC

These frequencies are used for short-range applications including the upward-looking ice profiling sonar, forward-looking ice avoidance sonar, and downward-looking depth sounders and bottommapping sonar. Torpedoes also operate at these frequencies and ice capture can be a problem, but Doppler processing can distinguish moving targets from the stationary ice (see Sonar Systems).

110 100

Spectrum level dB re IPa / Hz

90

1 4a

80

3

2a Knudsen Sea State zero

2b

70

2c 4b

60

5

50 40

Ambient Noise

30 20 10 1

10

2

10 Frequency (Hz)

3

10

10

4

Figure 6 The arrival time of a broadband acoustic pulse plotted as a function of frequency showing the modal dispersion with mode 1 arriving last, preceded by modes 2 and 3 respectively. The absolute travel time can be obtained by adding 396.5 s to the time shown on the vertical axis. This modeled result agreed very closely with the received data. (Adapted with permission from DiNapoli et al., 1978, in Von Winkle, 1984.)

by the longer travel times. This modal dispersion and arrival pattern is a classic Arctic acoustic result. The data upon which Figure 5 is based were taken in 1962; 32 years later in 1994 the TAP experiment measured the same arrival pattern, attesting to the very long-term stable water mass structure in the Arctic Ocean. Observations of the changes in the arrival times of these modes are directly related to the temperature changes of these water masses (since sound speed increases approximately 5 m/s per 11C). In particular, at 20 Hz mode 2 is most sensitive to temperature changes in the AIW, as can be seen from Figure 2. A decrease in the travel time of mode 2 by 2 s was the indicator of the warming in the AIW observed in the 1994 TAP experiment. In addition to the long-term stability of the Arctic sound speed profile, the short-term stability results in exceptional amplitude (B1 dB rms) and phase stability (B0.01 cycles rms) of acoustic signals at low frequencies even over ranges approaching 3000 km for up to 1 h (the longest continuous observations made to date). This applies to essentially fixed terminals. This was first observed in 1980 at 15–30 Hz and 300 km and then again in 1994 and 1999 at 20 Hz and 2700 km. This implies that coherent integration of up to 1 h with optimal gain is achievable (and has been demonstrated). At very high frequencies (more than tens of kHz) reflection occurs at the ice–water interface. Attenuation is rapid, due not only to the severe scattering from the ice, but to volumetric absorption as well.

The ambient noise in the Arctic is highly variable, exhibiting some of the quietest as well as the noisiest ocean noise conditions of all the world’s oceans. A composite of various measurements of Arctic ambient noise is shown in Figure 6. In Figure 6, Knudsen Sea State Zero refers to the ambient noise level in the temperate oceans at sea state zero (the quietest conditions) for comparison (see Acoustic Noise). Icegenerated ambient noise is the dominant mechanism contributing to the general character of ambient noise in the Arctic Ocean from a few tenths of Hz up to 10 000 Hz. Episodic noise is also present in the form of seismic events, such as earthquakes along the Mid Arctic Ridge (Figure 1), biologics (mostly marine mammals in the marginal ice zones), and man-made noise from ice breakers, and seismic exploration. Ice noise is generated when the sea ice deforms, fractures, and breaks in response to environmental forcing such as wind, current, thermal, and internal and surface wave-induced stresses. In general, during stable or warming temperature conditions with low winds the quietest conditions obtain, particularly under shore-fast ice. During periods of rapid cooling ice-fracturing events resulting from thermal-induced tensile stresses can lead to higher noise levels. Higher noise levels also occur in the pack ice when the ice is in motion due to nonthermal forcing such as high winds causing bumping, grinding and rubbing of the ice flows. In the marginal ice zone where the ice concentrations are typically less than in the central Arctic pack, ice concentration and surface gravity wave-induced flexural floe failure are the primary correlates with the ambient noise. The actual noisegenerating mechanics are complicated but fall into at least two general categories. In the low frequency range (5–100 Hz) the important noise-generating mechanism is the unloading motion of the ice immediately following breaking. This has been shown to have a dipole radiation characteristic. The breaking process itself is important at intermediate frequencies (100–2000 Hz) and has an octopole radiation characteristic likely resulting from a slip– dislocation process.


ACOUSTICS, ARCTIC

Noise from earthquakes and other seismic events occur with some regularity within the Arctic Ocean, concentrated along the Mid Arctic Ridge (Figure 1). These events have been recorded on acoustic arrays suspended from the ice within approximately 300 km of the Mid Arctic Ridge. The received frequencies are typically 20 Hz and below. Earthborne acoustic pressure and shear waves emanate from the source and couple through the ocean floor above the epicenter to compressional waterborne waves that propagate vertically and are reflected from the ice canopy into the Arctic sound channel. Most of the energy from these events arrives via this path, but there are weaker precursors associated with crustal propagating compressional and shear waves that couple to the water directly below the receive array. These arrivals are easily identified by their vertical arrival angle at the array as well as their arrival time. Biological noise in the Arctic is concentrated in the marginal ice zones and near the edge of the pack ice. The bowhead is the most numerous of the baleen whales in Arctic waters, migrating along the coast of Alaska in the Chukchi and Beaufort Seas in the spring and then exiting the Arctic waters in the fall. They can vocalize from 25 to 3500 Hz, but their dominant frequencies are between 100 and 400 Hz. Of the toothed whales the beluga and narwhal are the most common, with calls ranging from a few hundred Hz to as high as 20 kHz. Many species of pinnipeds, including hair seals such as the bearded, hooded, harp, ringed, and ribbon seal, and the walrus frequent the marginal ice zones. As a group their calls typically range from a few hundred Hz to 10 kHz. Unlike the temperate oceans where shipping typically dominates the ambient noise spectrum, manmade noise in the Arctic is a small contributor except for specific events that are isolated in time and space. Such events include icebreakers, seismic exploration, and some military and experimental activities. Icebreaker noise peaks in the 50–100 Hz range, but with broadband contributions up to 1000 Hz. Seismic exploration occurs during the summer and fall seasons when the ice extent is minimum (Figure 1), allowing easier access to the shelves. Most of the activity has been confined to the Beaufort Sea off the North Slope of Alaska near Prudhoe Bay.

Conclusions The presence of the year-round ice cap creates the upward refracting sound speed profile in the Arctic Ocean with the sound speed minimum at the ice–water interface. Reflection, scattering, mode

99

conversion, and absorption by the rough elastic sea ice cover causes high attenuation as frequency increases, limiting long-range propagation to very low frequencies. Bathymetric effects are important near the major ocean ridges, basin margins, and on the shelf areas where significant mode coupling can occur. The Arctic sound channel is very stable and predictable in the central Arctic basins and there is a close correspondence of propagating acoustic modes with the major water masses of the Arctic Ocean, especially the important AIW. This latter fact makes the use of acoustic thermometry for monitoring longterm Arctic Ocean temperature change particularly suitable. Ongoing research is exploring ways to relate changes in acoustic travel time and intensity to monitor other important variables in the Arctic Ocean including changes in the PW, DW, and the halocline, and average sea-ice thickness and roughness. The latter measurements, when combined with sea ice extent from satellite remote sensing, could provide an estimate of sea ice mass in the Arctic. The role of the Arctic Ocean in shaping and responding to global climate change is only beginning to be explored. Cost-effective, long-term, year-round synoptic observations in the Arctic Ocean require new measurement strategies. The year-round ice cover in the Arctic prevents the use of satellites for direct ocean observations common in the ice-free oceans. Shore-cabled mooring-based observations using advanced biogeochemical sensors and acoustic sources and hydrophone arrays, as well as instrumented autonomous underwater vehicles (AUVs) and under-ice drifters, represent new approaches for observing the Arctic Ocean. Interestingly the RAFOS concept (using nonexplosive sources) is being evaluated anew as a way to track AUVs and drifters in the Arctic as well as for acoustic communication of data. It is clear that Arctic acoustics will have as large a role to play in this important new endeavor in the future, as it has had in the submarine and military operations of the past.

See also Acoustics, Arctic. Acoustics in Marine Sediments. Acoustic Noise. Acoustics, Shallow Water. Platforms: Autonomous Underwater Vehicles. Bioacoustics. Ice–ocean interaction. Nepheloid Layers. North Atlantic Oscillation (NAO). Satellite Passive-Microwave Measurements of Sea Ice. Sea Ice. Sea Ice: Overview. Seals. Seismic Structure. Tomography. Under-Ice Boundary Layer. Water Types and Water Masses.


100

ACOUSTICS, ARCTIC

Further Reading Dyer I (1984) Song of sea ice and other Arctic Ocean melodies. In: Dyer I and Chryssostomidis C (eds.) Arctic Technology and Policy, pp. 11--37. Washington, DC: Hemisphere Publishing. Dyer I (1993) Source mechanisms of Arctic Ocean ambient noise. In: Kerman BR (ed.) Natural Physical Sources of Underwater Sound, pp. 537--551. Netherlands: Kluwer Academic. Leary WM (1999) Under Ice. College Station: Texas A&M University Press. LePage K and Schmidt H (1994) Modeling of lowfrequency transmission loss in the central Arctic. Journal of the Acoustic Society of America 96(3): 1783--1795. Mikhalevsky PN, Gavrilov AN, and Baggeroer AB (1999) The transarctic acoustic propagation experiment and

climate monitoring in the Arctic. IEEE Journal of Oceanic Engineering 24(2): 183--201. Newton JL (1989) Sound speed structure of the Arctic Ocean including some effects on acoustic propagation. US Navy Journal of Underwater Acoustics 39(4): 363--384. Richardson WJ, Greene CR Jr, Malme CI, and Thomson DH (1995) Marine Mammals and Noise. San Diego: Academic Press. Urick RJ (1975) Principles of Underwater Sound. New York: McGraw-Hill. Von Winkle WA (1984) Naval Underwater Systems Center (NUSC) Scientific and Engineering Studies: Underwater Acoustics in the Arctic. New London: Naval Underwater Systems Center Publisher.


ACOUSTICS, DEEP OCEAN W. A. Kuperman, Scripps Institution of Oceanography, University of California, San Diego, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 61–72, & 2001, Elsevier Ltd.

Introduction The acoustic properties of the ocean, such as the paths along which sound from a localized source will travel, are mainly dependent on its sound speed structure. The sound speed structure is dependent on the oceanographic environment described by variations in temperature, salinity, and density with depth or horizontal position. This article will review the ocean acoustic environment, sound propagation, ambient noise, scattering and reverberation, and the passive and active sonar equation.

Ocean Acoustic Environment Sound propagation in the ocean is governed by the spatial structure of the sound speed and the sound speed in the ocean is a function of temperature, salinity, and ambient pressure. Since the ambient pressure is a function of depth, it is customary to express the sound speed (c) in meters per second as an empirical function of temperature (T) in degrees celsius, salinity (S) in parts per thousand and depth (z) in meters, e.g. eqn [1]. c ¼ 14449:2 þ 4:6T 0:055T þ 0:00029T þ ð1:34 0:01T ÞðS 35Þ þ 0:016z

Units The decibel (dB) denotes a ratio of intensities (see Section 3.3) expressed in terms of a logarithmic (base _1

Sound speed (m s ) 1440 1460 1480 1500 1520 Warmer surface water profile

Surface duct profile

3

Mixed layer profile

½1

There exist more accurate formulas, if needed. Figure 1 shows a typical set of sound speed profiles, indicating greatest variability near the surface. In a warmer season (or warmer part of the day, sometimes referred to as the ‘afternoon effect’), the temperature increases near the surface and hence the sound speed increases toward the sea surface. In nonpolar regions where mixing near the surface due to wind and wave activity is important, a mixed layer of almost constant temperature is often created. In this isothermal layer, sound speed increases with depth because of the increasing ambient pressure, the last term in eqn [1]. This is the surface duct region. Below the mixed layer is the thermocline where the temperature and hence the sound speed decrease

1000

Water depth (m)

2

with depth. Below the thermocline, the temperature is constant and the sound speed increases because of increasing ambient pressure. Therefore, between the deep isothermal region and the mixed layer, there is a minimum sound speed; the depth at which this minimum takes place is referred to as the axis of the deep sound channel. However, in polar regions, the water is coldest near the surface so that the minimum sound speed is at the surface. Figure 2 is a contour display of the sound speed structure of the North and South Atlantic with the deep sound channel axis indicated by the heavy dashed line. Note that the deep sound channel becomes shallower toward the poles. Aside from sound speed effects, the ocean volume is absorbtive and will cause attenuation that increases with acoustic frequency. The ocean surface and bottom also have a strong influence on sound propagation. The ocean surface, though a perfect reflector when flat, causes scattering when its roughness becomes comparable in size with the acoustic wavelength. The ocean bottom, depending on its local structure will scatter and also attenuate the acoustic field.

Polar region profile

Thermocline Deep sound channel axis

2000

Deep isothermal layer

3000

4000

Figure 1 Generic sounds speed profiles.


101

102

ACOUSTICS, DEEP OCEAN

0

3000

1485 1495 1505

REYKJANES RIDGE

MID ATLANTIC RIDGE

4000

1515 1525

60˚N

50˚N

40˚N

30˚N

MID ATLANTIC RIDGE

CAPE VERDE BASIN

CANARY BASIN

5000 6000 70˚N

1475

1500

ANTARCTICA

2000

1510

GREENLAND

Depth (m)

1000

20˚N

10˚N

BRAZIL BASIN

0˚

10˚S

RIO GRANDE RISE

1535

20˚S

30˚S

ARGENTINE BASIN

40˚S

SCOTIA RIDGE EAST SCOTIA WEDDELL RIDGE SEA

50˚S

60˚S

70˚S

80˚S

Latitude Figure 2 Sound speed contours of the North and South Atlantic along 30.501W. Dashed line indicates axis of deep sound channel. (From Northrop and Colborn (1974).)

30

Depth (m)

Transmission loss (dB)

Ocean surface

0

40 50

Su

rfa

Source

Direct 100

ce

-re

fle

cte

dp

ath

path

Receiver range

60 ~r

70 ~r

_4

_2

80 90 100 0

1

2

3

4

5

6

7

8

9

10

Range (km) Figure 3 The insert shows the geometry of the Lloyd mirror effect. The plots show a comparison of Lloyd mirror to spherical spreading. Transmission losses are plotted in decibels corresponding to losses of 10 log r2 and 10 log r4, respectively, as explained in the test.

10) scale. The ratio of two intensities, I1/I2 is 10 log (I1/I2) in dB units. Absolute intensities are expressed using an accepted reference intensity of a plane wave having an rms pressure equal to 105 dyn cm2 or, equivalently, 1 mPa. Transmission loss is a decibel measure of relative intensity, the latter being proportional to the square of the acoustic amplitude.

Sound Propagation Very Short-range Propagation

The pressure amplitude from a point source in free space falls off with range r as r1; this geometric loss is called spherical spreading. Most sources of interest in the deep ocean are nearer the surface than the bottom. Hence, the two main short-range paths are the direct path and the surface-reflected path. When

these two paths interfere, they produce a spatial distribution of sound often referred to as a ‘Lloyd mirror pattern’ as shown in the insert of Figure 3. Basic Long-range Propagation Paths

Figure 4 is a schematic of propagation paths in the ocean resulting from the sound speed profiles (indicated by the dashed line) described above in Figure 1. These paths can be understood from Snell’s law eqn [2], which relates the ray angle y (z), with respect to the horizontal, to the local sound speed c(z) at depth z. cosyðzÞ ¼ constant cðzÞ

½2

The equation requires that the higher the sound speed, the smaller the angle with the horizontal,



Ocean basin

Arctic

Continental Continental shelf margin

Ice

F B C D

A E

A: Arctic B: Surface duct C: Deep sound channel

Ocean bottom

D: Convergence zone E: Bottom bounce F: Shallow water

Figure 4 Schematic representation of sound propagations paths in the ocean.

meaning that sound bends away from regions of high sound speed or, put another way, sound bends toward regions of low sound speed. Therefore, paths A, B, and C are the simplest to explain since they are paths that oscillate about the local sound speed minima. For example, path C depicted by a ray leaving a source near the deep sound channel axis at a small horizontal angle propagates in the deep sound channel. This path, in temperate latitudes where the sound speed minimum is far from the surface, permits propagation over distances of thousands of kilometers. Path D, which is at slightly steeper angles and is usually excited by a near surface source, is convergence zone propagation, a spatially periodic (35–65 km) refocusing phenomenon producing zones of high intensity near the surface due to the upward refracting nature of the deep sound speed profile. Regions between these zones are referred to as shadow regions. Referring back to Figure 1, there may be a depth in the deep isothermal layer at which the sound speed is the same as it is at the surface; this depth is called the critical depth and is the lower limit of the deep sound channel. A positive critical depth specifies that the environment supports longdistance propagation without bottom interaction, whereas a negative critical depth specifies that the ocean bottom is the lower boundary of the deep sound channel. The bottom bounce path E is also a periodic phenomenon but with a shorter cycle distance and shorter propagation distance because of losses when sound is reflected from the ocean bottom. An alternative way of describing paths is by denoting that they are composed of combinations of refraction (R), surface reflection (SR) and bottom reflection (BR) processes. Thus, Figure 4 also represents some of these paths. Note from Snell’s law that refractive paths involve a path turning around at the highest speed in its duct of confinement. Because such ray paths spend much of their propagation in

103

regions of high sound speed, larger launch angle paths with longer path lengths arrive earlier than shorter paths launched at shallower angles. This is just the opposite of boundary-limited propagation such a bottom bounce (or shallow water) in which a reflection occurs before refraction in a high-speed region can take place. Hence, for deep water refractive paths, those paths that penetrate to a deeper depth have a greater group speed (the horizontal speed of energy propagation) than those paths that do not go as deep, with the axial, most horizontal path along the deep sound channel axis having the slowest group speed. For pulse propagation, this axial arrival is the last arrival. Geometric Spreading Loss

The energy per unit time emitted by a sound source flows through a larger area with increasing range. Intensity is the power flux through a unit area, which translates to the energy flow per unit time through a unit area. Hence, the simplest example of geometric loss is spherical spreading for a point source in free space, for which the area increases as 4pr2, where r is the range from the point source. Thus spherical spreading results in an intensity decay proportional to r2. Since intensity is proportional to the square of the pressure amplitude, the fluctuations in pressure p induced by the sound decay as r1. For range-independent ducted propagation, that is, where rays are refracted or reflected back toward the horizontal direction (which is the case for most long-range propagation), there is no loss associated with the vertical dimension. In this case, the spreading surface is the area of a cylinder whose axis is in the vertical direction passing through the source: 2prH where H is the depth of the duct and is constant. Geometric loss in the near-field Lloyd mirror regime requires consideration of interfering beams from direct and surface reflected paths. To summarize, the geometric spreading laws for the pressure field (recall that intensity is proportional to the square of the pressure) are

• • •

Spherical spreading loss: ppr1 Cylindrical spreading loss: ppr1/2 Lloyd mirror loss: ppr2.

Volume Attenuation

Volume attenuation increases with frequency. In Figure 3, the losses associated with path C include only volume attenuation and scattering because this path does not involve boundary interactions. The volume scattering can be biological in origin or can arise from interaction with large internal wave activity in the vicinity of the upper part of the deep


104


sound channel where paths are refracted before they would interact with the surface. Both of these effects are small for low frequencies. This same internal wave region is also on the lower boundary of the surface duct, allowing scattering out of the surface duct and thereby also constituting a loss mechanism for the surface duct. This mechanism also leaks sound into the deep sound channel, a region that without scattering would be a shadow zone for a surface duct source. This type of scattering from internal waves is also thought to be a major source of fluctuation of the sound field. Attenuation is characterized by an exponential decay of the sound field. If A0 is the rms amplitude of the sound field at unit distance from the source, then the attenuation of the sound field causes the amplitude to decay with distance r along the path according to eqn [3], where the unit of a is nepers/ distance (nepers is a unitless quantity). A ¼ A0 expðarÞ

½3

This attenuation coefficient can be expressed in decibels per unit distance by the conversion a0 ¼ 0.686a. The frequency dependence of attenuation can be roughly divided into four regimes as displayed in Figure 5. In region I, leakage out of the sound channel is believed to be the main cause of attenuation. The main mechanisms associated with regions II and III are boric acid and magnesium sulfate chemical relaxation. Region IV is dominated by the shear and bulk viscosity associated with fresh

1000

Region I Leakage

Attenuation, (dB/km)

100

A' B'

Region II Chemical relaxation B (OH)3

30:1

0:11f 2 43f 2 þ a0 dBkm1 ¼ 3:3 103 þ 1 þ f 2 4100 þ f 2 þ 2:98 104 f 2

½4

Bottom Loss

The structure of the ocean bottom affects those acoustic paths that interact with the ocean bottom. This bottom interaction is summarized by bottom reflectivity, the amplitude ratio of reflected and incident plane waves at the ocean-bottom interface as a function of grazing angle, y (see Figure 6A). For a simple bottom that can be represented by a semiinfinite half-space with constant sound speed cb and density rb, the reflectivity is given by eqn [5] with the subscript w denoting water RðyÞ ¼

rb kwz rw kbz ; rb kwz þ rw kbz

½5

The wavenumbers are given by eqn [6]. kiz ¼ ðo=ci Þ sin yi ¼ k sin yi ;

i ¼ w; b

½6

The incident and transmitted grazing angles are related by Snell’s law according to eqn [7] and the incident grazing angle yw is also equal to the angle of the reflected plane wave. cb cos yw ¼ cw cos yb

½7

For this simple water-bottom interface for which we take cb >cw, there exists a critical grazing angle yc below which there is perfect reflection (eqn [8]).

3:1

10

water. A summary of the approximate frequency dependence (f in kHz) of attenuation (in units of dB km1) is given in eqn [4] with the terms sequentially associated with regions I–IV in Figure 5.

A B

cosyc ¼

cw cb

½8

1 _1

10

10:1

_2

10

_3

?

?

10

Region III MgSO4 relaxation

Region IV Shear & volume viscosity

_4

10

10

100 1000 10 K 100K

1M

10 M

Frequency (Hz) Figure 5 Regions of different dominant processes of attenuation of sound in sea water. (From Urick (1979).) The attenuation a is given in dB per 1000 yards.

For a lossy bottom, there is no perfect reflection, as also indicated in a typical reflection curve in Figure 6B. These results are approximately frequency independent. However, for a layered bottom, the reflectivity has a complicated frequency dependence as shown in the example in Figure 6C, where the contours are in decibels. This example shows the simple reflectivity result below 200 Hz and then a more complicated frequency dependence at higher frequencies. It should be pointed out that if the density of the second medium vanishes, the reflectivity reduces to the pressure release case of R(y) ¼ 1.



exp [i (k⊥ .r + k1z z)]

R exp [i (k⊥. r + k1z z)]

z 1c1 2c 2

r

i

R T

(A)

T exp [i (k⊥. r + k 2z z)]

Nonlossy bottom Lossy bottom Phase

0

0 c

0

90

Grazing angle (°)

5

_2

10

_4

_ 1

_3

_ 4_ 5

_ 3

15 20 0.05

_ 4

_43 _

_2

Grazing angle (°)

0

_ 4

(B)

_ 3

Phase (°)

Reflectivity R ()

1

_ _ 32

0.1

(C)

0.2

_3

0.4

_4

0.8

1.6

3.2

Frequency (kHz)

Figure 6 The reflection and transmission processes. Grazing angles are defined relative to the horizontal. (A) A plane wave is incident on an interface separating two media with densities and sound speeds rc. R(y) and T (y) are reflection and transmission coefficients. Snell’s law is a statement that k>, the horizontal component of the wavevector, is the same for all three waves. (B) Rayleigh reflection curve (eqn [5]) as a function of the grazing angle (y in (A)) indicating critical yc. The dashed curve shows that if the second medium is lossy, there is less than perfect reflection below the critical angle. (C) Examples of contour of reflection loss (20 log R) for a layered bottom, showing frequency and grazing angle dependence. The simpler reflectivity curve for each frequency is obtained from a vertical slice.

105

acoustically relevant parameters of sound speed, density, and attenuation. Sound propagation in the ocean is mathematically described by the wave equation whose coefficients and boundary conditions are derived from the ocean environment. There are essentially four types of models (computer solutions to the wave equation) to describe sound propagation in the sea: ray, spectral or fast field program (FFP), normal mode (NM), and the parabolic equation (PE). Ray theory is an asymptotic high-frequency approximation to the wave equation, whereas the latter three models are more or less direct solutions to the wave equations under an assortment of milder restrictions. The highfrequency limit does not include diffraction phenomena. All of these models can handle depth variation of the ocean acoustic environment. A model that also takes into account horizontal variations in the environment (i.e., sloping bottom or spatially variable oceanography) is termed rangedependent. For high frequencies (a few kHz or above), ray theory is the most practical. The other three model types are more applicable and usable at lower frequencies (below 1 kHz). The hierarchy of underwater acoustic models is shown in schematic form in Figure 7. The output of these models is typically propagation loss, which is the intensity relative to a unit source at unit distance, expressed in decibels. Transmission loss is the negative of propagation loss, and hence, a positive quantity. An example of the output of propagation models is shown in Figure 8, indicating agreement between the models. However, we also see a difference among the models in that ray theory predicts a sharper shadow zone than the wave theory model (i.e., the 10–30 km region in Figure 8B); this is an expected result from the infinite-frequency ray approximation.

Scattering and Reverberation Propagation Models

An ocean acoustic environment is often very complex, with range- and depth-dependent properties. Such an environment does not in general lend itself to simple analytic predictions of sound propagation. Even in range-independent environments there are many paths (multipaths) and these paths combine to form a complex interference pattern. For example, the convergence zones are an example of a more complex structure that cannot be described by a monotonic geometric spreading law. Acoustic models play an important role in predicting sound propagation; the inputs to these models are oceanographic quantities ultimately translated into the

Scattering caused by rough boundaries or volume inhomogeneities is a mechanism for loss (attenuation), reverberant interference, and fluctuation. Attenuation from volume scattering was addressed above. In most cases, it is the mean or coherent (or specular) part of the acoustic field that is of interest for a sonar or communications application, and scattering causes part of the acoustic field to be randomized. Rough surface scattering out of the ‘specular direction’ can be thought of as an attenuation of the mean acoustic field and typically increases with increasing frequency. A formula often used to describe reflectivity from a rough boundary is eqn [9], where R(y) is the reflection coefficient of the


106


Nonlinear wave equation

Linear wave equation

TDFFP FFP

NPE

TDPE

Frequency-domain wave equation

Normal modes

WKBJ approximation

Coupled modes

Adiabatic modes

Asymptotic

Rangeindependent

Ray theory

Rangedependent

PE

Asymptotic

Spectral Figure 7 Heirarchy of underwater acoustic models. TD refers to time domain. NPE is the nonlinear parabolic equation that describes high-amplitude (e.g., shockwave) propagation. The arrows are directed toward the flow of derivation of the model.

G2 2

½9

The scattered field is often referred to as reverberation. Surface, bottom, or volume scattering strength, SS,B,V is a simple parametrization of the production of reverberation and is defined as the ratio in decibels of the sound scattered by a unit surface area or volume referenced to a unit distance, (Insert equqtion), to the incident plane wave intensity, Iinc (eqn [10]). SS;B;V

Iscat ¼ 10 log Iinc

½10

The Chapman–Harris curves predicts the ocean surface scattering strength in the 400–6400 Hz region; eqn [11], where y is the grazing angle in degrees, w the wind speed in m s1 and f is the frequency in Hz.

Depth (m)

R0 ðyÞ ¼ RðyÞexp

1520 _ 1510 1530 m s 1

y 42:4 log b þ 2:6; 30 0:58 b ¼ 107 wf 1=3

½11

A more elaborate formula exists that is more accurate for lower wind speeds and lower frequencies. The simple characterization of bottom backscattering strength utilizes Lambert’s rule for diffuse scattering, given by eqn [12] where the first term is determined empirically. SB ¼ A þ 10 log sin2 y

½12

20

10

30

Range (km)

(A) 60 70 80

60 70 80 90 100

PE (250 Hz) Ray (2000 Hz) Experiment

110 120 0

(B)

SS ¼ 3:3b log

Source = 25 m, Receiver = 38 m

0 200 400 600 800 1000 0

Propagation loss (dB)

smooth interface and G is the Rayleigh roughness parameter defined as GR2ks sin y where k ¼ 2p/l, l is the acoustic wavelength, and s is the rms roughness (height).

10

20

30

Range (km)

Figure 8 Model and data comparison for a range-dependent deep water case. (A) Sound speed profiles as a function of range together with a ray trace showing the breakdown of surface duct propagation. (B) Parabolic equation comparison with data at 250 Hz and ray theory comparison with data at 2000 Hz.

Under the assumption that all incident energy is scattered into the water column with no transmission in to the bottom, A is 5 dB. Typical realistic values for A that have been measured are 17 dB for big Basalt Mid-Atlantic Ridge cliffs and 27 dB for sediment ponds.



Volume scattering strength is typically reduced to a surface scattering strength by taking (Insert equqtion) as an average volume scattering strength within some layer at a particular depth; then the corresponding surface scattering strength is given by eqn [13], where H is the layer thickness. SS ¼ Sv þ 10 log H

½13

The column or integrated scattering strength is defined as the case for which H is the total water depth. Volume scattering usually decreases with depth (about 5 dB per 300 m) with the exception of the deep scattering layer. For frequencies less than 10 kHz, fish with air-filled swimbladders are the main scatterers. Above 20 kHz, zooplankton or smaller animals that feed upon phytoplankton and the associated biological chain are the scatterers. The deep scattering layer (DSL) is deeper in the day than in the night, changing most rapidly during sunset and sunrise. This layer produces a strong scattering increase of 5–15 dB within 100 m of the surface at night and virtually no scattering in the daytime at the surface since it migrates down to hundreds of meters. Since higher pressure compresses the fish swimbladder, the backscattering acoustic resonance tend to be at a higher frequency during the day when the DSL migrates to greater depths. Examples of day and night scattering strengths are shown in Figure 9.

107

Finally, near-surface bubbles and bubble clouds can be thought of as either volume or surface scattering mechanisms acting in concert with the rough surface. Bubbles have resonances (typically greater than 10 kHz) and at these resonances, scattering is strongly enhanced. Bubble clouds have collective properties; among these properties is that a bubbly mixture, as specified by its void fraction (total bubble gas volume divided by water volume) has a considerable lower sound speed than water.

Ambient Noise There are essentially two types of ocean acoustic noise: man-made and natural. Generally, shipping is the most important source of man-made noise, though noise from offshore oil rigs is becoming more and more prevalent. Typically, natural noise dominates at low frequencies (below 10 Hz) and high frequencies (above a few hundred hertz). Shipping fills in the region between ten and a few hundred hertz. A summary of the spectrum of noise is shown in Figure 10. The higher-frequency noise is usually parametrized according to sea state (also Beaufort number) and/or wind. Table 1 summarizes the description of sea state. The sound speed profile affects the vertical and angular distribution of noise in the deep ocean. When there is a positive critical depth, sound from

≤100 m

≤ 850 m

_ 50

≤ 780 m

≤ 340 m

≤ 580 m

d

10 log ∫0 MZ d Z

_ 45

_ 55

≤ 140 m

_ 60

_ 65

Day

2

5

10

Night

15 2

5

10

15

Frequency (kHz) (A)

(B)

Figure 9 Day and night scattering strength measurements using an explosive source as a function of frequency. The spectra measured at various times after the explosion are labeled with the depth of the nearest scatterer that could have contributed to the reverberation. The ordinate corresponds to Sv in eqn [13]. (From Chapman and Marchall (1966))


108


Intermittent and local effects Earthquakes and explosions

146

Biologics Precipitation Ships, industrial activity Sea ice

KEY 126

Limits of prevailing noise Wind-dependent bubble and spray noise Low-frequency very-shallow-water wind dependence

Sound pressure spectrum level (dB re 1 Pa)

Heavy precipitation Heavy traffic noise Usual traffic noise _ deep water

106

Usual traffic noise _ shallow water Thermal noise General pattern of noise from earthquakes and explosions Extrapolations

86

Wind force (Beaufort)

66 8 5

46 3 2

Prevailing noises Turbulent-pressure fluctuations

26

1 Oceanic traffic Bubbles and spray (surface agitation)

Surface waves _ second-order pressure effects (seismic background)

Molecular Agitation

6 1

10

10

2

10

3

4

10

5

10

Frequency (Hz) Figure 10 Composite of ambient noise spectra. (From Wenz (1962).)

surface sources can travel long distances without interacting with the ocean bottom, but a receiver below this critical depth should sense less surface noise because propagation involves interaction with lossy boundaries, surface and/or bottom. This is illustrated in Figure 11, which shows a deep water environment with measured ambient noise. Figure 12 is an example of vertical directivity of noise that also follows the propagation physics discussed above. The shallower depth is at the axis of the deep sound channel, while the other is at the critical depth. The pattern is narrower at the critical depth where the sound paths tend to be horizontal since the rays are turning around at the lower boundary of the deep sound channel.

In a range-independent ocean, Snell’s law predicts a horizontal noise notch at depths where the speed of sound is less than the near-surface sound speed. Returning to eqn [2] and reading off the sound speeds from Figure 11 at the surface (c ¼ 1530 m s1) and say, 300 m (1500 m s1), a horizontal ray (y ¼ 0) launched from ocean surface would have an angle with respect to the horizontal of about 111 at 300 m depth. All other rays would arrive with greater vertical angles. Hence we expect this horizontal notch. However, the horizontal notch is often not seen at shipping noise frequencies. This is because shipping tends to be concentrated in continental shelf regions; range-dependent propagation couples such noise sources to the deep ocean. Thus, for example,



Table 1

109

Descriptions of the ocean sea surface. Approximate relation between scales of wind speed, wave height, and sea state

Sea criteria

Wind speed

12-h wind Fully risen sea

Beaufort Range scale (m s1)

Mirrorlike Ripples Small wavelets Large wavelets, scattered whitecaps Small waves, frequent whitecaps Moderate waves, many whitecaps Large waves, whitecaps everywhere, spray Heaped-up sea, blown spray, streaks Moderately high, longwaves, spindrift

0 1 2 3 4 5 6 7 8

o0.5 0.5–1.7 1.8–3.3 3.4–5.4 5.5–8.4 8.5–11.1 11.2–14.1 14.2–17.2 17.3–20.8

Mean Wave (m s1) height a, (m)

1.1 2.5 4.4 6.9 9.8 12.6 15.7 19.0

b

Wave height a, (m)

b

Durationb, (h)

c

Fetchb, (km)

c

Seastate scale

0 1/2 1 o19 2 19–74 3 74–185 4 185–370 5 370–740 6 740–1300 7

o0.30 o0.30 0.30–0.61 0.30–0.61 o2.5 0.61–1.5 0.61–1.8 2.5–6.5 1.5–2.4 1.8–3.0 6.5–11 2.4–3.7 3.0–5.2 11–18 3.7–5.2 5.2–7.9 18–29 5.2–7.3 7.9–11.9 29–42

a

The average height of the highest one-third of the waves (significant wave height). Estimated from data given in US Hydrographic Office (Washington, DC) publications HO 604 (1951) and HO 603 (1955). c The minimum fetch and duration of the wind needed to generate a fully risen sea. From Wenz (1962). b

50 Hz

25 Hz

100 Hz

150 Hz

1000

250 Hz

500 Hz

18 Hz

0

Depth (m)

2000

3000

4000

CRITICAL DEPTH = 4420 m

5000 BOTTOM 5322 m BOTTOM 5322 m

6000 1480

1500

1520

1540

50

_1

Sound velocity (m s ) (B)

(A)

60

70

80

90

Spectrum level (dB relative to 1 μPa)

Figure 11 (A) Sound speed profile and (B) noise level as a function of depth in the Pacific. (From Morris (1978))

propagation down a continental slope converts highangle rays to lower angles at each bounce. There are also deep sound channel shoaling effects that result in the same trend in angle conversion.

methodology of the sonar equation is analogous to an accounting procedure involving acoustic signal, interference, and system characteristics. It is instructive, beyond the specific application to conventional sonars, to understand this accounting methodology and below is a simplified summary.

Sonar Equation A major application of underwater acoustics is sonar technology. The performance of a sonar is often described simply in terms of the sonar equation. The

Passive Sonar Equation

A passive sonar system uses the radiated sound from a target to detect and locate the target. A radiating


110


object of source level SL (all units are in decibels) is received at a hydrophone of a sonar system at a lower signal level S because of the transmission loss ‘TL’ it suffers (e.g., cylindrical spreading plus attenuation or a TL computed from one of the propagation models) (eqn [14]). S ¼ SL TL

½14

The noise, N, at a single hydrophone is subtracted from eqn [14] to obtain the signal-to-noise ratio (SNR) at a single hydrophone (eqn [15]). SNR ¼ SL TL N

½15

Typically a sonar system consists of an array or antenna of hydrophones that provides signal-to-noise enhancement through a beamforming process; this process is quantified in decibels by array gain AG that is therefore added to the single hydrophone SNR to give the SNR at the output of the beamformer (eqn [16]). SNRBF ¼ SL TL N þ AG

½16

Because detection involves additional factors including sonar operator ability, it is necessary to specify a detection threshold (DT) level above the SNRBFat which there is a 50% (by convention) probability of detection. The difference between these two quantities is called signal excess (SE) (eqn [17]). SE ¼ SL TL N þ AG DT

½17

This decibel bookkeeping leads to an important sonar engineering descriptor called the figure of merit, FOM, which is the transmission loss that gives a zero signal excess (eqn [18]). FOM ¼ SL N þ AG DT

½18

Spectrum level (dB / µPa / steradian)

The FOM encompasses the various parameters a sonar engineer must deal with: expected source level, the noise environment, array gain and the detection

threshold. Conversely, since the FOM is a transmission loss, one can use the output of a propagation model (or, if appropriate, a simple geometric loss plus attenuation) to estimate the minimum range at which a 50% probability of detection can be expected. This range changes with oceanographic conditions and is often referred to as the ‘range of the day’ in navy sonar applications. Active Sonar Equation

A monostatic active sonar transmits a pulse to a target and its echo is detected at a receiver co-located with the transmitter. A bistatic active sonar has the receiver in a different location from the transmitter. The main differences between the passive and active cases is that the source level is replaced by a target strength, TS; reverberation and hence reverberation level, RL, is usually the dominant source of interference as opposed noise; and the transmission loss is over two paths: transmitter to target and target to receiver. In the monostatic case the transmission loss is 2TL, where TL is the one-way transmission loss; and in the bistatic case the transmission loss is the sum (in dB) over paths from the transmitter to the target and the target to the receiver, TL1 þ TL2. The concept of the detection threshold is useful for both passive and active sonars. Hence, for signal excess, we have eqn [19]. S ¼ SL TL1 þ TS TL2 ðRL þ N Þ þ AG DT

½19

The corresponding FOM for an active system is defined for the maximum allowable two-way transmission loss with TS ¼ 0 dB.

See also Acoustic Noise. Acoustic Scattering by Marine Organisms. Acoustics, Arctic. Acoustics, Shallow Water. Acoustics in Marine Sediments. Bioacoustics. Bubbles. Deep-Sea Drilling Methodology. Seismic Structure. Seismology Sensors. Sonar Systems. Tomography.

80 713 m 3781 m 70 _ 20 DOWN

_ 10

0

10

Elevation angle (°)

20 UP

Figure 12 The vertical directionality of noise at the axis of the deep sound channel and at the critical depth in the Pacific. (From Anderson (1979))

Further Reading Anderson VC (1979) Variations of the vertical directivity of noise with depth in the North Pacific. Journal of the Acoustical Society of America 66: 1446--1452. Brekhovskikh LM and Lysanov YP (1991) Fundamentals of Ocean Acoustics. Berlin: Springer-Verlag. Chapman RP and Harris HH (1962) Surface backscattering strengths measured with explosive sound sources.



Journal of the Acoustical Society of America 34: 1592-1597. Chapman RP and Marchall JR (1966) Reverberation from deep scattering layers in the Western North Atlantic. Journal of the Acoustical Society of America 40: 405--411. Collins MD and Siegmann WL (2001) Parabolic Wave Equations with Applications. New York: Springer-AIP. Dushaw BD, Worcester PF, Cornuelle BD, and Howe BM (1993) On equations for the speed of sound in seawater. Journal of the Acoustical Society of America 93: 255--275. Jensen FB, Kuperman WA, Porter MB, and Schmidt H (1994) Computational Ocean Acoustics. Woodbury: AIP Press. Keller and Papadakis JS (eds.) (1977) Wave Propagation in Underwater Acoustics. New York: Springer-Verlag. Makris NC, Chia SC, and Fialkowski LT (1999) The biazimuthal scattering distribution of an abyssal hill. Journal of the Acoustical Society of America 106: 2491--2512. Medwin H and Clay CS (1997) Fundamentals of Acoustical Oceanography. Boston: Academic Press. Morris GB (1978) Depth dependence of ambient noise in the Northeastern Pacific Ocean. Journal of the Acoustical Society of America 64: 581--590.

111

Munk W, Worcester P, and Wunsch C (1995) Acoustic Tomography. Cambridge: Cambridge University Press. Nicholas M, Ogden PM, and Erskine FT (1998) Improved empirical descriptions for acoustic surface backscatter in the ocean. IEEE-JOE 23: 81--95. Northrup J and Colborn JG (1974) Sofar channel axial sound speed and depth in the Atlantic Ocean. Journal of Geophysical Research 79: 5633--5641. Ross D (1976) Mechanics of Underwater Noise. New York: Pergamon. Ogilvy JA (1987) Wave scattering from rough surfaces. Reports on Progress in Physics 50: 1553--1608. Urick RJ (1979) Sound Propagation in the Sea. Washington, DC: US GPO. Urick RJ (1983) Principles of Underwater Sound. New York: McGraw Hill. Spiesberger JL and Metzger K (1991) A new algorithm for sound speed in seawater. Journal of the Acoustical Society of America 89: 2677--2688. Wenz GM (1962) Acoustics ambient noise in the ocean: spectra and sources. Journal of the Acoustical Society of America 34: 1936--1956.


ACOUSTICS, SHALLOW WATER F. B. Jensen, SACLANT Undersea Research Centre, La Spezia, Italy Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 89–96, & 2001, Elsevier Ltd.

away from the bottom and therefore can propagate to long ranges with little attenuation. Moreover, the environmental variability is much higher in coastal regions than in the deep ocean, with the result that there is much more acoustic variability in shallow water than in deep water.

Introduction

The Ocean Acoustic Environment

Using the commonly accepted definition of shallow water to mean coastal waters with depth up to 200 m, the shallow-water regions of the world constitute around 8% of all oceans and seas. These regions are particularly important since they are national economic zones and also more assessible. Sound waves in the sea play the role of light in the atmosphere, i.e. acoustics is the only means of ‘seeing’ objects at distances beyond a few hundred meters in seawater. All forms of electromagnetic waves (light, radar) are rapidly attenuated in seawater. Lowfrequency acoustic signals, on the other hand, propagate with little attenuation and can be heard over thousands of kilometers in the deep ocean. The use of sound in the sea is ubiquitous. It is employed by the military to detect mines and submarines, and ship-mounted sonars measure water depth, ship speed and the presence of fish shoals. Side-scan systems are used to map bottom topography, sub-bottom profilers for getting information about the deeper layering, and other sonar systems for locating pipelines and cables on and beneath the seafloor. Sound is also used for navigating submerged vehicles, for underwater communications and for tracking marine mammals. In an inverse sense sound is used for measuring physical parameters of the ocean environment and for monitoring oceanic processes through the techniques of acoustical oceanography and ocean acoustic tomography. Optimal sonar design for this great variety of applications demands using a wide range of acoustic frequencies. Practical shallow-water systems cover a frequency range from 50 Hz to 500 kHz, which, with a mean sound speed of 1500 m s1, correspond to acoustic wavelengths from 30 m down to 3 mm. The principal characteristic of shallow-water propagation is that the sound-speed profile is nearly constant over depth or downward refracting, meaning that long-range propagation takes place exclusively via lossy bottom-interacting paths. This is very different from deep-water scenarios, where the sound-speed structure is such that sound is refracted

The ocean is an acoustic waveguide limited above by the sea surface and below by the seafloor. The speed of sound in the waveguide plays the same role as the index of refraction does in optics. Sound speed is normally related to density and compressibility. In the ocean, density is related to static pressure, salinity and temperature. The sound speed in the ocean is an increasing function of temperature, salinity, and pressure, the latter being a function of depth. It is customary to express sound speed (c) as an empirical function of three independent variables: temperature (T) in degrees centigrade, salinity (S) in parts per thousand (%), and depth (D) in meters. A simplified expression for this dependence is

112

c ¼ 1449:2 þ 4:6T 0:055T 2 þ 0:00029T 3 ½1 þð1:34 0:010T ÞðS 35Þ þ 0:016D In shallow water, where the depth effect on sound speed is small, the primary contributor to sound speed variations is the temperature. Thus, for a salinity of 35%, the sound speed in seawater varies between 1450 m s1 at 01C and 1545 m s1 at 301C. Seasonal and diurnal changes affect the oceanographic parameters in the upper ocean. In addition, all of these parameters are a function of geography. In a warmer season (or warmer part of the day) in shallow seas where tidal mixing is weak, the temperature increases near the surface and hence the sound speed increases toward the sea surface. This near-surface heating (and subsequent cooling) has a profound effect on surface-ship sonars. Thus the diurnal heating causes poorer sonar performance in the afternoon – a phenomenon known as the afternoon effect. The seasonal variability, however, is much greater and therefore more important acoustically. A ray picture of propagation in a 100-m deep shallow water duct is shown in Figure 1. The soundspeed profile in the upper panel is typical of the Mediterranean in the summer. There is a warm surface layer causing downward refraction and hence repeated bottom interaction for all ray paths. Since


ACOUSTICS, SHALLOW WATER

SV

SEA SURFACE

Depth

0 2.5˚ slope

Depth (m)

100

(A)

Range SV

113

200

Depth

300

400 0

Range

(B)

1

2

3

Range (km)

Figure 1 Ray paths in shallow water for typical Mediterranean summer and winter profiles. (A) In summer sound interacts repeatedly with the seabed but not with the sea surface. (B) In winter sound interacts with both the sea surface and the seabed, except for shallow rays emitted near the horizontal.

the seafloor is a lossy boundary, propagation in shallow water is dominated by bottom reflection loss at low and intermediate frequencies (o1 kHz) and scattering losses at high frequencies. The seasonal variation in sound-speed structure is significant with winter conditions being nearly iso-speed (Figure 1B). The result is that there is less bottom interaction in winter than in summer, which again means that propagation conditions are generally better in winter than in summer. Of course, the ocean sound-speed structure is neither frozen in time nor space. On the contrary, the ocean has its own weather system. There are currents, internal waves and thermal microstructure present in most shallow-water areas. Figure 2 illustrates the sound speed variability along a 15 km-long track in the Mediterranean Sea. The data were recorded on a towed thermistor chain covering depths between 5 and 90 m. In general, this type of timevarying oceanographic structure has an effect on sound propagation, both as a source of attenuation (acoustic energy being scattered into steeperangle SV

Depth

Summer

Figure 3 Seismic profile of bottom layering in coastal-water area of the Mediterranean.

propagation paths suffers increased bottom reflection loss) and of acoustic signal fluctuations with time. Turning to the upper and lower boundaries of the ocean waveguide, the sea surface is a simple horizontal boundary and a nearly perfect reflector. The seafloor, on the other hand, is a lossy boundary with varying topography. Both boundaries have smallscale roughness associated with them which causes scattering and hence attenuation of sound due to the increased bottom reflection loss associated with steep-angle propagation paths. In terms of propagation physics, the seafloor is definitely the most complex boundary, exhibiting vastly different reflectivity characteristics in different geographical locations. The structure of the ocean bottom in shallow water generally consists of a thin stratification of sediments overlying the continental crust. The nature of the stratification is dependent on many factors, including geological age and local geological activity. Thus, relatively recent sediments will be characterized by plane stratification parallel to the sea bed, whereas older sediments and sediments close to the crustal plate boundaries may have undergone significant deformation. An example of a complicated bottom layering is given in Figure 3, which displays a seismic section from the coastal Mediterranean. The upper stratification here is almost parallel to the seafloor, whereas deeper layers are strongly inclined.

Transmission Loss Range

Figure 2 Spatial variability of sound speed in shallow-water area of the Mediterranean. The depth covered is around 100 m and the range 15 km.

The decibel (dB) is the dominant unit in ocean acoustics and denotes a ratio of intensities (not pressures) expressed on a log10 scale. An acoustic signal traveling through the ocean becomes distorted due to multipath effects and


114


weakened due to various loss mechanisms. The standard measure in underwater acoustics of the change in signal strength with range is transmission loss defined as the ratio in decibels between the acoustic intensity I(r,z) at a field point and the intensity I0 at 1 m distance from the source, i.e. Iðr; zÞ I0 jpðr; zÞj ½dB re 1 m: ¼ 20 log j p0 j

TL ¼ 10 log

½2

Here use has been made of the fact that the intensity of a plane wave is proportional to the square of the pressure amplitude. The major contributors to transmission loss in shallow water are: geometrical spreading loss, water volume attenuation, bottom reflection loss, and various scattering losses.

The cylindrical spreading loss is therefore given by TL ¼ 10 log r

½dB re 1 m

½4

Note that for a point source in a waveguide, there is spherical spreading in the nearfield (rrD) followed by a transition region toward cylindrical spreading which applies only at longer ranges (rbD). As an example consider propagation in a shallowwater waveguide to a range of 20 km with spherical spreading applying on the first 100 m. The total propagation loss (neglecting attenuation) then becomes: 40 dB þ 23 dB ¼ 63 dB. This figure represents the minimum loss to be expected at 20 km. In practice, the total loss will be higher due both to the attenuation of sound in seawater, and to various reflection and scattering losses. Sound Attenuation in Seawater

Geometrical Spreading

The spreading loss is simply a measure of the signal weakening as it propagates outward from the source. Figure 4 shows the two geometries of importance in underwater acoustics. First consider a point source in an unbounded homogeneous medium (Figure 4A). For this simple case the power radiated by the source is equally distributed over the surface area of a sphere surrounding the source. If the medium is assumed to be lossless, the intensity is inversely to proportional the surface of the sphere, i.e. Ip1= 4pR2 . Then from eqn [2] the spherical spreading loss is given by TL ¼ 20 log r

½dB re 1 m

½3

where r is the horizontal range in meters. When the medium has plane upper and lower boundaries as in the waveguide case in Figure 4B, the farfield intensity change with horizontal range becomes inversely proportional to the surface of a cylinder of radius R and depth D, i.e. Ip1=ð2pRDÞ.

*

R

D

*

R

When sound propagates in the ocean, part of the acoustic energy is continuously absorbed, i.e. the energy is transformed into heat. Moreover, sound is scattered by different kinds of inhomogeneities, also resulting in a decay of sound intensity with range. As a rule, it is not possible in real ocean experiments to distinguish between absorption and scattering effects; they both contribute to sound attenuation in seawater. A simplified expression for the frequency dependence (f in kHz) of the attenuation is given by Thorp’s formula, a¼

0:11f 2 44f 2 þ ½dB km1 ; 1 þ f 2 4100 þ f 2

½5

where the two terms describe absorption due to chemical relaxations of boric acid, B(OH)3, and magnesium sulphate, MgSO4, respectively. According to eqn [5] the attenuation of low-frequency sound in seawater is indeed very small. For instance, at 100 Hz a tenfold reduction in sound intensity ( 10 dB) occurs over a distance of around 8300 km. Even though attenuation increases with frequency (r10 dB C150 km at 1 kHz and C9 km at 10 kHz), no other kind of radiation can compete with sound waves for long-range propagation in the ocean. Bottom Reflection Loss

I∝ (A)

1 I∝ 2πRD

1 4πR 2 (B)

Figure 4 Geometrical spreading laws. (A) Spherical spreading; (B) Cylindrical spreading.

Reflectivity, the ratio of the amplitudes of a reflected plane wave to a plane wave incident on an interface separating two media, is an important measure of the effect of the bottom on sound propagation. Ocean bottom sediments are often modeled as fluids which



rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi ðc1 =c2 Þ2 cos2 y1 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Rðy1 Þ ¼ ffi 2 2 ðr2 =r1 Þsin y1 þ ðc1 =c2 Þ cos y1 ðr2 =r1 Þsin y1

½6

where y1 denotes the grazing angle of the incident plane wave of unit amplitude. The reflection coefficient has unit magnitude, meaning perfect reflection, when the numerator and denominator of eqn [6] are complex conjugates. This can only occur when the square root is purely imaginary, i.e. for cos y14c1/c2 (total internal reflection). The associated critical grazing angle below which there is perfect reflection is found to be c1 yc ¼ arccos c2

½7

Note that a critical angle only exists when the sound speed of the second medium is higher than that of the first. A closer look at eqn [6] shows that the reflection coefficient for lossless media is real for yc4yc, which means that there is loss ðjRj ¼ 1Þ but no phase shift associated with the reflection process. On the other hand, for ycoyc we have perfect reflection ðjRj ¼ 1Þ but with an angle-dependent phase shift. In the general case of lossy media (ci complex), the reflection coefficient is complex, and, consequently, there is both a loss and a phase shift associated with each reflection. The critical-angle concept is very important for understanding the waveguide nature of shallowwater propagation. Figure 5 shows bottom loss curves ðBL ¼ 10 log jRj2 Þ for a few simple fluid bottoms with different compressional wave speeds (Cp), densities and attenuations. Note that for a lossy bottom we never get perfect reflection. However, there is in all cases an apparent critical angle (yCC331 for cp ¼ 1800 m s1 in Figure 5), below which the reflection loss is much smaller than for supercritical incidence. With paths involving many bottom bounces such as in shallow-water propagation, bottom losses even as small as a few tenths of a decibel per bounce accumulate to significant total losses since the propagation path may involve many tens or even hundreds of bounces.

10

Bottom loss (dB)

means that they support only one type of sound wave – a compressional wave. The expression for reflectivity at an interface separating two homogeneous fluid media with density ri and sound speed ci, i ¼ 1, 2, was first worked out by Rayleigh as

115

Cp = 1550 m s−1 1600 1800

5

θc

0 0

60 30 Grazing angle (°)

90

Figure 5 Bottom reflection loss curves for different bottom types. Note that low-speed bottoms (clay, silt) are more lossy than high-speed bottoms (sand, gravel). Cp is the compressional wave speed.

Real ocean bottoms are complex layered structures of spatially varying material composition. A geoacoustic model is defined as a model of the real seafloor with emphasis on measured, extrapolated, and predicted values of those material properties important for the modeling of sound transmission. In general, a geoacoustic model details the true thicknesses and properties of sediment and rock layers within the seabed to a depth termed the effective acoustic penetration depth. Thus, at high frequencies (41 kHz), details of the bottom composition are required only in the upper few meters of sediment, whereas at low frequencies (o100 Hz) information must be provided on the whole sediment column and on properties of the underlying rocks. The information required for a complete geoacoustic model should include the following depthdependent material properties: the compressional wave speed, Cp; the shear wave speed, Cp; the compressional wave attenuation, Cp; the shear wave attenuation, Cp; and the density, Cp. Moreover, information on the variation of all of these parameters with geographical position is required. The amount of literature dealing with acoustic properties of seafloor materials is vast. Table 1 lists the geoacoustic properties of some typical seafloor materials, as an indication of the many different types of materials encountered just in continental shelf and slope environments. Boundary and Volume Scattering Losses

Scattering is a mechanism for loss, interference and fluctuation. A rough sea surface or seafloor causes


116


Table 1

Geoacoustic properties of continental shelf environments

Bottom type

p (%)

rb/rw –

Cp/Cw –

Cp (m s1)

Cs (m s1)

ap (aBl1 p )

as (aBl1 s )

Clay Silt Sand Gravel Moraine Chalk Limestone Basalt

70 55 45 35 25 – – –

1.5 1.7 1.9 2.0 2.1 2.2 2.4 2.7

1.00 1.05 1.1 1.2 1.3 1.6 2.0 3.5

1500 1575 1650 1800 1950 2400 3000 5250

o100 Csa Csb Csc 600 1000 1500 2500

0.2 1.0 0.8 0.6 0.4 0.2 0.1 0.1

1.0 1.5 2.5 1.5 1.0 0.5 0.2 0.2

a

Cs ¼ 80 z˜0.3 cs ¼ 110 z˜0.3 c cs ¼ 180 z˜0.3 Cw ¼ 1500 ms 1, rw ¼ 1000 kg m 3.

b

attenuation of the mean acoustic field propagating in the ocean waveguide. The attenuation increases with increasing frequency. The field scattered away from the specular direction, and, in particular, the backscattered field (called reverberation) acts as interference for active sonar systems. Because the ocean surface moves, it will also generate acoustic fluctuations. Bottom roughness can also generate fluctuations when the sound source or receiver is moving. The importance of boundary roughness depends on the sound-speed profile which determines the degree of interaction of sound with the rough boundaries. Often the effect of scattering from a rough surface is thought of as simply an additional loss to the specularly reflected (coherent) component resulting from the scattering of energy away from the specular direction. If the ocean bottom or surface can be modeled as a randomly rough surface, and if the roughness is small with respect to the acoustic wavelength, the reflection loss can be considered to be modified in a simple fashion by the scattering process. A formula often used to describe reflectivity from a rough boundary is: R0 ðyÞ ¼ RðyÞe0:5G

2

½8

where R0 (y) is the new reflection coefficient, reduced because of scattering at the randomly rough interface. G is the Rayleigh roughness parameter defined as G ¼ 2ks sin y

½9

where k ¼ 2p/l is the acoustic wavenumber and s is the rms roughness. Note that the reflection coefficient for the smooth ocean surface is simply 1 (the pressure-release condition is obtained from eqn [6] by setting r2 ¼ 0) so that the rough-sea-surface reflection coefficient for the coherent field is R0 ðyÞ ¼ exp 0:5G2 . For the ocean bottom, the

appropriate geoacoustic parameters (see Table 1) are used for evaluating R0 (y), and the rough-bottom reflection coefficient is then obtained from eqn [8]. Volume scattering is thought to arise primarily from biological organisms. For lower frequencies (less than 10 kHz), fish with air-filled swim bladders are the main scatterers whereas above 20 kHz, zooplankton or smaller animals that feed on the phytoplankton, and the associated biological food chain, are the scatterers. Many of the organisms undergo a diurnal migration rising towards the sea surface at sunset and descending to depth at sunrise. Since the composition and density of the populations vary with the environmental conditions, the scattering characteristics depend on geographical location, time of day, season and frequency. As an example, data from the Mediterranean Sea for volume scattering losses due to fish shoals show excess losses of 10–15 dB for a propagation range of 12 km and frequencies between 1 and 3 kHz. Finally, scattering off bubbles near the surface is sometimes referred to as either a volume- or surfacescattering mechanism. These bubbles arise not only from sea surface action, but also from biological origins and from ship wakes. Furthermore, bubbles are not the only scattering mechanism, but bubble clouds may have significantly different sound speed than plain seawater thereby altering local refraction conditions. At the sea surface, the relative importance of roughness versus bubble effects is not yet resolved.

Transmission-loss Data Figure 6 gives an example of transmission-loss variability in shallow water. The graph displays a collection of experimental data from different shallow-water areas (100–200 m deep) all over the



40

SD = 50.0 m RD = 50.0 m

3200 1600 Frequency (Hz)

50 60 70

800 400 200 100 50

80

25 (A)

90

SD = 50.0 m RD = 59.0 m 3200

100

1600 Frequency (Hz)

Transmission loss (dB)

117

110 120

800 400 200 100 50

130

25

140 1

2

5

10 20 Range (km)

50

100

Figure 6 Transmission loss variability in shallow water.

world. The data refer to downward-refracting summer conditions in the frequency band 0.5–1.5 kHz. Two features are of immediate interest. One is the spread of the data amounting to around 50 dB at 100 km and caused primarily by the varying bottomloss conditions in different areas of the world. The second feature is the fact that transmission is generally better than free-field propagation (20 log r) at short and intermediate ranges but worse at longer ranges. This peculiarity is due to the trapping of energy in the shallow-water duct, which improves transmission at shorter ranges (cylindrical versus spherical spreading), but, at the same time, causes increased boundary interaction, which degrades transmission at longer ranges. A second example of transmission-loss variability in shallow water is given in Figure 7, where broadband data from two different geographical areas are compared. The data set in Figure 7A, was collected in the Barents Sea in 60 m water depth. Note the high transmission losses recorded below 200 Hz, where energy levels fall off rapidly indicating that most of the acoustic energy emitted by the source is lost to the seabed. It is believed that this excess attenuation is caused by the coupling of acoustic energy into shear waves in the seabed. In contrast to the highloss environment in the Barents Sea, Figure 7B shows a data set from the English Channel in 90 m water depth. Here propagation conditions are excellent

10

0

(B)

20 Range (km)

30

40

Figure 7 Examples of frequency-dependent propagation losses measured in two shallow-water areas: (A) Barents Sea, (B) English Channel. Note the presence of an optimum frequency of propagation between 200 and 400 Hz.

over the entire frequency band. This second data set represents typical propagation conditions for thick sandy sediments with negligible shear-wave effects. Cutoff Frequency and Optimum Frequency

A common feature of all acoustic ducts is the existence of a low-frequency cutoff. Hence, there is a critical frequency below which the shallow-water channel ceases to act as a waveguide, causing energy radiated by the source to propagate directly into the bottom. The cutoff frequency is given by, f0 ¼

cw rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4D 1 ðcw =cb Þ2

½10

This expression is exact only for a homogeneous water column of depth D and sound speed cw overlying a homogeneous bottom of sound speed cb. As an example, let us take D ¼ 100 m, cw ¼ 1500 m s1, and cb ¼ 1600 m s1 (sand–silt), which yields f0C11 Hz. Sound transmission in shallow water has the characteristic frequency-dependent behavior shown in Figure 7, i.e. there is an optimum frequency of propagation at longer ranges. Thus the 80 dB contour line extends farthest in range for frequencies around 400 Hz in Figure 7A and around 200 Hz in


118


Figure 7B, implying that transmission is best at these frequencies – the optimum frequencies of propagation for the two sites. Optimum frequency is a general feature of ducted propagation in the ocean. It occurs as a result of competing propagation and attenuation mechanisms at high and low frequencies. In the high-frequency regime we have increasing volume and scattering loss with increasing frequency. At lower frequencies the efficiency of the duct to confine sound decreases (the cutoff phenomenon). Hence propagation and attenuation mechanisms outside the duct (in the seabed) become important. In fact, the increased penetration of sound into a lossy seabed with decreasing frequency causes the overall attenuation of waterborne sound to increase with decreasing frequency. Thus we get high attenuation at both high and low frequencies, whereas intermediate frequencies have the lowest attenuation. It can be shown that the optimum frequency for shallow-water propagation is strongly dependent on water depth (foptp D1), has some dependence on the sound-speed profile, but is only weakly dependent on the bottom type. Typically, the optimum frequency is in the range 200–800 Hz for a water depth of 100 m.

differences, and individual pulse shapes being modified due to frequency-dependent amplitude and phase changes associated with each boundary reflection. From simple geometrical considerations, the time dispersion is found to be Dt C

R 1 1 c¯ cosy

½11

where R is the range between source and receiver, c¯ is the mean sound speed in the channel, and y is the maximum propagation angle with respect to the horizontal. This angle will be determined either by the source beamwidth or by the critical angle at the bottom (the smaller of the two). Since the dispersion considered here is solely due to the geometry of the waveguide, it is called geometrical dispersion. An example of measured pulse arrivals over a 10 h period in the Mediterranean is given in Figure 9. Note that the time-varying ocean (internal waves, currents, tides) causes strong signal fluctuations with time, particularly in the earlier part of the signal. The time dispersion is 15–20 ms and at least four main energy packets, each consisting of several ray arrivals can be identified.

Signal Transmission in the Time Domain

1

S

2 3 4

10 9 8

Geo-time (h)

Even though underwater acousticians have traditionally favored spectral analysis techniques for gaining information about the band-averaged energy distribution within a shallow-water waveguide, additional insight into the complication of multipath propagation can be obtained by looking at signal transmission in the time domain. Figure 8 indicates that the signal structure measured downrange will consist of a number of arrivals with time delays determined by the pathlength

7 6 5

4

D′ D

3

R 2 BOTTOM

1

Figure 8 Schematic of ray arrivals in shallow-water waveguide. A series of multipath arrivals are expected, represented by eigenrays connecting source (S) and receiver (R). The shortest path is the direct arrival D followed by the surface-reflected arrival D0 . Next comes a series of four arrivals all with a single bottom bounce; then four rays with two bottom bounces as illustrated in the figure, followed by four rays with three bottom bounces, etc.

0 0

10

20 Signal time (ms)

30

40

Figure 9 Measured pulse arrivals versus geo-time over a 10 km shallow-water propagation track in the Mediterranean Sea. The bandwidth is 200–800 Hz.



119

Wave eq.

WI

NM

Coupled WI

Coupled NM

Range independent

Adiabatic NM

Ray

PE

FD / FE

Range dependent

Figure 10 Hierarchy of numerical models in underwater acoustics. WI, wavenumber integration; NM, normal modes; PE, parabolic equation; FD, finite difference; FE, finite element.

Numerical Modeling The advent of computers has resulted in an explosive growth in the development and use of numerical models since the mid-1970s. Numerical models have become standard research tools in acoustic laboratories, and computational acoustics is becoming an evermore important branch of the ocean acoustic science. Only the numerical approach permits an analysis of the full complexity of the acoustic problem. An assortment of models has been developed over the past 25 years to compute the acoustic field in shallow-water environments in both the frequency and time domains. Entire textbooks are dedicated to the development of theoretical and numerical formalisms which can provide quantitative acoustic predictions for arbitrary ocean environments. Sound propagation is mathematically described by the wave equation, whose parameters and boundary conditions are descriptive of the ocean environment. As shown in Figure 10, there are essentially five types of models (computer solutions to the wave equation) to describe sound propagation in the sea: wavenumber integration (WI); normal mode (NM); ray; parabolic equation (PE) and direct finite-difference (FD) or finite-element (FE) solutions of the full wave equation. All of these models permit the ocean environment to vary with depth. A model that also permits horizontal variations in the environment, i.e. sloping bottom or spatially varying oceanography, is termed range dependent. As shown in Figure 10, an a priori assumption about the environment being range independent, leads to solutions based on spectral techniques (WI) or normal modes (NM); both of these techniques can, however, be extended to treat range dependence. Ray, PE and FD/FE solutions are applied directly to rangevarying environments. For high frequencies (a few kilohertz or above), ray theory, the infinite frequency approximation, is still the most practical, whereas the other five model types become more and more applicable below, say, a kilohertz in shallow water.

Models that handle ocean variability in three spatial dimensions have also been developed, but these models are used less frequently than two-dimensional versions because of the computational cost involved.

Conclusions The acoustics of shallow water has been thoroughly studied both experimentally and theoretically since World War II. Today the propagation physics is well understood and sophisticated numerical models permit accurate simulations of all processes (reflection, refraction, scattering) that contribute to the complexity of the shallow-water problem. Sonar performance predictability, however, is limited by knowledge of the controlling environmental inputs. The current challenge is therefore how best to collect relevant environmental data from the world’s enormously variable shallow-water areas.

See also Acoustic Noise. Acoustic Scattering by Marine Organisms. Acoustics, Arctic. Acoustics in Marine Sediments. Sonar Systems. Surface Gravity and Capillary Waves. Tomography.

Further Reading Brekhovskikh LM and Lysanov YP (1990) Fundamentals of Ocean Acoustics, 2nd edn. New York: SpringerVerlag. Etter PC (1996) Underwater Acoustic Modeling, 2nd edn. London: E & FN Spon. Jensen FB, Kuperman WA, Porter MB, and Schmidt H (2000) Computational Ocean Acoustics. New York: Springer-Verlag. Medwin H and Clay CS (1998) Fundamentals of Acoustical Oceanography. San Diego: Academic Press. Urick RJ (1996) Principles of Underwater Sound, 3rd edn. Los Altos: Peninsula Publishing.


AEOLIAN INPUTS R. Chester, Liverpool University, Liverpool, Merseyside, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 97–103, & 2001, Elsevier Ltd

Introduction The oceans are an important reservoir in the global biogeochemical cycles of many elements, but until the recent past it was thought that material fluxes to the reservoir were dominated by fluvial inputs. Over the last two or three decades, however, it has become apparent that the atmosphere is a major transport pathway in the land–sea exchange of material. This atmospherically transported material differs from that introduced by fluvial inputs in two important ways. (i) It is delivered, albeit at different flux magnitudes, to all areas of the sea surface, whereas river inputs are initially delivered to the land–sea margins. (ii) It does not pass through the biogeochemically dynamic estuarine filter; a region of intense dissolved/particulate reactivity, which, under present day conditions, retains B90% of fluvial particulate material. As a result, the atmosphere is the most important pathway for the long-range transport of much of the particulate material delivered directly to open-ocean regions. This material has an important influence on marine sedimentation; for example, in equatorial North Atlantic deep-sea sediments deposited to the east of the Mid-Atlantic Ridge and in central North Pacific deep-sea sediments, essentially all the land-derived components are aeolian in origin. Atmospheric aerosols can also exert an influence on climatic forcing by acting as cloud condensation nuclei and by processes such as the scattering of short-wave radiation by both anthropogenic and natural aerosols. Further, the presence of anthropogenic sulfate aerosols in the atmosphere can lead to an increase in albedo and so cool the planet; an effect, which, on a global scale, is comparable to that induced by the ‘greenhouse’ gases, but is opposite in sign. The aeolian material delivered to the sea surface by the atmosphere is dispersed from the source regions via the major wind systems, such as the Trades and the Westerlies, within which relatively smallscale winds (e.g. the Sirocco and the Mistral in the Mediterranean) can be important locally. Large-scale meteorological phenomena can also affect aeolian

120

transport; for example, long-term inter-annual variability in dust transport out of Africa to the Atlantic Ocean and the Mediterranean Sea has been linked to precipitation patterns induced by the North Atlantic Oscillation. Material in the marine atmosphere consists of gaseous and particulate components, both of which can originate from either natural or anthropogenic sources. Gas-to-particle conversions are important in the generation of particulate material, especially that derived from anthropogenic sources; however, air/sea gaseous exchange is covered in articles on air–sea interactions, and attention here is largely confined to the particulate aerosol. The sea surface itself is a major source of particulate material to the marine atmosphere in the form of sea salt. However, these sea salts are re-cycled components, and the globally important terrestrial sources of material to the marine atmosphere, i.e. those supplying material involved in land– sea exchange, are (i) the Earth’s crust (mineral dust), and (ii) anthropogenic processes (sulfates, nitrates, etc). Other terrestrial sources, which include volcanic activity and the biosphere (e.g. direct release from vegetation, biomass burning), can also supply components to the atmosphere. Material is removed from the atmosphere by a combination of two depositional modes; (i) the ‘dry’ mode, which does not involve an aqueous phase, and (ii) the ‘wet’ (precipitation scavenging) mode, either by cloud droplets (in-cloud processes) or by falling rain (below-cloud processes).

Land–sea Exchange of Individual Components Marine Aerosol: Major Components, Sources, and Distribution

Data are available on the concentrations of aerosols over many marine regions, and it is now apparent that there is an ‘aerosol veil’ over all oceans. Concentrations of material in the aerosol veil, however, vary from B103 ng m3 of air close to continental sources to B102 ng m3 of air over pristine oceanic regions. The aerosol veil is composed mainly of mineral dust and anthropogenic components. On a global scale, the anthropogenic material is dominated by sulfate aerosols, together with smaller amounts of nitrates. The mineral dust aerosol consists of a wide variety of minerals, with quartz, the clay minerals and feldspars usually being


AEOLIAN INPUTS

predominant. The signatures of the major clay minerals (chlorite, kaolinite, illite, and montmorillonite) can be used as tracers to identify the sources of the dusts, the extent to which the material has been transported over the oceans, and its contribution to marine sedimentation. The principal continental sources of both the mineral dust aerosol and anthropogenic (mainly sulfate) aerosol are concentrated in specific latitudinal belts, predominantly in the northern hemisphere (see Figure 1A). Quantitatively, the crust-

121

derived mineral dust, which is derived mainly from the arid and semi-arid desert regions of the world (Figure 1A), imposes the strongest fingerprints on the marine aerosol. Mineral dust fluxes to the world ocean are listed in Table 1, and are illustrated in Figure 2. From this figure it can be seen that the highest dust fluxes to the sea surface are found off the major deserts, e.g. the Sahara in the North Atlantic and the Asian deserts in the North Pacific. Much of the material injected into the atmosphere from these arid sources is transported in the form of

40°N

0

40°S

(A)

120°E

60°E

60°E

60°N

180

120°E

_2

60°W

0

00

0

100

00

10

0 00

120°E

10

_1

y )

30°N 0

120°W

Dust flux (mg m

60°E

0

00

10

1000 10 000

100

1

1000 100

30°S 60°S

10 10

10

(B) Figure 1 Aerosols: terrestrial sources and fluxes to the world ocean. (A) Terrestrial sources of aerosol production; light gray areas indicate regions of anthropogenic emissions, and dark gray areas indicate regions of mineral aerosol production. (Reproduced with permission from Gilman C and Garrett C (1994) Heat flux parameterizations for the Mediterranean Sea: The role of atmospheric aerosols and constraints from the water budget. Journal of Geophysical Research 99: 5119–5134.) (B) Mineral aerosol fluxes to the world ocean (units, mg m2 y 1). (Reproduced with permission from Duce RA et al. (1991)).


122

AEOLIAN INPUTS

Table 1

Dust deposition rates to the world ocean

Ocean region

Deposition rate

North Atlantic, north of trades North Atlantic trades South Atlantic Indian Ocean North Pacific: western North Pacific central and eastern South Pacific Entire Pacific

106g cm2y1

1012g y1

82 – 85 450

12 100–400 18–37 336

5000 11–62 5–64

300 30 18 350

All oceans (minimum–maximum)

816–1135

Adapted from Prospero JM (1981) and Prospero JM, Uematso M and Savoie DL (1989). In: Riley JP and Chester R (eds) Chemical Oceanography, vol. 10, pp. 137–218. London: Academic Press.

dust ‘pulses’; these are related to dust storms on the continental source regions, and are superimposed on background aerosol concentrations. Organic Matter and Organic Compounds

Various classes of particulate organic carbon (POC) and vapour-phase organic carbon (VOC) are present in the atmosphere. The non-methane atmospheric global VOC burden has been estimated to be B50 1012 g, and the total POC burden to be B1– 5 1012 g. The principal terrestrial sources of organic matter to the atmosphere are vegetation, soils, biomass burning, and the freshwater biomass, together with a variety of anthropogenic processes. The sea surface also contributes to the organic matter 60°E

120°E

180

burden in the marine atmosphere, with B14 1012 g y1 of organic carbon being produced by the ocean surface; 490% being on particles 41 mm in diameter. Particulate organic matter (POM) is removed from the air via ‘dry’ and ‘wet’ deposition, and in addition the removal of VOM includes conversion to POM and transformation to inorganic gaseous products. Estimates of the ‘wet’ atmospheric depositional flux of carbon to the ocean surface range between B2.2 1014 g y1 and B10 1014 g y1, and for the ‘dry’ flux a value of B6 1012 g y1 has been proposed. These are of the same order of magnitude as the estimates of fluvially transported POC entering the oceans (B1–2.5 1014 g y1). The estimates of atmospherically transported carbon must be regarded with extreme caution, but nonetheless, even when the marine source is taken into account, the ‘wet’ and ‘dry’ flux estimates indicate that the oceans act as a major sink for organic carbon in the atmosphere. Atmospherically transported carbon, however, makes up a maximum of only B2% of the carbon produced by primary productivity (B30–50 1015 g y1). Viable POC in the marine atmosphere includes material such as fungi, bacteria, pollen, algae, insects, yeasts, molds, mycoplasma, viruses, phages, protozoa, and nemotodes. Non-viable POC includes carbonaceous material (which has a refractory ‘soot’ component) and individual organic species. Concentrations of carbonaceous aerosols in the marine atmosphere vary over the range B0.05–1.20 mg C m3 of air and display a distinct latitudinal distribution. In the northern hemisphere the carbon, which ranges in concentration between B0.4 and B1.2 mg C m3 of air, has a predominantly anthropogenic continental source 120°W

60°W

0 100

60°N

Lead flux (μg m

_2

_1

y )

1000

1000

100

1000

30°N

100 100

0

10 100

100

10

10

30°S

1000

100

100

60°S

(A) Figure 2 Fluxes of lead to the world ocean (units, mg m2 y1). From Duce RA et al. (1991).


10

AEOLIAN INPUTS

from combustion processes. In the southern hemisphere POC concentrations are lower (range B0.05– 0.30 mg C m3 of air), and natural continental and marine sources are about equal. A wide variety of individual species of both natural and synthetic organic compounds are transported from the continents to the sea surface via the atmosphere. The classes of organic compounds that have received particular attention include aliphatic hydrocarbons, wax esters, fatty alcohols, sterols, fatty acids, and long-chain unsaturated ketones. The most extensive investigation of atmospherically transported organic species on an ocean-wide basis was carried out in the Pacific as part of the SEXREX Program (Sea-Air Exchange Program). On the basis of data obtained from the Pacific it is apparent that in the remote marine atmosphere organic carbon accounts for B10% of the total aerosol, although only B1% of this has been characterized. The most abundant terrestrially derived components are the nalkanes and the C21–C36 fatty alcohols, which are common in the epicuticular waxes of vascular plants, and the most abundant marinederived species are the C13–C18 fatty acid salts. The atmosphere is a major pathway for the transport of a number of organic pollutants that enter sea water. These include the synthetic trace organics, such as high molecular-weight halogenated hydrocarbons of the following compounds, or compound classes; chlorobenzenes (e.g. hexachlorobenzene, HCB), chlorocyclohexanes (e.g. hexachlorocyclohexanes, HCHs), polychlorobiphenyls (PCBs), dichlorodiphenyltrichloroethanes (DDTs), and non-halogenated coumpounds (e.g. polynuclear aromatic hydrocarbons, PAHs). Global atmospheric fluxes of HCHs, HCBs, DDTs, and PCBs to the world ocean are listed in Table 2, and two important conclusions can be drawn from the data. (i) The dominant deposition of the organochlorines is to the North Atlantic and North Pacific, which is consistent with their source derivations; with HCH and DDT compounds having their highest deposition rates in the Table 2

123

North Pacific and PCBs in the North Atlantic. (ii) The atmospheric inputs of the organochlorines to the world ocean exceed those from fluvial inputs by 1 to 2 orders of magnitude. Nutrients

These include nitrate, phosphate, and micro-nutrients such as iron. The main features in the atmospheric input of nitrogen nutrients to the world ocean can be summarized as follows. (i) The atmospheric input of total nitrogen (N) to the global ocean is B30.2 1012 g N y1, made up of B13.4 1012 g N y1 oxidized nitrogen species and B16.8 1012 g N y1 reduced nitrogen species, which is similar in magnitude to the total (i.e. natural þ anthropogenic) fluvial nitrogen flux (B21–49 1012 g N y1). (ii) The overall flux of nitrogen to the sea surface is B87 mg N m2 y1; the largest fluxes being to the North Atlantic and North Pacific. (iii) The ‘wet’ removal of reduced nitrogen species may be an important source of nutrients to the oceans. The atmospheric fluxes of nitrogen species usually include only inorganic forms, such as nitrate and ammonia; however, if dissolved organic nitrogen (DON), most of which has an anthropogenic source, is included in ‘wet’ deposition it would increase the anthropogenic input of fixed nitrogen to the oceans by a factor of B1.5. The atmospheric input of phosphate has been studied in detail in the Mediterranean Sea. However, the role played by atmospherically transported phosphorus in nutrient cycles is not clearly understood; for example, it has been proposed that Saharan dust may act as a sink for the removal of dissolved phosphorus in the water column by adsorption onto iron-rich particles, or as a source with up to B8% of the phosphorus in the dusts being soluble in sea water. There is evidence that in summer months the atmosphere does provide a source of phosphorus to the Western Mediterranean and that this may account for new production, since the stratification of the surface waters prevents the input

The atmospheric input of some organochlorine compounds to the world oceans (units, 106 g y1)

Compounda

North Atlantic

South Atlantic

North Pacific

South Pacific

Indian Ocean

Global atmospheric

Global fluvial

HCHs HCB DDTs PCBs

850 17 16 100

97 10 14 14

2600 20 66 36

470 19 26 29

700 11 43 52

4800 77 170 240

40–80 4 4 40–80

a HCHs, hexachlorocyclohexanes; HCB, hexachlorobenzene; DDTs, dichlorodiphenyltrichloroethanes; PCBs, polychlorobiphenyls. Adapted from Duce RA et al. (1991).


124

AEOLIAN INPUTS

of nutrients from deep waters. For example, in the Ligurian Sea a strong summer desert dust transport episode was followed 10 days later by a significant increase in phytoplankton concentration, which could have resulted from the solubilization of phosphorus from the dusts. In recent years there has been a renewed interest in the role played by iron as a limiting nutrient in primary production, especially in high-nitrogen, lowproductivity (HNLP) regions in which there is sufficient light and nutrient concentrations but low productivity. The concentrations of iron in openocean waters are generally low, and it is thought that the element would run out before nitrate is exhausted. To provide sufficient iron it therefore has been suggested that it must be added to sea water from other sources, one of which is the long-range transport of atmospheric dust. However, iron oxyhydroxide particles and iron colloids are not directly available to phytoplankton, and the bioavailable forms of the metal are thought to be dissolved Fe(II), which is rapidly converted into Fe(III), and Fe(III) itself. As a result, iron must be dissolved from atmospheric dust before it becomes available to phytoplankton, and although adsorption onto particulate matter is the dominant control on the concentrations of dissolved iron in open-ocean waters, the deposition of mineral dust from high concentration episodic atmospheric events can result in a net addition of dissolved iron to surface waters. Further, during ‘wet’ deposition iron may undergo reductive dissolution to Fe(II), a form of iron that would be immediately available to phytoplankton. It also has been suggested that siderophores (compounds with a high affinity for ferric iron which are secreted by organisms) may play a role in the bioavailability of iron. Various small-scale laboratory simulations have shown that phytoplankton growth rates increase in response to the addition of iron to the system, but there is considerable disagreement over the interpretation of the results. Laboratory experiments also have been criticized on the grounds that they do not represent planktonic community response on an ocean-wide scale. To overcome this, the Fe-limitation hypothesis has been tested by largescale intervention experiments, such as IronEx I and IronEx II, which were carried out in the equatorial Pacific. In these experiments patches of sea (defined by SF6 tracer) were seeded with iron in the concentrations expected from natural events. Although there was a doubling of the plant biomass following iron addition, interpretation of the data from IronEx I was hampered as a result of the subduction of the seeded seawater patch below a layer of less dense water. During IronEx II, however, a massive

phytoplankton bloom was triggered, providing direct evidence that in these HNLP waters phytoplankton growth is iron limited.

Trace Metals

Particulate trace metals in the marine atmosphere that are involved in land–sea exchange are derived from two principal terrestrial sources. The first is the Earth’s crust. Crustal weathering involves low-temperature generation processes, and crust-derived elements are found on particles with mass median diameters (MMDs) in the range B1–3 mm. Relative to other sources, crust-derived elements are referred to as the non-enriched elements (NEEs). The second source comprises a variety of anthropogenic processes, which often involve high temperatures (e.g. fuel combustion, ore smelting). Anthropogenically derived elements (e.g. Pb, Cu, Zn, Cd, As, Hg) are largely found on smaller particles with MMDso0.5 mm, and are termed the anomalously enriched elements (AEEs). Less important terrestrial sources of trace metals to the marine atmosphere include volcanic activity and the biosphere. In addition, the sea surface supplies recycled elements to the marine atmosphere; this is a low-temperature source, and sea salt-associated elements are located on particles with MMDs in the range B3–7 mm. Global elemental atmospheric emission rates from natural and anthropogenic sources are listed in Table 3. The overall trace metal composition of the marine aerosol is dependent on the extent to which material from the various sources are mixed together in the atmosphere, and trace metal concentrations are a function of factors such as the distance the air mass Table 3 Global elemental emission rates to the atmosphere (units, 109 g y1) Element

Natural

Anthropogenic

Al As Cd Co Cr Cu Fe Hg Mn Pb V Zn

20940 12 1.4 6.1–7.3 44 22–28 10370 2.5 221–317 11.5–12 41–45 45–280

4000 19 7.7 2.9 30.5 35–52 6000 3.6 39–408 332–404 22–86 132–280

Adapted from Chester R (2000) (full references given in original table).


AEOLIAN INPUTS

Table 4

125

Concentrations of particulate trace metals in the marine atmosphere (units, ng m3 of air) Coastal regions

Open-ocean regions

Close to anthropogenic sources

Close to crustal sources

Trace metal

North Sea

W. Black Sea

N. Atlantic N. Arabian north-east Sea trades

Tropical Tropical Indian Tropical N. Atlantic Ocean N. Pacific

Tropical S. Pacific

Al Fe Mn Ni Cr V Cu Zn Pb

294 353 14.5 3.8 4.7 – 6.3 41 34.5

540 420 17 4.9 9.0 3.2 – 46 60

5925 3685 65 6.6 10 15 4.5 16 6.9

160 100 2.2 0.64 0.43 0.54 0.79 4.4 9.9

– – – 0.013 0.07 0.016

1227 790 17 2.0 3.0 6.3 2.6 10 4.3

11 8.8 0.16 0.043 0.066 0.023 0.077 0.10 0.17

21 17 0.29 – 0.09 0.08 0.045 0.17 0.12

Adapted from Chester R (2000) (full references given in original table).

transporting them has travelled from the source, the ‘aging’ of the aerosol in the air, and the relative effectiveness of the processes that remove material from the atmosphere. As a result, trace metals in the marine atmosphere have concentrations ranging over several orders of magnitude – see Table 4. It is apparent from the data in this table that, in general, the trace metal concentrations decrease with increasing remoteness from continental sources in the general rank order: coastal seas4North Atlantic4North Pacific and tropical Indian Ocean4South Pacific. This is reflected in trace metal fluxes, which decrease in the same sequence; lead atmospheric fluxes to the world ocean are illustrated in Figure 2. Despite the fact that trace metal concentrations decrease towards more pristine oceanic environments, atmospheric inputs can be the dominant source of some trace metals to the mixed layer in open-ocean regions where inputs from other sources are minimal. This is especially the case for the ‘scavenged-type’ metals (e.g. Al, Mn, Pb), which have a surface source and a relatively short residence time in sea water. Aerosol-associated trace metals are removed from the air by ‘dry’ or ‘wet’ depositional processes, and once deposited at the sea surface the initial constraint on the manner in which the metals enter the major marine biogeochemical cycles is a function of the extent to which they undergo solubilization in sea water. There is considerable size-dependent fractionation between the parent aerosol and the deposited material in both the ‘dry’ and the ‘wet’ depositional modes. However, with respect to the seawater solubility of trace metals, the major difference between the two depositional modes is that they

follow separate geochemical routes. In ‘dry’ deposition material is delivered directly to the sea surface and trace metal solubility is largely constrained by particle 2 seawater reactivity; estimates of the seawater solubility of trace metals from aerosols are listed in Table 5. In contrast, in ‘wet’ deposition there is an initial particle 2 rainwater reactivity; much of this is pH-dependent, and can involve the dissolution of some trace metals prior to the deposition of the scavenged aerosol at the sea surface. Data on the trace metal composition of marine rainwaters from a number of regions are now available, and a selection are listed in Table 6, from which it can be seen that trace metals in rainwaters, like those in aerosols, have their highest concentrations in coastal regions and decrease towards more pristine oceanic regions.

Table 5 Seawater solubility of trace metals from particulate aerosols; % total element soluble Trace metal

Solubility (%)

Al Fe Mn Ni Cr V Cu Zn Pb

B1–10 B1–50 B20–50 t20–50 t10–20 t20–85 t10–85 t10–75 t10–90

Data, which are from various sources, include both crustal and anthropogenic aerosols.


126

AEOLIAN INPUTS

Table 6

Trace metal concentrations in marine-influenced rainwaters (units, ng l1) Mediterranean Sea

North Atlantic

Trace metal

North Sea coast (VWM)

Irish Sea coast

South coast, France (VWM)

Sardinia (VWM)

Bermuda (VWM)

Bantry Bay, Ireland (VWM)

North Pacific

South Pacific

Al Fe Mn Cu Zn Pb

– 88 3.8 2.3 13 4.0

43 48 2.0 8.7 9.3 5.2

144 – – 2.8 – 3.7

883 519 8.0 2.9 16 1.6

– 4.8 0.27 0.66 1.15 0.77

3.62 8.06 0.13 0.86 8.05 0.51

2.1 1.0 0.012 0.013 0.052 0.035

16 0.42 0.020 0.021 1.6 0.014

VWM, volume-weighted mean concentrations; this normalizes the trace metal concentration in a rain to the total amount of rainfall over the sampling period. Adapted from Chester R (2000) (full references given in original table).

Table 7 Atmospheric and fluvial trace metal fluxes to the world ocean (units, 109 g y1) Atmospheric input

Fluvial input

Element

Dissolved

Particulate

Dissolved

Particulate

Fe P Ni Cu Pb Zn Cd As

3.2 103 310 8–11 14–45 80 33–170 1.9–3.3 2.3–5.0

29 103 640 14–17 2–7 10 11–60 0.4–0.7 1.3–2.9

1.1 103 Total 300a 11 10 2 6 0.3 10

110 103 Total 300 1400 1500 1600 3900 15 80

Total phosphorus input to marine sediments. Adapted from Duce RA et al. (1991).

Estimates of particulate and dissolved atmospheric and fluvial trace metal fluxes to the world ocean are listed in Table 7, from which a number of overall conclusions can be drawn. (i) Rivers are the principal source of particulate trace metals to the oceans; phosphorus being an exception. (ii) For iron, nickel, copper, and phosphorus the dissolved atmospheric and the dissolved fluvial inputs are the same order of magnitude. (iii) For lead, zinc, and cadmium the dissolved atmosphere fluxes are dominant; this will still be the case for lead even when allowance is made for the phasing out of leaded gasoline, which has accounted for a large fraction of the anthropogenic lead previously released into the atmosphere.

Conclusions There is an ‘aerosol veil’ over all marine regions, and the atmosphere is a major transport route for the supply of mineral dust, organic matter, nutrients, and

trace metals to the world ocean. Atmospheric fluxes decrease in strength away from the continental source regions, but in remote open-ocean areas they can be the dominant supply route for the deposition of land-derived particulate material and some trace metals to the ocean surface. Unlike fluvial inputs, which are delivered to the land–sea margins, the atmosphere supplies material to the ‘mixed layer’ over the whole ocean surface, and the material plays an important role in oceanic biogeochemical cycles and in the formation of marine sediments.

See also Air–Sea Gas Exchange. Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens. Air–Sea Transfer: N2O, NO, CH4, CO; Anthropogenic Trace Elements in the Ocean. Atmospheric Input of Pollutants. Metal Pollution. Nitrogen Cycle. Phosphorus Cycle. Photochemical Processes. Refractory Metals. Transition Metals and Heavy Metal Speciation.

Further Reading Buat-Menard P (ed.) (1986) The Role of Air-Sea Exchange in Geochemical Cycling. Dordrecht: Kluwer Academic Publishers. Charlson RJ and Heintzenberg (eds.) (1995) Aerosol Forcing of Climate. Berlin: Dahlem Workshop. Chester R (2000) Marine Geochemistry 2nd edn. Oxford: Blackwell Science. Duce RA, Mohnen VA, Zimmerman PR, et al. (1983) Organic material in the global troposphere. Reviews of Geophysics and Space Physics 21: 921--952. Duce RA, Liss PS, Merrill JT, et al. (1991) The atmospheric input of trace species to the World Ocean. Global Biogeochemical Cycles 5: 193--529.


AEOLIAN INPUTS

Guerzoni S and Chester R (eds.) (1996) The Impact of Desert Dust Across the Mediterranean. Dordrecht: Kluwer Academic Publishers. Knap AH (ed.) (1990) The Long Range Atmospheric Transport of Natural and Contaminant Substances. Dordrecht: Kluwer Academic Publishers. Peltzer ET and Gagosian RB (1989) Organic chemistry of aerosols over the Pacific Ocean. In: Riley JP and Chester R (eds.) Chemical Oceanography, vol. 10, pp. 282--338. London: Academic Press.

127

Prospero JM (1981) Eolian transport to the World Ocean. In: Emiliani C (ed.) The Sea, vol. 7, pp. 801--874. New York: John Wiley, Interscience. Prospero JM (1996) The atmospheric transport of particles to the ocean. In: Ittekkot V, Schafer P, Honjo S, and Depetris PJ (eds.) Particle Flux to the Oceans, pp. 19--52. New York: John Wiley.


AGULHAS CURRENT J. R. E. Lutjeharms, University of Cape Town, Rondebosch, South Africa Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 104–113, & 2001, Elsevier Ltd.

Introduction The greater Agulhas Current forms the western boundary system of the circulation in the South Indian Ocean. Contrary to the flow of comparable subtropical gyres in other ocean basins, the sources of the Agulhas Current are interrupted by a substantial barrier, the island of Madagascar. This leads to the formation of two minor western boundary flows, the East Madagascar Current and the Mozambique drift. Once fully constituted off the coast of south-eastern Africa, the Agulhas Current proper can be considered to consist of two distinct parts: the northern and the southern current. The northern part flows along a steep continental shelf and its trajectory is extremely stable. The southern part flows along the wide shelf expanse of the Agulhas Bank and by contrast meanders widely. South of the African continent the Agulhas Current retroflects in a tight loop, with most of its waters subsequently flowing eastward as the Agulhas Return Current. This loop configuration is unstable and at irregular intervals it is pinched off to form a detached Agulhas ring. These rings, carrying warm and salty Indian Ocean water, drift into the South Atlantic Ocean. Some cross the full width of this ocean in the next 2– 3 years, whereas many are dissipated within 5 months of being spawned. The Agulhas Return Current flows back into the South Indian Ocean along the Subtropical Convergence. This juxtaposition generates considerable mesoscale turbulence in the form of meanders and an assortment of eddies. Water from the Agulhas Return Current leaks northward, back into the subtropical gyre, along its full length. By about 701E all Agulhas water has been lost to the eastward flow that subsequently continues as the South Indian Ocean Current.

Importance Historically the Agulhas Current was one of the first ocean currents to receive a great deal of scientific attention. It was described in some detail as early as 1766 by Major James Rennell, preeminent British

128

geographer at the time. This was followed by wideranging investigations by Dutch mariners such as Van Gogh and Andrau in the 1850s. This early interest was motivated purely by nautical concerns, the Agulhas Current constituting a formidable impediment to vessels sailing to India and to the East. Fundamental studies by German investigators dominated research on the Agulhas Current region in the 1930s, but this endeavour was terminated by the Second World War. During the past few decades a renaissance in interest in this current has occurred for totally different reasons. It has been demonstrated that the greater Agulhas Current system (Figure 1) has a marked influence on the climate variability over the southern African subcontinent. It has also been shown that this current is a key link in the exchanges of water between ocean basins and thus probably has a special role in the oceans’ influence on global climate. This renewed interest has stimulated a number of research cruises, the placement of current meter moorings, investigations by satellite remote sensing, as well as theoretical and modeling studies, all leading to an enormous increase in knowledge of the Agulhas Current.

Large-scale Circulation The Agulhas Current forms part of the overall circulation of surface waters in the South Indian Ocean that is anticyclonic, i.e., anticlockwise in the Southern Hemisphere (Figure 2). On its eastern side, the equatorward flow is weak and dispersed, whereas the recirculation in the South West Indian Ocean is particularly strongly developed, penetrating to 1000 m depth at its centre. The southern border to the circulation is the Subtropical Convergence. This strong thermohaline front at roughly 411S separates the characteristic flows and water masses of the subtropical gyre and those of the Antarctic Circumpolar Current that lies to the south. Along the Subtropical Convergence, the Agulhas Return Current and the South Indian Ocean Current carry their respective water masses eastward. To the north the gyral circulation is closed by the South Equatorial Current that is found between about 101 and 251S and carries water from east to west. At the eastern shores of Madagascar, the South Equatorial Current splits into a northern and a southern limb of the East Madagascar Current; about 70% going north along this shoreline, 30% heading south. Most of that heading north eventually


AGULHAS CURRENT

E 30˚

20˚

40˚

50˚ 20˚ g as c

ar

10˚

129

da

Walvisbaai

Ridge Cu

ul

Port Elizabeth

Ag

Curren

ha

s

la t

Cape Town Agulhas Bank

30˚

Mada gasc Ridge ar

ue

rre

nt

ng

Mozam

Be Durban

a t t- M ren Eas Cur

bique

Maputo

S

Ag

has ul

l i ca

R e t ur n C u r r e n

t

40˚

o v e r g e nc e

C

p S u btr o

Agulhas Plateau

Current

Circumpolar

Antarctic

Figure 1 A conceptual portrayal of the greater Agulhas Current. Ocean regions shallower than 3000 m have been shaded. Intense currents are black, whereas the general background circulation is shown by open arrows. Cyclonic eddies are open; anticyclonic rings and eddies are black. Note the stability of the northern Agulhas Current, the meanders of the southern Agulhas Current, the tight retroflection loop, and the continuously weakening eastward flow of the Agulhas Return Current along the Subtropical Convergence. Agulhas rings (black) are advected by the Benguela Current past the extensive coastal upwelling off south-western Africa.

10˚

10 15

5

10

5

10

5

10

20˚

5

25

S

5

10 5 5

30˚

5

65 40˚

10

10

5 5

10

5

10

60 40

20˚

5

40˚

5 5

5 5

20 60˚

80˚

100˚

120˚

E Figure 2 The baroclinic volume transport for the upper 1000 m of the South Indian Ocean. Values are in units of 106 m3 s1. Note the very small contribution coming from the Mozambique Channel and the concentration of the recirculation in the South West Indian Ocean west of 701E. This transport pattern, averaged over a long period, is not to be confused with the depictions of instantaneous currents in Figures 1 and 4. (After Stramma and Lutjeharms (1997) Journal of Geophysical Research 102(C3): 5513–5530. ^ American Geophysical Union.)

reaches the east coast of the African continent, where it forms the East African Coastal Current. There is some leakage from the Mozambique Channel and from the southern limb of the East Madagascar Current into the Agulhas Current, but most of the Agulhas Current’s waters come from the subgyre of the South West Indian Ocean (see Figure 2). The

Agulhas Current itself is narrow, deep, and fast; a typical western boundary current. All these surface currents are driven largely by the reigning wind systems. The wind systems over the South West Indian Ocean fall largely outside the influence of the seasonally varying monsoonal winds of the North


130

AGULHAS CURRENT

Indian Ocean. The average air motion over the South Indian Ocean is dominated by a large-scale, anticyclonic circulation around a high-pressure system centered south-east of the island of Madagascar. This flow generally is stronger in austral summer than in winter. Strongest winds (24 m s1) are found at 501S latitude in summer, and weakest at 351S (2 m s1). A band of minimum wind stress extends across this ocean at 351S. Next to the continental land masses the winds are usually aligned with the coasts. Apart from driving the surface currents, the atmosphere also has a considerable effect on the formation of certain water masses (Figure 3) that are typical for the region. This temperature–salinity portrayal indicates the characteristic for each specific water mass in this ocean region. The fresher Tropical Surface Water is formed north of 201S where there is an excess of precipitation over evaporation; Subtropical Surface Water is formed between 281 and 381S, where this ratio is reversed. Subtropical Surface Water is found as a shallow subsurface salinity maximum in regions to the north and south of its region of formation. Intermediate waters lie at depths between 1000 and 2000 m and consist of Antarctic Intermediate Water and North Indian Intermediate Water (also called Red Sea Water). The former subducts between the Antarctic Polar Front and the Subtropical Convergence;

30 Tropical Surface Water

26

Potential temperature (˚C)

22 Tropical Thermocline Water

18

Subtropical Surface Water

the latter is formed owing to very high rates of evaporation in the Red Sea, the Arabian Sea and the Persian Gulf. Central Water, lying between the surface and the intermediate waters, is a mixture of these two. Below the intermediate waters are Indian Deep Water and North Atlantic Deep Water, each formed by subduction in the respective ocean regions after which they are named. All these respective water masses are involved in some way or other in the source currents of the Agulhas Current.

Sources of the Agulhas Current According to the transport portrayed in Figure 2, 30% of the volume flux of the Agulhas Current derives from east of Madagascar, only 13% comes through the Mozambique Channel, and 67% is recirculated in a South West Indian Ocean subgyre. Note, however, that the inflow from east of Madagascar does not necessarily come from the East Madagascar Current (see Figure 1). The southern limb of the East Madagascar Current starts at the bifurcation point of the South Equatorial Current at about 171S along the east coast of Madagascar. Its surface speed here is roughly 1 m s1, increasing downstream to about 1.5 m s1. The current is very stable in both flux and trajectory and exhibits no clear seasonality in any of its characteristics. Using the 0.5 m s1 isotach as the outer limits of the current, it is 75 km wide, 200 m deep, and its core lies 50 km offshore. It carries Tropical as well as Subtropical Surface Water with a total volume transport of 21 106 m3 s1 (i.e., 21 Sv). Where it overshoots the end of the shelf of Madagascar it retroflects, with most of its waters subsequently heading eastwards (Figure 4). There may be some E

14

40˚

10

S SAASW

6 2 _2 33.8

Indian Deep Water

34.2

W

34.6

50˚

SAMW

South Equatorial Current

Red Sea Water

AAIW

AASW

IC

20˚

Mozambique Current

Reunion

NADW

East Madagascar Current

AABW

35.0 35.4 35.8 36.2 Salinity (PSU)

36.6

Figure 3 The relationship of potential temperature and salinity for waters in the western Indian Ocean. Some of the characteristic water masses to be found here are SICW, South Indian Central Water; SAMW, Subantarctic Mode Water; NADW, North Atlantic Deep Water; AABW, Antarctic Bottom Water; AASW, Antarctic Surface Water; AAIW, Antarctic Intermediate Water; and SAASW, Subantarctic Surface Water. (After Gordon et al. (1987) Deep-Sea Research 34(4): 565–599. ^ Elsevier Science.)

S

Retroflection Agulhas Current

30˚

Figure 4 A conceptual portrayal of the flow regime in the source regions of the Agulhas Current. Narrow, intense currents are shown by black arrows. The East Madagascar Current is a miniature western boundary current that retroflects south of Madagascar. (After Lutjeharms et al. (1981) Deep-Sea Research 28(9): 879–899. ^ Elsevier Science.)


AGULHAS CURRENT

Northern Agulhas Current Water from the South West Indian Ocean subgyre feeds into the Agulhas Current along its full length, but at a latitude of 271S this current is nonetheless thought to be fully constituted (Figure 5). This northern part of the Agulhas Current is characterized in particular by an extremely stable

30

0 20 10 _1 cm s 0

500

1000

1500

Depth (m)

leakage of East Madagascar Water into the Agulhas Current by way of rings and filaments, but this constitutes an insignificant contribution. The wider flow east of Madagascar, up to 240 km offshore, is 41 106 m3 s1. A substantial part of this more extensive flow may eventually make its way into the Agulhas Current (Figures 2 and 4). The flow through the Mozambique Channel was once thought to be the major contributor to the flux of the Agulhas Current. Now it is considered to be minor (Figure 2). The entire existence of a consistent, continuous Mozambique Current, flowing along the African coastline, has in fact been called into question. The northern mouth of the Mozambique Channel is largely closed to subsurface flow by bottom ridges shallower than 2000 m, except at its western side. The surface circulation in the northern part of the channel is anticyclonic to an estimated depth of 1000 m. The eastern side of this flow, which might be the start of a Mozambique Current, is 250 km wide, has a surface speed of 0.3 m s1 and a volume flux of 6 106 m3 s1. The water masses here are characteristic of the monsoonal regime to the north with no Antarctic Intermediate Water. There is some evidence for the presence of this water mass in the central part of the channel, but there still is no North Atlantic Deep Water. The rest is essentially Subtropical Surface Water of the South Indian variety. Occasional strong flows of 2 m s1 have been observed off the African coastline in this central part of the channel, but this is extremely variable. The southern third of the channel has all the thermohaline characteristics of the South West Indian Ocean, including the presence of North Atlantic Deep Water. Net volume flux through the southern mouth of the channel seems to be very changeble. Calculations have varied from 26 106 m3 s1 southward to 5 106 m3 s1 northward, above 1000 m. The best-substantiated characteristic of the circulation in the Mozambique Channel is therefore its very high mesoscale variability. This argues against a persistent western boundary current and instead suggests a series of eddies moving southward down the channel. This scenario is consistent with most available observations and also some numerical models.

131

2000

0 10 20 Di sta 30 nc 40 e of fs ho 50 re (k 60 m ) 70

2500

3000

80

Figure 5 The spatial velocity structure of the northern Agulhas Current at Durban based on direct current measurements during a research cruise. Speeds below 1000 m have been estimated using a geostrophic calculation and are less certain. (After Duncan (1970) PhD dissertation, University of Hawaii.)

trajectory; its core meanders less than 15 km to either side. This stability is thought to be due to the strong slope of the continental shelf along which it flows. The current has a well-developed, inshore thermal front that meanders somewhat more extensively. The surface characteristics of the current are given in Table 1. Surface temperatures decrease by about 21C downstream. In the north they are at a maximum of 281C in February and a minimum of 231C in July. At Port Elizabeth, where the southern Agulhas Current starts, the maximum temperature is 251C in January, with a minimum of 211C in August. Surface salinities decrease from 35.5 PSU in the north to 35.3 PSU in the south. At Durban the core of the current usually lies 20 km offshore and penetrates to a depth of 2500 m (see Figure 5). Between the 0.5 m s1 isotachs it is 90 km wide; its offshore termination being more disperse than its strong inshore edge. Its core slopes so that at 900 mdepth it lies 65 km offshore. Its total volume flux is 73 106 m3 s1 and this increases by an estimated 6 106 m3 s1 for every 100 km distance downstream in the northern Agulhas Current. Surface speeds at its core usually lie between 1.4 and 1.6 m s1, with occasional peaks of up to 2.6 m s1.


132

AGULHAS CURRENT

Table 1

Kinematic characteristics of the upper layers of the northern Agulhas Current

Peak speed in current core (m s1) Current core offshore distance (km) Distance offshore 0.5 m s1 (km) Distance offshore 1.0 m s1 (km) Core width, between 1.0 m s1 isotachs (km) Distance offshore temperature max. (km) Distance offshore 151C/200 m intersection (km) Distance offshore 35.35 PSU salinity/200 m (km)

Mean

SD

Minimum

1.36 52 35 42 34 58 50 47

0.30 14 14 14 15 20 15 13

30 10 25 10 35 25 23

Maximum 2.45 >100 70 95 >60 >100 90 >100

After Pearce (1997), Journal of Marine Research 35(4): 731–753.

Neither these velocities nor the volume fluxes show any discernible seasonality. The northern Agulhas Current is underlain by an opposing undercurrent at 1200 m that carries 6 106 m3 s1 water equatorward at a rate of about 0.3 m s1. It consists partially of modified North Indian Intermediate Water. The invariant path of the northern Agulhas Current is interrupted during about 20% of the time by an intermittent, solitary meander – the Natal Pulse – that originates at the Natal Bight, an offset in the coast north of 301S (see Figure 6). It translates downstream at a very steady 20 km per day, continuously growing in its lateral dimensions. On its

landward side it encloses a cyclonic eddy that creates a strong coastal countercurrent as the Natal Pulse passes. This meander is triggered at the Natal Bight whenever the current intensity there exceeds a certain threshold, allowing baroclinic instability to develop away from the constraining shelf slope. The Natal Pulse is an important component of the current system since it may cause upstream retroflection at the Agulhas Plateau (see Figure 1) and may precipitate ring shedding at the Agulhas retroflection, far downstream. Flow over the shelf adjacent to the northern Agulhas Current is dependent on the shelf morphology.

Cape St Lucia

Cape St Lucia

Richards Bay

6 5 4 3 2

29˚ 4 5

1.

2

15 1 134

12 11 1098 7

Richards Bay

1.5

1.2 1.0

29˚ 0.5

1

1

S

2

S

Durban

Durban

30˚ 20 0m

1

20

0m

2

30˚

32˚

31˚

32˚

31˚

E

E

Figure 6 Current-induced upwelling in the Natal Bight off south-eastern Africa. The left panel gives the distribution of nitrate (in mmol l1) at 10 m depth and the right hand panel the simultaneous distribution of chlorophyll a. The active upwelling cell off Cape St Lucia with high nitrates and chlorophyll a is well circumscribed. (After Lutjeharms et al. (2000) Continental Shelf Research 20(14): 1907–1939. ^ Elsevier Science.)


AGULHAS CURRENT

Where the shelf is narrow, the flow is mostly parallel to the current. Over the Natal Bight, where the shelf is wider, the flow consist of cyclonic eddies. At the northern end of the Natal Bight, the current forces inshore upwelling (Figure 6). The water in this upwelling cell may be 51C colder than the adjacent current, have a high nutrient content, and exhibit enhanced biological primary productivity. The cold water thus upwelled flows over the bottom of the whole Natal Bight, strengthening the vertical layering over this shelf region. Off Durban a recurrent lee eddy is often observed. Seaward of the Agulhas Current, off the Mozambique Ridge (see Figure 1), a large number of very intense deep-sea eddies have been observed. They may be at least 2000 m deep, 100 km in diameter, have surface speeds of 1 m s1 and circular transports of between 6 106 and 18 106 m3 s1. Their lifetimes are estimated to be 1–3 years. Most of those observed are cyclonic, although a few anticyclonic ones have been seen. They seem to come from both the Mozambique Channel and from east of Madagascar, but their true origins remain unknown.

133

Meanders usually have a trailing plume and an embedded, cyclonic lee eddy (Figure 7). There is evidence that these eddies are preferentially clustered in the eastern bight of the Agulhas Bank (see Figure 1). The dimensions of all these shear edge features change markedly with distance downstream (see Table 2). Sea surface temperatures also change more readily with distance downstream here than they do in the northern Agulhas Current. Sea surface temperatures in the southern Agulhas Current reach a maximum of 261C at Algoa Bay in February; 231C off the southern tip of the Agulhas Bank. In August these temperatures are 211C and 171C, respectively. The volume flux of the Agulhas Current off the southern tip of the Agulhas Bank has been estimated at 70 106 m3 s1 down to 1500 m, i.e., about the same as that of the current to its full depth at Durban. Even though there is this increase, it seems that the Table 2 Dimensions of shear edge features along the landward edge of the southern Agulhas Current

Southern Agulhas Current In contrast to the northern Agulhas Current, the southern Agulhas Current is characterized by wide meanders as it flows past the Agulhas Bank south of Africa (see Figure 1). Meanders are present along this shelf edge at least 65% of the time. They have an average wavelength of 300 km and a phase speed that varies from 5 to 23 km per day.

Plume lengths at surface (km) Plume widths at surface (km) Diameter of enclosed eddy (km) Plume dispersion from current (km)

Port Elizabeth

Tip of Agulhas Bank

100 27 27 50

162 37 51 150

After Lutjeharms et al. (1989) Continental Shelf Research 9(7): 597–616.

Mossel Bay 17

15

16

14

17 18

16 17

0 16

18

Plume

18 18

17

14

14

50 Depth (m)

18

17 16 13

100

15

12 13 13 10

150

12 11

Eddy

200 Figure 7 Three vertical temperature sections across the Agulhas Bank picture the thermal composition of the southern Agulhas Current, a shear-edge eddy, and its associated plume. The Agulhas Current lies outside the 181C envelope. The plume has water warmer than 181C while the core of the eddy has water colder than 101C. (After Lutjeharms et al. (1989) Continental Shelf Research 9(7): 1570–1583. ^ Elsevier Science.)


134

AGULHAS CURRENT

increase per unit distance downstream found in the northern Agulhas Current is not maintained in its southern part. Water masses in the current are generally the same in the northern and the southern part. The presence of Tropical Surface Water is maintained in the southern part, as are remnants of North Indian Intermediate Water (or Red Sea Water). Tropical Surface Water is mostly found at the inshore side of the current and derives from the Mozambique Current. Its presence and volume seem to be intermittent. Some of the surface plumes generated by meanders in the far southern reaches of the Agulhas Current are advected past the western edge of the Agulhas Bank as Agulhas filaments (see Figure 1). They are present about 60% of the time and carry substantial amounts of heat that are rapidly lost to the much colder atmosphere. They also carry about 3–9 1012 kg of salt per year into the South Atlantic; salt in excess to that of the waters already present there. On average they are 50 km wide and 50 m deep. The southern Agulhas Current influences the water masses over the adjacent Agulhas Bank in three ways. First, plumes of warm Agulhas surface water may extend over the bank, heating the top layers (see Figure 7). Second, current-driven upwelling takes place off the far eastern side of the Agulhas Bank (see Figure 1) and this water flows westward and covers the greater part of the bottom of the shelf. This process cools the water column from below, leading to intense seasonal thermoclines over the Agulhas Bank. Third, most of the mean flow over the eastern part of this shelf is parallel to the current. At the southern tip of the Agulhas Bank, the current detaches from the shelf edge.

The Agulhas Retroflection The region where the Agulhas Current then terminates south of Africa is characterized by its extremely high levels of mesoscale variability (Figure 8). The measured eddy kinetic energy is higher here than in any comparable western boundary current such as the Kuroshio or the Gulf Stream. This is due to a number of dynamical traits of the current retroflection. First, the continuous progradation, or westward penetration, of the Agulhas retroflection loop into the South Atlantic Ocean causes substantial levels of variability. The loop has an average diameter of 340 km (770 km) and progrades westward at about 10 km per day. The outer limits to its movement are 101 and 211E. At its furthest extent a ring is shed by loop occlusion. This happens between 4 and 9 times per year. This ring spawning activity adds considerably to the general variability. Ring shedding events are usually preceded by the arrival of a Natal

30˚

35˚ S 40˚ 45˚ 10˚

15˚

20˚

25˚ E

30˚ 35˚ 40˚

Figure 8 The high levels of mesoscale variability that are characteristic of the Agulhas Current retroflection and the Agulhas Return Current are here portrayed by the superimposed thermal borders at the sea surface as observed for a period of one year. The inshore border of the northern Agulhas Current is particularly stable, that of the southern Agulhas Current less so, whereas the Agulhas Return Current exhibits a tendency to prefer certain meanders. The Agulhas retroflection has a range of locations and is attended by a host of rings (to the north) and eddies (to the south). (After Lutjeharms and van Ballegooyen (1988) Journal of Physical Oceanography 18(11): 1570–1580.)

Pulse on the Agulhas Current. The average lag time between initiation of a Natal Pulse off the Natal Bight and the shedding of a ring at the retroflection is 165 days. Between a newly formed ring and the reconstituted retroflection loop, a wedge of Subtropical Surface Water usually penetrates northward. Its water has a temperature of 171C and salinity lower than 34.9 over the top 100 m, also adding to the variability of the region. Newly formed rings retain the hydrographic and kinematic characteristics of the southern Agulhas Current (Table 3). Since the heat loss from such a ring may be between 80 and 160 W m2 and since there is substantial evaporation in the region, the temperature and salinity of the upper layers of the features are considerably modified near the retroflection region. Rings are about 320 km (7100 km) in diameter at the sea surface. Estimated by the location of their maximum azimuthal speeds, the diameters are a reduced 240 km (740 km). These radial speeds lie between 0.3 and 0.9 m s1. The mean depth of the 101C isotherm in Agulhas rings, a proxy for the geostrophic speed of their water masses, is 650 m (7130 m). Further properties, as estimated by a number of investigators, are given in Table 4. Rings move off into the South Atlantic Ocean at speeds of 4–8 km per day. Maximum translation rates of up to 16 km per day have been observed. They lose 50% of their sea surface height – and therefore by inference of their energy – during the


AGULHAS CURRENT

first 4 months of their lifetime. A full 40% of rings never seem to leave the Cape Basin, off the southeastern coast of Africa, at all but totally disintegrate here. This decay may well be enhanced by the splitting of rings. This process seems to be largely induced by rings passing over prominent features of the bottom topography such as seamounts. This rapid dissipation of these features means that a considerable part of all the excess salt, heat, energy, and vorticity carried by the rings is deposited exclusively in this corner of the South Atlantic. The remaining rings seem to have lifetimes between 2 and 3 years. They move westward across the full width of the South Atlantic Ocean, slightly to the left of the general background flow. A few of them interact with the upwelling front off the south-eastern coast of Africa, with upwelling filaments occasionally Table 3 Thermohaline characteristics of the principal water masses found at the Agulhas Current retroflection and vicinity Temperature range (1C) Surface Water Central Water South East Atlantic Ocean South West Indian Ocean Antarctic Intermediate Water South East Atlantic Ocean South West Indian Ocean Deep Water North Atlantic Deep Water (SE Atlantic) Circumpolar Deep Water (SW Indian) Antarctic Bottom Water

being wrapped around passing rings. These Agulhas rings play a crucial role in the interbasin exchange of waters between the South Indian and South Atlantic Oceans. This is partially quantified in Table 5.

Agulhas Return Current That part of the Agulhas Current not involved in ring production flows back in an easterly direction on having successfully negotiated the retroflection. Here also there are very high levels of mesoscale variability with substantial meandering and eddies being shed to both sides of the Agulhas Return Current/ Subtropical Convergence. Much of this meandering

Table 5 Estimates of interbasin volume transport between the South Indian and the South Atlantic Oceans caused by ring shedding. The values were calculated by the investigators named here. For full references see De Ruijter et al. (1999) Investigators

Volume transport per ring (106m3s1)

Referenced to

Olson and Evans (1986) Duncombe Rae et al. (1989) Gordon and Haxby (1990)

0.5–0.6 1.2 1.0–1.5 2.0–3.0 0.4–1.1

T > 101C Total T > 101C Total Total

1.1 0.8–1.7 0.45–0.90 0.65 1.0

T > 101C 1000 db 1500 db Total T > 101C

Salinity range (PSU)

16.0 to 26.0

>35.50

6.0 to 16.0 8.0 to 15.0

34.50 to 35.50 34.60 to 35.50

2.0 to 6.0 2.0 to 10.0

33.80 to 34.80 33.80 to 34.80

1.5 to 4.0

34.80 to 35.00

0.1 to 2.0

34.63 to 34.73

0.9 to 1.7

135

McCartney and Woodgate-Jones (1991) Van Ballegooyen et al. (1994) Byrne et al. (1995) Clement and Gordon (1995) Duncombe Rae et al. (1996) Goni et al. (1997)

34.63 to 34.72

After Valentine et al. (1993) Deep-Sea Research 40(6): 1285–1305. ^ Elsevier Science.

After De Ruijter WPM et al. (1999) Indian–Atlantic inter-ocean exchange: dynamics, estimation and impact. Journal of Geophysical Research 104(C9): 20885–20911. ^ American Geophysical Union.

Table 4 Physical properties of Agulhas rings as furnished by a number of independent investigators, calculated with respect to the characteristics of water in the South East Atlantic Ocean Investigators

Olson and Evans (1986) Duncombe Rae et al. (1989) Duncombe Rae et al. (1992) Van Ballegooyen et al. (1994) Byrne et al. (1995) Clement and Gordon (1995) Duncombe Rae et al. (1996) Goni et al. (1997) Garzoli et al. (1996)

Heat flux (103 PW)

Salt flux (105 kg/s)

25

6.3

7.5

Available potential energy (1015 J)

Kinetic energy (1015 J)

30.5–51.4

6.2–8.7

38.8

2.3

18 7.0 11.3 24 2.8–3.8

4.5 7.0 2.0

4.2

1.7

1.1

1.0–1.6

0.7–1.0

After De Ruijter WPM et al. (1999) Indian–Altantic inter-ocean exchange: dynamics, estimation and impact. Journal of Geophysical Research 104(C9): 20885–29911. ^ American Geophysical Union, where full references can be found.


AGULHAS CURRENT

velocity has reduced to 0.4 m s1 and the volume flux to about 19 106 m3 s1. By a longitude of 701E, at the Kerguelen Plateau, little of the Agulhas characteristics remain along the Subtropical Convergence, all of the Agulhas water having leaked off into the South West Indian Ocean subgyre. The speed, volume transport, as well as mesoscale variability associated with the Agulhas Return Current will all have declined here to values observed in the southeastern Atlantic Ocean, upstream of any influence from the Agulhas Current. The Agulhas Return Current can therefore be considered to have terminated here. The continuing flow along the Subtropical Convergence east of here is known as the South Indian Ocean Current.

80 _1

is brought about by the variable bathymetry over which the current has to pass. The first obstacle to a purely zonal flow for the Agulhas Return Current is the Agulhas Plateau (see Figure 1). Here it carries out a northward meander of 290765 km. Cold eddies are frequently formed here with diameters of 280750 km, but rapidly warm to become indistinguishable from ambient surface waters. Warm eddies may in turn be shed to the south. One such eddy that has been observed closely remained in roughly the same position for 2 months, rotated every 3 days, and had a volume flux of 32 106 m3 s1 to a depth of 1500 m. Downstream of the Agulhas Plateau the next meander lies at a distance of 4507110 km. These meanders move upstream at about half a wavelength per season. Upstream of the Plateau there are also westwardpropagating Rossby waves on the Agulhas Return Current/Subtropical Convergence. However, the direct correlation between the Agulhas Return Current and the Subtropical Convergence is not always straightforward. Sometimes they are in close juxtaposition, sometimes not. When they are not, two separate fronts, or even multiple fronts, may be formed (Table 6). The Agulhas Current has a considerable effect on the Subtropical Convergence, forcing it to lie 51 of latitude farther south than in the South Atlantic Ocean and increasing its surface gradients so that a meridional gradient of 51C in 35 km is not unknown. The Agulhas Return Current starts off with characteristics nearly identical to those of the Agulhas Current. South of Africa it may exhibit a surface speed of 1.3 m s1 and a volume transport of 40 106 m3 s1 to 1000 m. The parallel flow along the Subtropical Convergence would be 16–20 106 m3 s1 at the same time. All these flow characteristics decrease rapidly as the current progresses eastward (Figure 9). By about 551E, the surface

Volume transport (106 m3 s )

136

AJAX SCARC ARC Marathon Discovery SUZIL

70 60 50 40 30 20 10 0

0

10

20

30

40

50

60

70

80

E

Figure 9 The nature of the volume flux along the Subtropical Convergence south of Africa, as established by a number of individual research cruises (shown in the box). Transport is in units of 106 m3 s1 and has been calculated to a depth of 1500 m. The influence of the Agulhas Current is felt from 201E eastwards, but by the 701E meridian it has been dissipated completely. This may therefore be considered the termination of the Agulhas Return Current and the start of the South Indian Ocean Current. (After Lutjeharms and Ansorge Journal of Marine Systems, in press.)

Table 6 The geographic location and the thermal characteristics of the surface expressions of the Agulhas Front as well as the Subtropical Convergence south of Africa. Values for the Agulhas Front were based on 24 crossings, that of the Subtropical Convergence on 70. Values in parentheses denote standard deviations for the calculated averages Latitudinal position From

Agulhas Front Subtropical Convergence

391090 (011160 ) 401350 (011230 )

To

401010 (011060 ) 421360 (011320 )

Temperature Middle

Width (km)

From

To

Middle

Range

Gradient (C km1)

391370 (011140 ) 411400 (011190 )

96.3 69.1 225.1 140.6

21.0 (1.6) 17.9 (2.1)

15.7 (1.5) 10.6 (1.8)

18.4 (1.2) 14.2 (1.7)

5.4 (1.6) 7.3 (1.9)

0.102 0.106 0.047 (0.043)

After Lutjeharms and Valentine (1984) Deep-Sea Research, 31(12): 1461–1476. ^ Elsevier Science.


AGULHAS CURRENT

Conclusion

137

Further Reading

The Agulhas Current is unusual as a western boundary current for a number of reasons. First, because the African continent terminates at relatively low latitudes, the current penetrates freely into the adjacent ocean basin and a substantial leakage between basins is feasible. Second, through the process of ring and filament shedding, an interaction between a western boundary current and an extensive coastal upwelling regime is brought about that is geographically not possible elsewhere. Third, the very stable nature of the northern Agulhas Current and its characteristic Natal Pulse creates a dynamic environment in which mesoscale disturbances can have profound circulatory effects downstream. The contemporary ignorance about the East Madagascar Current, about the circulation of the Mozambique Channel, and about the origin of midocean eddies in the South West Indian Ocean needs to be eliminated. Only then will a more realistic concept of the interactions between elements of the greater Agulhas Current system become possible.

De Ruijter WPM, Biastoch A, and Drijfhout SS et al. (1999) Indian–Atlantic inter-ocean exchange: dynamics, estimation and impact. Journal of Geophysical Research 104: 20885--20911. Lutjeharms JRE (1996) The exchange of water between the South Indian and the South Atlantic. In: Wefer G, Berger WH, Siedler G, and Webb D (eds.) The South Atlantic: Present and Past Circulation, pp. 125--162. Berlin: Springer-Verlag. Lutjeharms JRE (2001) The Agulhas Current. Berlin: Springer-Verlag. Shannon LV (1985) The Benguela Ecosystem. 1. Evolution of the Benguela, physical features and processes. Oceanography and Marine Biology, An Annual Review 23: 105--182. Shannon LV and Nelson G (1996) The Benguela: large scale features and processes and system variability. In: Wefer G, Berger WH, Siedler G, and Webb D (eds.) The South Atlantic: Present and Past Circulation, pp. 163--210. Berlin: Springer-Verlag.

See also Mesoscale Eddies. Arctic Ocean Circulation. Water Types and Water Masses


AIRCRAFT REMOTE SENSING L. W. Harding, Jr and W. D. Miller, University of Maryland, College Park, MD, USA R. N. Swift, and C. W. Wright, NASA Goddard Space Flight Center, Wallops Island, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 113–122, & 2001, Elsevier Ltd.

Introduction The use of aircraft for remote sensing has steadily grown since the beginnings of aviation in the early twentieth century and today there are many applications in the Earth sciences. A diverse set of remote sensing uses in oceanography developed in parallel with advances in aviation, following increased aircraft capabilities and the development of instrumentation for studying ocean properties. Aircraft improvements include a greatly expanded range of operational altitudes, development of the Global Positioning System (GPS) enabling precision navigation, increased availability of power for instruments, and longer range and duration of missions. Instrumentation developments include new sensor technologies made possible by microelectronics, small, high-speed computers, improved optics, and increased accuracy of digital conversion of electronic signals. Advances in these areas have contributed significantly to the maturation of aircraft remote sensing as an oceanographic tool. Many different types of aircraft are currently used for remote sensing of the oceans, ranging from balloons to helicopters, and from light, single engine piston-powered airplanes to jets. The data and information collected on these platforms are commonly used to enhance sampling by traditional oceanographic methods, giving increased spatial and temporal resolution for a number of important properties. Contemporary applications of aircraft remote sensing to oceanography can be grouped into several areas, among them ocean color, sea surface temperature (SST), sea surface salinity (SSS), wave properties, near-shore topography, and bathymetry. Prominent examples include thermal mapping using infrared (IR) sensors in both coastal and open ocean studies, lidar and visible radiometers for ocean color measurements of phytoplankton distributions and ‘algal blooms’, and passive microwave sensors to make observations of surface salinity structure of estuarine plumes in the coastal ocean. These topics

138

will be discussed in more detail in subsequent sections. Other important uses of aircraft remote sensing are to test instruments slated for deployment on satellites, to calibrate and validate space-based sensors using aircraft-borne counterparts, and to make ‘under-flights’ of satellite instruments and assess the efficacy of atmospheric corrections applied to data from space-based observations. Aircraft have some advantages over satellites for oceanography, including the ability to gather data under cloud cover, high spatial resolution, flexibility of operations that enables rapid responses to ‘events’, and less influence of atmospheric effects that complicate the processing of satellite data. Aircraft remote sensing provides nearly synoptic data and information on important oceanographic properties at higher spatial resolution than can be achieved by most satellite-borne instruments. Perhaps the greatest advantage of aircraft remote sensing is the ability to provide consistent, high-resolution coverage at larger spatial scales and more frequent intervals than are practical with ships, making it feasible to use aircraft for monitoring change. Disadvantages of aircraft remote sensing include the relatively limited spatial coverage that can be obtained compared with the global coverage available from satellite instruments, the repeated expense of deploying multiple flights, weather restrictions on operations, and lack of synopticity over large scales. Combination of the large-scale, synoptic data that are accessible from space with higher resolution aircraft surveys of specific locations is increasingly recognized as an important and useful marriage that takes advantages of the strengths of both approaches. This article begins with a discussion of sensors that use lasers (also called active sensors), including airborne laser fluorosensors that have been used to measure chlorophyll (chl-a) and other properties; continues with discussions of lidar sensors used for topographic and bathymetric mapping; describes passive (sensors that do not transmit or illuminate, but view naturally occurring reflections and emissions) ocean color remote sensing directed at quantifying phytoplankton biomass and productivity; moves to available or planned hyper-spectral aircraft instruments; briefly describes synthetic aperture radar applications for waves and wind, and closes with a discussion of passive microwave measurements of salinity. Readers are directed to the Further Reading section if they desire additional information on individual topics.


AIRCRAFT REMOTE SENSING

Active Systems Airborne Laser Fluorosensing

The concentrations of certain waterborne constituents, such as chl-a, can be measured from their fluorescence, a relationship that is exploited in shipboard sensors such as standard fluorometers and flow cytometers that are discussed elsewhere in this encyclopedia. NASA first demonstrated the measurement of laser-induced chl-a fluorescence from a low-flying aircraft in the mid-1970s. Airborne laser fluorosensors were developed shortly thereafter in the USA, Canada, Germany, Italy, and Russia, and used for measuring laser-induced fluorescence of a number of marine constituents in addition to chl-a. Oceanic constituents amenable to laser fluorosensing include phycoerythrin (photosynthetic pigment in some phytoplankton taxa), chromophoric dissolved organic matter (CDOM), and oil films. Airborne laser fluorosensors have also been used to follow dyes such as fluorescein and rhodamine that are introduced into water masses to trace their movement. The NASA Airborne Oceanographic Lidar (AOL) is the most advanced airborne laser fluorosensor. The transmitter portion features a dichroic optical device to spatially separate the temporally concurrent 355 and 532 nm pulsed-laser radiation, followed by individual steering mirrors to direct the separated beams to respective oceanic targets separated by B1 m when flown at the AOL’s nominal 150 m operational altitude. The receiver focal plane containing both laser-illuminated targets is focused onto the input slits of the monochromator. The monochromator output focal planes are viewed by custommade optical fibers that transport signal photons from the focal planes to the photo-cathode of each photo-multiplier module (PMM) where the conversion from photons to electrons takes place with a substantial gain. Time-resolved waveforms are collected in channels centered at 404 nm (water Raman) and 450 nm (CDOM) from 355 nm laser excitation, and at 560 nm and 590 nm (phycoerythrin), 650 (water Raman), and 685 nm (chl-a) from the 532 nm laser excitation. The water Raman from the respective lasers is the red-shifted emission from the OH bonds of water molecules resulting from radiation with the laser pulse. The strength of the water Raman signal is directly proportional to the number of OH molecules accessed by the laser pulse. Thus, the water Raman signal is used to normalize the fluorescence signals to correct for variations in water attenuation properties in the surface layer of the ocean. The AOL is described in more detail on http:// lidar.wff.nasa.gov.

139

The AOL has supported major oceanographic studies throughout the 1980s and 1990s, extending the usefulness of shipboard measurements over wide areas to permit improved interpretation of the shipderived results. Examples of data from the AOL show horizontal structure of laser-induced fluorescence converted to chl-a concentration (Figure 1). Prominent oceanographic expeditions that have benefited from aircraft coverage with the AOL include the North Atlantic Bloom Experiment (NABE) of the Joint Global Ocean Flux Study (JGOFS), and the Iron Enrichment Experiment (IRONEX) of the Equatorial Pacific near the Galapagos Islands. This system has been flown on NASA P-3B aircraft in open-ocean missions that often exceeded 6 h in duration and collected hundreds of thousands of spectra. Each ‘experiment’ is able to generate both active and passive data in ‘pairs’ that are used for determining ocean color and recovering chl-a and other constituents, and that are also useful in the development of algorithms for measuring these constituents from oceanic radiance spectra. Airborne Lidar Coastal Mapping

The use of airborne lidar (light detection and ranging) sensors for meeting coastal mapping requirements is a relatively new and promising application of laser-ranging technology. These applications include high-density surveying of coastline and beach morphology and shallow water bathymetry. The capability to measure distance accurately with lidar sensors has been available since the early 1970s, but their application to airborne surveying of terrestrial features was seriously hampered by the lack of knowledge of the position of the aircraft from which the measurement was made. The implementation of the Department of Defense GPS constellation of satellites in the late 1980s, coupled with the development of GPS receiver technology, has resulted in the capability to provide the position of a GPS antenna located on an aircraft fuselage in flight to an accuracy approaching 5 cm using kinematic differential methodology. These methods involve the use of a fixed receiver (generally located at the staging airport) and a mobile receiver that is fixed to the aircraft fuselage. The distance between mobile and fixed receivers, referred to as the baseline, is typically on the order of tens of kilometers, and can be extended to hundreds of kilometers by using dual frequency survey grade GPS receivers aided by tracking the phase code of the carrier from each frequency. Modern airborne lidars are capable of acquiring 5000 or more discrete range measurements per second. Aircraft attitude and heading information are


140


Figure 1 Cross-section profiles flown across a large oceanic front west of the Galapagos Islands on 25 October and 3 November 1993. (A) A dramatic horizontal displacement of the front as measured with an infrared radiometer; (B) and (C) show corresponding changes in laser-induced fluorescence of chl-a and phycoerythrin. This flight was made as part of the original IRONEX investigation in late 1993. NFU, normalized fluorescence units.

used along with the GPS-determined platform position to locate the position of the laser pulse on the Earth’s surface to a vertical accuracy approaching 10 cm with some highly accurate systems and o30 cm for most of these sensors. The horizontal accuracy is generally 50–100 cm. Depending on the pulse repetition rate of the laser transmitter, the off-nadir pointing angle, and the speed of the aircraft platform, the density of survey points can exceed one sample per square meter. NASA’s Airborne Topographic Mapper (ATM) is an example of a topographic mapping lidar used for coastal surveying applications. An example of data from ATM shows shoreline features off the east coast of the USA (Figure 2). ATM was originally developed to measure changes in the elevation of Arctic ice sheets in response to global warming. The sensor was applied to measurements of changes in coastal morphology beginning in 1995. Presently, baseline

topographic surveys exist for most of the Atlantic and Gulf coasts between central Maine and Texas and for large sections of the Pacific coast. Affected sections of coastline are re-occupied following major coastal storms, such as hurricanes and ‘Nor’easters’, to determine the extent of erosion and depositional patterns resulting from the storms. Additional details on the ATM and some results of investigations on coastal morphology can be found on websites (http:// lidar.wff.nasa.gov and http://aol.wff.nasa.gov/aoltm/ projects/beachmap/98results/). Other airborne lidar systems have been used to survey coastal morphology, including Optec lidar systems by Florida State University and the University of Texas at Austin. Beyond these airborne lidar sensors, there are considerably more instruments with this capability that are currently in use in the commercial sector for surveying metropolitan areas, flood plains, and for other terrestrial



141

Figure 2 (Left) Map of coastal topography around Pacifica, CA, USA, derived from lidar data obtained in April 1998, after a winter of severe storms associated with El Ninõ. Map insert shows the Esplanade Drive area of Pacifica rendered from lidar data gridded at 2 m resolution and colored according to elevation. (Right) Cross-sections derived from lidar data of October 1997 and April 1998 at locations marked in the map inset. The profiles in (A) show a stable cliff and accreting beach, whereas about 200 m to the south the profiles in (B) show erosion of the sea cliff and adjacent beach resulting in undermining of houses. Each profile shows individual laser spot elevations that fall within a 2 m wide strip oriented approximately normal to shoreline. (Reproduced with permission from Sallenger et al., 1999.)

applications. At the last count (early 2000) there were approximately 60 airborne lidars in operation worldwide, with most engaged in a variety of survey applications generally outside the field of coastal mapping. Pump and Probe Fluorometry

Several sections in this chapter describe recoveries of phytoplankton biomass as chl-a by active and passive measurements. Another recent accomplishment is an active, airborne laser measurement intended to aid in remote detection of photosynthetic performance, an important ingredient of primary productivity computations. Fluorometric techniques, such as fast repetition rate (FRR) fluorometry (explained in another article of this encyclopedia), have provided an alternative approach to 14C assimilation and O2 evolution in the measurement of primary productivity. This

technology has matured with the commercial availability of FRR instruments that can give vertical profiles or operate in a continuous mode while underway. There have been several attempts to develop airborne lidar instruments to determine phytoplankton photosynthetic characteristics from aircraft in the past decade. NASA scientists have deployed a pump and probe fluorometer from aircraft, wherein the AOL laser (described above) acts as the pump and a second laser with variable power options and rapid pulsing capabilities (10 ns) functions as the probe.

Passive Systems Multichannel Ocean Color Sensor (MOCS)

Passive ocean color measurements using visible radiometers to measure reflected natural sunlight from the ocean have been made with a number of


142


instruments in the past two decades. These instruments include the Multichannel Ocean Color Sensor (MOCS) that was flown in studies of Nantucket Shoals in the early 1980s, the passive sensors of the AOL suite that have been used in many locations around the world, and more recently, simple radiometers that have been deployed on light aircraft in regional studies of Chesapeake Bay (see below). MOCS was one of the earliest ocean color sensors used on aircraft. It provided mesoscale data on shelf and slope chl-a in conjunction with shipboard studies of physical structure, nutrient inputs, and phytoplankton primary productivity. Ocean Data Acquisition System (ODAS) and SeaWiFS Aircraft Simulator (SAS)

Few aircraft studies have obtained long time-series sufficient to quantify variability and detect secular trends. An example is ocean color measurements made from light aircraft in the Chesapeake Bay region for over a decade, providing data on chl-a and SST from 4250 flights. Aircraft over-flights of the Bay using the Ocean Data Acquisition System (ODAS) developed at NASA’s Goddard Space Flight Center commenced in 1989. ODAS was a nadirviewing, line-of-flight, three-band radiometer with spectral coverage in the blue-green region of the visible spectrum (460–520 nm), a narrow 1.51 fieldof-view, and a 10 Hz sampling rate. The ODAS instrument package included an IR temperature sensor (PRT-5, Pyrometrics, Inc.) for measuring SST. The system was flown for B7 years over Chesapeake Bay on a regular set of tracks to determine chl-a and SST. Over 150 flights were made with ODAS between 1989 and 1996, coordinated with in situ observations from a multi-jurisdictional monitoring program and other cruises of opportunity. ODAS was flown together with the SeaWiFS Aircraft Simulator (SAS II, III, Satlantic, Inc., Halifax, Canada) beginning in 1995 and was retired soon thereafter and replaced with the SAS units. SAS III is a multi-spectral (13-band, 380–865 nm), line-offlight, nadir viewing, 10 Hz, passive radiometer with a 3.51 field-of-view that has the same wavebands as the SeaWiFS satellite instrument, and several additional bands in the visible, near IR, and UV. The SAS systems include an IR temperature sensor (Heimann Instruments, Inc.). Chl-a estimates are obtained using a curvature algorithm applied to water-leaving radiances at wavebands in the blue-green portion of the visible spectrum with validation from concurrent shipboard measurements. Flights are conducted at B50–60 m s1 (100–120 knots), giving an alongtrack profile with a resolution of 5–6 m averaged to

50 m in processing, and interpolated to 1 km2 for visualization. Imagery derived from ODAS and SAS flights is available on a web site of the NOAA Chesapeake Bay Office for the main stem of the Bay (http://noaa.chesapeakebay.net), and for two contrasting tributaries, the Choptank and Patuxent Rivers on a web site of the Coastal Intensive Sites Network (CISNet) (http://www.cisnet-choptank.org). Data from ODAS and SAS have provided detailed information on the timing, position, and magnitude of blooms in Chesapeake Bay, particularly the spring diatom bloom that dominates the annual phytoplankton cycle. This April–May peak of chl-a represents the largest accumulation of phytoplankton biomass in the Bay and is a proximal indicator of over-enrichment by nutrients. Data from SeaWiFS for spring 2000 show the coast-wide chl-a distribution for context, while SAS III data illustrate the high-resolution chl-a maps that are obtained regionally (Figure 3). A well-developed spring bloom corresponding to a year of relatively high freshwater flow from the Susquehanna River, the main tributary feeding the estuary, is apparent in the main stem Bay chl-a distribution. Estimates of primary productivity are now being derived from shipboard observations of key variables combined with high-resolution aircraft measurements of chl-a and SST for the Bay.

Hyper-spectral Systems Airborne Visible/Infrared Imaging Spectrometer (AVIRIS)

The Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) was originally designed in the late 1980s by NASA at the Jet Propulsion Laboratory (JPL) to collect data of high spectral and spatial resolution, anticipating a space-based high-resolution imaging spectrometer (HIRIS) that was planned for launch in the mid-1990s. Because the sensor was designed to provide data similar to satellite data, flight specifications called for both high altitude and high speed. AVIRIS flies almost exclusively on a NASA ER-2 research aircraft at an altitude of 20 km and an airspeed of 732 km h1. At this altitude and a 301 field of view, the swath width is almost 11 km. The instantaneous field of view is 1 mrad, which creates individual pixels at a resolution of 20 m2. The sensor samples in a whiskbroom fashion, so a mirror scans back and forth, perpendicular to the line-of-flight, at a rate of 12 times per second to provide continuous spatial coverage. Each pixel is then sent to four separate spectrometers by a fiber optic cable. The spectrometers are arranged so that they each cover a part of the



Chesapeake Bay chlorophyll 14 May 2000

Mid-Atlantic chlorophyll 8 May 2000 64.0

_ Chl [mg m 3]

39˚30'

40 39

10.0

80 60

38 39˚00'

1.0

40

37

30

0.1

36

25

38˚30'

35 0.01

34

143

76

74

72

20

70 16

(A)

12

38˚00'

10 8 37˚30'

6 4 2

37˚00' 77˚00'

(B)

76˚30'

76˚00'

75˚30'

Patuxent chlorophyll 1 May 2000 38˚35' Choptank chlorophyll 1 May 2000 38˚45'

38˚30'

38˚40'

38˚25'

38˚35'

38˚20'

(C)

76˚40'

76˚30'

(D)

76˚20'

76˚10'

76˚00'

Figure 3 Spring chl-a (mg m3) in: (A) the mid-Atlantic region from SeaWiFS; (B) Chesapeake Bay; (C) Patuxent R; (D) Choptank R from SAS III.

spectrum from 0.40 to 2.4 mm, providing continuous spectral coverage at 10 nm intervals over the entire spectrum from visible to near IR. Data are recorded to tape cassettes for storage until rectification, atmospheric correction, and processing at JPL. A typical AVIRIS ‘scene’ is a 40 min flight line. At ER-2

flight parameters, this creates an image roughly 500 km long and 11 km wide. Data are encoded at 12-bits for a high degree of discrimination. The physical dimensions of AVIRIS are quite large, 84 cm wide 160 cm long 117 cm tall at a weight of 720 pounds.


144


Data collected with AVIRIS have been used for terrestrial, marine, and atmospheric applications. Accomplishments of AVIRIS include separation of the chl-a signature from bottom reflectance for clear lake waters of Lake Tahoe and turbid waters near Tampa Bay, interpretation of spectral signals from resuspended sediment and dissolved organic materials in W. Florida, and of suspended sediment and kelp beds in S. California. Recent efforts have focused on improving atmospheric correction procedures for both AVIRIS and satellite data, providing inputs for bio-optical models which determine inherent optical properties (IOPs) from reflectance, algorithm development, and sporadic attempts at water quality monitoring (e.g., chl-a, suspended sediment, diffuse attenuation coefficient, kd). AVIRIS data have recently been used as an input variable to a neural network model developed to estimate water depth. The model was able to separate the contributions of different components to the total waterleaving radiance and to provide relatively accurate estimates of depth (rms error ¼ 0.48 m). Compact Airborne Spectrographic Imager (CASI)

The Compact Airborne Spectrographic Imager (CASI) is a relatively small, lightweight hyper-spectral sensor that has been used on a variety of light aircraft. CASI was developed by Itres Research Ltd (Alberta, Canada) in 1988 and was designed for a variety of remote sensing applications in forestry, agriculture, land-use planning, and aquatic monitoring. By allowing user-defined configurations, the 12-bit, push-broom-type sensor (333 scan lines per second) using a charge-coupled detector (CCD) can be adapted to maximize either spatial (37.81 across track field of view, 0.0771 along-track, 512 pixels – pixel size varies with altitude) or spectral resolution (288 bands at 1.9 nm intervals between 400 and 1000 nm). Experiments have been conducted using CASI to determine bottom type, benthic cover, submerged aquatic vegetation, marsh type, and in-water constituents such as suspended sediments, chl-a, and other algal pigments. Portable Hyper-spectral Imager for Low-Light Spectroscopy (PHILLS)

The Portable Hyper-spectral Imager for Low-Light Spectroscopy (PHILLS) has been constructed by the US Navy (Naval Research Laboratory) for imaging the coastal ocean. PHILLS uses a backside-illuminated CCD for high sensitivity, and an all-reflective spectrograph with a convex grating in an Offner configuration to produce a distortion-free image. The instrument benefits from improvements in large-

format detector arrays that have enabled increased spectral resolution and higher signal-to-noise ratios for imaging spectrographs, extending the use of this technology in low-albedo coastal waters. The ocean PHILLS operates in a push-broom scanned mode whereby cross-track ground pixels are imaged with a camera lens onto the entrance slit of the spectrometer, and new lines of the along-track ground pixels are attained by aircraft motion. The Navy’s interest in hyper-spectral imagers for coastal applications centers on the development of methods for determining shallow water bathymetry, topography, bottom type composition, underwater hazards, and visibility. PHILLS precedes a planned hyper-spectral satellite instrument, the Coastal Ocean Imaging Spectrometer (COIS) that is planned to launch on the Naval Earth Map Observer (NEMO) spacecraft.

Radar Altimetry Ocean applications of airborne radar altimetry systems include several sensors that retrieve information on wave properties. Two examples are the Radar Ocean Wave Spectrometer (ROWS), and the Scanning Radar Altimeter (SRA), systems designed to measure long-wave directional spectra and near-surface wind speed. ROWS is a Ku-band system developed at NASA’s Goddard Space Flight Center in support of present and future satellite radar missions. Data obtained from ROWS in a spectrometer mode are used to derive two-dimensional ocean spectral wave estimates and directional radar backscatter. Data from the pulse-limited altimeter mode radar yield estimates of significant wave height and surface wind speed.

Synthetic Aperture Radar (SAR) Synthetic Aperture Radar (SAR) systems emit microwave radiation in several bands and collect the reflected radiation to gain information about sea surface conditions. Synthetic aperture is a technique that is used to synthesize a long antenna by combining signals, or echoes, received by the radar as it moves along a flight track. Aperture refers to the opening that is used to collect reflected energy and form an image. The analogous feature of a camera to the aperture would be the shutter opening. A synthetic aperture is constructed by moving a real aperture or antenna through a series of positions along a flight track. NASA’s Jet Propulsion Laboratory and Ames Research Center have operated the Airborne SAR (AIRSAR) on a DC-8 since the late 1980s. The radar



of AIRSAR illuminates the ocean at three microwave wavelengths: C-band (6 cm), L-band (24 cm), and Pband (68 cm). Brightness of the ocean (the amount of energy reflected back to the antenna) depends on the roughness of the surface at the length scale of the microwave (Bragg scattering). The primary source of roughness, and hence brightness, at the wavelengths used is capillary waves associated with wind. Oceanographic applications derive from the responsiveness of capillary wave amplitude to factors that affect surface tension, such as swell, atmospheric stability, and the presence of biological films. For example, the backscatter characteristics of the ocean are affected by surface oil and slicks can be observed in SAR imagery as a decrease of radar backscatter; SAR imagery appears dark in an area affected by an oil spill, surface slick, or biofilm, as compared with areas without these constituents.

Microwave Salinometers Passive microwave radiometry (L-band) has been tested for the recovery of SSS from aircraft, and it may be possible to make these measurements from space. Salinity affects the natural emission of EM radiation from the ocean, and the microwave signature can be used to quantify SSS. Two examples of aircraft instruments that have been used to measure SSS in the coastal ocean are the Scanning Low-Frequency

145

Microwave Radiometer (SLFMR), and the Electronically Thinned Array Radiometer (ESTAR). SLFMR was used recently in the estuarine plume of Chesapeake Bay on the east coast of the USA to follow the buoyant outflow that dominates the nearshore density structure and constitutes an important tracer of water mass movement. SLFMR is able to recover SSS at an accuracy of about 1 PSU (Figure 4). This resolution is too coarse for the open ocean, but is quite suitable for coastal applications where significant gradients occur in regions influenced by freshwater inputs. SLFMR has a bandwidth of 25 MHz, a frequency of 1.413 GHz, and a single antenna with a beam width of approximately 161 and six across-track positions at 761, 7221 and 7391. Tests of SLFMR off the Chesapeake Bay demonstrated its effectiveness as a ‘salinity mapper’ by characterizing the trajectory of the Bay plume from surveys using light aircraft in joint operations with ships. Flights were conducted at an altitude of 2.6 km, giving a resolution of about 1 km. The accuracy of SSS in this example is B0.5 PSU. ESTAR is an aircraft instrument that is the prototype of a proposed space instrument for measuring SSS. This instrument relies on an interferometric technique termed ‘aperture synthesis’ in the across-track dimension that can reduce the size of the antenna aperture needed to monitor SSS from space. It has been described as a ‘hybrid of a real and a synthetic aperture radiometer.’ Aircraft surveys of

(PSU)

Figure 4 Sea surface salinity from an airborne microwave salinity instrument for (A) 14 September 1996; (B) 20 September 1996. Images reveal strong onshore-offshore gradients in salinity from the mouth of Chesapeake Bay to the plume and shelf, and the effect of high rainfall and freshwater input on the salinity distribution over a 1-week interval. (Adapted with permission from Miller et al., 1998.)


146


SSS using ESTAR in the coastal current off Maryland and Delaware showed good agreement with thermosalinograph measurements from ships in the range of 29–31 PSU.

See also Beaches, Physical Processes Affecting. Bio-Optical Models. Fluorometry for Biological Sensing. Inherent Optical Properties and Irradiance. IR Radiometers. Iron Fertilization. Optical Particle Characterization. Phytoplankton Blooms. Primary Production Distribution. River Inputs. Satellite Oceanography, History and Introductory Concepts. Satellite Remote Sensing: Ocean Color. Satellite Remote Sensing of Sea Surface Temperatures. Satellite Remote Sensing: Salinity Measurements. Upper Ocean Time and Space Variability.

Further Reading Blume H-JC, Kendall BM, and Fedors JC (1978) Measurement of ocean temperature and salinity via

microwave radiometry. Boundary Layer Meteorology 13: 295--308. Campbell JW and Esaias WE (1985) Spatial patterns in temperature and chlorophyll on Nantucket Shoals from airborne remote sensing data, May 7–9, 1981. Journal of Marine Research 43: 139--161. Harding LW Jr, Itsweire EC, and Esaias WE (1994) Estimates of phytoplankton biomass in the Chesapeake Bay from aircraft remote sensing of chlorophyll concentrations, 1989–92. Remote Sensing Environment 49: 41--56. Le Vine DM, Zaitzeff JB, D’Sa EJ, et al. (2000) Sea surface salinity: toward an operational remote-sensing system. In: Halpern D (ed.) Satellites, Oceanography and Society, pp. 321--335. Elsevier Science. Miller JL, Goodberlet MA, and Zaitzeff JB (1998) Airborne salinity mapper makes debut in coastal zone. EOS Transactions of the American Geophysical Union 79: 173--177. Sallenger AH Jr, Krabill W, Brock J, et al. (1999) EOS Transactions of the American Geophysical Union 80: 89--93. Sandidge JC and Holyer RJ (1998) Coastal bathymetry from hyper-spectral observations of water radiance. Remote Sensing Environment 65: 341--352.


AIR–SEA GAS EXCHANGE a wavy water surface are still not known. A number of new imaging techniques are described which give direct insight into the transfer processes and promise to trigger substantial theoretical progress in the near future.

B. Ja¨hne, University of Heidelberg, Heidelberg, Germany & 2009 Elsevier Ltd. All rights reserved.

Introduction

Theory

The exchange of inert and sparingly soluble gases, including carbon dioxide, methane, and oxygen, between the atmosphere and oceans is controlled by a 20–200-mm-thick boundary layer at the top of the ocean. The hydrodynamics in this layer is significantly different from boundary layers at rigid walls since the orbital motion of the waves is of the same order as the velocities in the viscous boundary layer. Laboratory and field measurements show that wind waves and surfactants significantly influence the gastransfer process. Because of limited experimental techniques, the details of the mechanisms and the structure of the turbulence in the boundary layer at

Mass Boundary Layers

Table 1

The transfer of gases and volatile chemical species between the atmosphere and oceans is driven by a concentration difference and the transport by molecular and turbulent motion. Both types of transport processes can be characterized by ‘diffusion coefficients’, denoted by D and Kc, respectively (Table 1). The resulting flux density jc is proportional to the diffusion coefficient and the concentration gradient. Thus, jc ¼ ðD þ Kc ðzÞÞrc

½1

Diffusion coefficients for various gases and volatile chemical species in deionized water and in some cases in seawater

Species

Heat 3 Heb,c 4 He 4 Hea Ne Kr Xe 222 Rnb H2 H2a CH4 CO2 DMSb CH3Brb F12b (CCl2F2) F11b (CCl3F) SF6b

Molecular mass

3.02 4.00 20.18 83.80 131.30 222.00 2.02 16.04 44.01 62.13 94.94 120.91 137.37 146.05

A (10 5 cm2 s 1)

379.2 941 818 886 1608 6393 9007 15 877 3338 1981 3047 5019 2000 3800 4100 3400 2900

Ea (kJ mol 1)

2.375 11.70 11.70 12.02 14.84 20.20 21.61 23.26 16.06 14.93 18.36 19.51 18.10 19.10 20.50 20.00 19.30

s(Fit) (%)

2.1 2.1 1.8 3.5 1.6 3.5 11 1.6 4.3 2.7 1.3

Diffusion coefficient (10 5 cm2 s 1) 5 1C

15 1C

25 1C

35 1C

135.80 5.97 5.10 4.86 2.61 1.02 0.77 0.68 3.17 3.05 1.12 1.07 0.80 0.98 0.58 0.60 0.69

140.72 7.12 6.30 5.88 3.28 1.41 1.12 0.96 4.10 3.97 1.48 1.45 1.05 1.31 0.79 0.81 0.92

145.48 8.39 7.22 7.02 4.16 1.84 1.47 1.34 5.13 4.91 1.84 1.91 1.35 1.71 1.05 1.07 1.20

150.08 9.77 8.48 8.03 4.82 2.40 1.94 1.81 6.23 5.70 2.43 2.43 1.71 2.20 1.37 1.38 1.55

a

In seawater. Values of diffusion coefficients from fit, not measured values. c Set 15% higher than 4He. Columns 3 and 4 contain the parameters for the fit of the diffusion coefficient: D ¼ A exp[ Ea/(RT )], the last four columns the diffusion coefficients for 5, 15, 25, and 35 1C. Data collected from Ja¨hne B, Heinz G, and Dietrich W (1987) Measurement of the diffusion coefficients of sparingly soluble gases in water. Journal of Geophysical Research 92: 10767–10776; and King DB, De Bryun WJ, Zheng M, and Saltzman ES (1995) Uncertainties in the molecular diffusion coefficient of gases in water for use in the estimation of air–sea exchange. In: Ja¨hne B and Monahan E (eds.) Air–Water Gas Transfer, pp. 13–22. Hanau: Aeon. b


147

148

AIR–SEA GAS EXCHANGE

In a stationary homogeneous case and without sinks and sources by chemical reactions, the flux density j is in vertical direction and constant. Then integration of [1] yields vertical concentration profiles Z 0

Zr

1 dz D þ Kc ðzÞ

The molecular diffusion coefficient is proportional to the velocity of the molecules and the free length between collisions. The same concept can be applied to turbulent diffusion coefficients. Far away from the interface, the free length (called ‘mixing length’) is set proportional to the distance from the interface and the turbulent diffusion coefficient Kc for mass transfer is Kc ¼

k u z Sct

Air-side mass boundary layer (100–1000 μm)

½2 Kc < D

Ca

C Water C w s = Ca s surface Aqueous mass boundary layer (20–200 μm)

s

Viscous boundary layer (600–2000 μm)

Water phase

zr

Cw

Reference level b

z

½3

where k ¼ 0.41 is the von Ka´rmań constant, u , the friction velocity, a measure for the velocity fluctuations in a turbulent flow, and Sct ¼ Km/Kc the turbulent Schmidt number. Closer to the interface, the turbulent diffusion coefficients are decreasing even faster. Once a critical length scale l is reached, the Reynolds number Re ¼ u l=v (n is the kinematic viscosity, the molecular diffusion coefficient for momentum) becomes small enough so that turbulent motion is attenuated by viscosity. The degree of attenuation depends on the properties of the interface. At a smooth solid wall, Kcpz3, at a free water interface it could be in the range between Kcpz3 and Kcpz2 depending on surface conditions. Boundary layers are formed on both sides of the interface (Figure 1). When the turbulent diffusivity becomes equal to the kinematic viscosity, the edge of the ‘viscous boundary layer’ is reached. As the name implies, this layer is dominated by viscous dissipation and the velocity profile becomes linear because of a constant diffusivity. The edge of the ‘mass boundary layer’ is reached when the turbulent diffusivity becomes equal to the molecular diffusivity. The relative thickness of both boundary layers depends on the dimensionless ratio Sc ¼ v/D (Schmidt number). The viscous and mass boundary layers are of about the same thickness in the air, because values of D for various gaseous species and momentum are about the same (Scair is 0.56 for H2O, 0.63 for heat, and 0.83 for CO2). In the liquid phase the situation is completely different. With Schmidt numbers in the range from 100 to 3000 (Figure 2 and Table 2), molecular diffusion for a dissolved volatile chemical species is two to three orders of magnitude slower than diffusion of momentum. Thus the mass boundary layer

z∼

Kc < υ

Figure 1 Schematic graph of the mass boundary layers at a gas–liquid interface for a tracer with a solubility a ¼ 3. 1

102

10

103

104

0

105

0

Air-side control

H2O

Atrazine

30

0 Pentachlorophenol

Momentum

DDT

103

3)

Sm

Transition zone

n= ce (

2/ (n =

SO2 (pH 60

½11

This equation establishes the basic analogy between momentum transfer and gas exchange. The transfer coefficient is proportional to the friction velocity in water, which describes the shear stress (tangential force per unit area) t ¼ rw u2w applied by the wind field at the water surface. Assuming stress continuity at the water surface, the friction velocity in water is related to the friction velocity in air by uw ¼ ua

ra rw

1=2 ½12

The friction velocity in air, ua , can further be linked via the drag coefficient to the wind speed UR at a reference height: cD ¼ ðua =UR Þ2 . Depending on the roughness of the sea surface, the drag coefficient has values between 0.8 and 2.4 10 3. In this way the gas exchange rate is directly linked to the wind speed. The gas exchange further depends on the chemical species and the water temperature via the Schmidt number. Gas Exchange at Rough and Wavy Water Surfaces

A free water surface is neither solid nor is it smooth as soon as short wind waves are generated. On a free water surface velocity fluctuations are possible. Thus, there can be convergence or divergence zone at the surface; surface elements may be dilated or contracted. At a clean water surface dilation or



contraction of a surface element does not cause restoring forces, because surface tension only tries to minimize the total free surface area, which is not changed by this process. As a consequence of this hydrodynamic boundary condition, the turbulent diffusivity normal to the interface can now increase with the distance squared from the interface, Kcpz2. Then 1 kw ¼ uw Sc1=2 b

½13

where b is a dimensionless constant. In comparison to the smooth case in [11], the exponent n of the Schmidt number drops from 2/3 to 1/2. This increases the transfer velocity for a Schmidt number of 600 by about a factor of 3. The total enhancement depends on the value of the constant b. Influence of Surface Films

A film on the water surface creates pressure that works against the contraction of surface elements. This is the point at which the physicochemical structure of the surface influences the structure of the near-surface turbulence as well as the generation of waves. As at a rigid wall, a strong film pressure at the surface maintains a two-dimensional continuity at the interface just as at a rigid wall. Therefore, [11] should be valid for a smooth film-covered water surface and has indeed been verified in wind/ wave tunnel studies as the lower limit for the transfer velocity. As a consequence, both [11] and [13] can only be regarded as limiting cases. A more general approach is required that has not yet been established. One possibility is a generalization of [11] and [13] to kw ¼ uw

1 ScnðsÞ bðsÞ

½14

where both b and n depend on dimensionless parameters describing the surface conditions s. Even films with low film pressure may easily decrease the gas transfer rate to half of its value at clean water surface conditions. But still too few measurements at sea are available to establish the influence of surfactants on gas transfer for oceanic conditions more quantitatively. Influence of Waves

Wind waves cannot be regarded as static roughness elements for the liquid flow because their characteristic particle velocity is of the same order of magnitude as the velocity in the shear layer at the surface.

151

This fact causes a basic asymmetry between the turbulent processes on the air and on the water sides of the interface. Therefore, the wave effect on the turbulent transfer in the water is much stronger and of quite different character than in the air. This basic asymmetry can be seen if the transfer velocity for CO2 is plotted against the transfer velocity for water vapor (Figure 3(a)). At a smooth water surface the points fall well on the theoretical curve predicted by the theory for a smooth rigid wall. However, as soon as waves occur at the water surface, the transfer velocity of CO2 increases significantly beyond the predictions. Even at high wind speeds, the observed surface increase is well below 20%. When waves are generated by wind, energy is not only transferred via shear stress into the water but a second energy cycle is established. The energy put by the turbulent wind into the wave field is transferred to other wave numbers by nonlinear wave–wave interaction and finally dissipated by wave breaking, viscous dissipation, and turbulence. The turbulent wave dissipation term is the least-known term and of most importance for enhanced near-surface turbulence. Evidence for enhanced turbulence levels below wind waves has been reported from field and laboratory measurements. Experimental results also suggest that the gas transfer rate is better correlated with the ‘mean square slope’ of the waves as an integral measure for the nonlinearity of the wind wave field than with the wind speed. It is not yet clear, however, to what extent ‘microscale wave breaking’ can account for the observed enhanced gas transfer rates. A gravity wave becomes unstable and generates a steep train of capillary waves at its leeward face and has a turbulent wake. This phenomenon can be observed even at low wind speeds, as soon as wind waves are generated. At higher wind speeds, the frequency of microscale wave breaking increases. Influence of Breaking Waves and Bubbles

At high wind speeds, wave breaking with the entrainment of bubbles may enhance gas transfer further. This phenomenon complicates the gas exchange between atmosphere and the oceans considerably. First, bubbles constitute an additional exchange surface. This surface is, however, only effective for gases with low solubility. For gases with high solubility, the gas bubbles quickly come into equilibrium so that a bubble takes place in the exchange only for a fraction of its lifetime. Thus, bubble-mediated gas exchange depends – in contrast to the exchange at the free surface – on the solubility of the gas tracer.


152


(a)

(b)

+ with waves (unlimited fetch) • no waves

0.001 0.80

10−2 Schmidt number exponent n

10−4

0.75 0.70

0.65

0.65

0.60

0.60

0.55

0.55

0.50

0.50

0.45

0.45

0.40 0.001

1

0.3 0.80

0.70

kH2O (cm s−1) 0.1

0.1 Small circular facility Large circular facility

0.75

kCO2 (Sc = 600) (cm s−1)

10−3

0.01

0.01

0.1

0.40 0.3

Mean square surface slope s 2

10

Figure 3 (a) Transfer velocity of CO2 plotted against the transfer velocity of water vapor as measured in a small circular wind wave facility. (b) Schmidt number exponent n as a function of the mean square slope. (a) From Ja¨hne B (1980) Zur Parameterisierung des Gasaustausches mit Hilfe von Laborexperimenten. Dissertation, University of Heidelberg. (b) From Ja¨hne B and HauXecker H (1998) Air–water gas exchange. Annual Review of Fluid Mechanics 30: 443–468.

Second, bubble-mediated gas transfer shifts the equilibrium value to slight supersaturation due to the enhanced pressure in the bubbles by surface tension and hydrostatic pressure. Third, breaking waves also enhance near-surface turbulence during the breaking event and the resurfacing of submerged bubbles. Experimental data are still too sparse for the size and depth distribution of bubbles and the flux of the bubbles through the interface under various sea states for a sufficiently accurate modeling of bubblemediated air–sea gas transfer and thus a reliable estimate of the contribution of bubbles to the total gas transfer rate. Some experiments from wind/wave tunnels and the field suggest that significant enhancements can occur, other experiments could not observe a significant influence of bubbles. Empiric Parametrization

Given the lack of knowledge all theories about the enhancement of gas transfer by waves are rather speculative and are not yet useful for practical application. Thus, it is still state of the art to use semiempiric or empiric parametrizations of the gas exchange rate with the wind speed. Most widely used is the parametrization of Liss and Merlivat. It identifies three physically well-defined regimes (smooth,

wave-influenced, and bubble-influenced) and proposes a piecewise linear relation between the wind speed U and the transfer velocity k: k ¼ 106 8 > 0:472UðSc=600Þ2=3 ; Ur3:6 m s1 > < 7:917ðU 3:39ÞðSc=600Þ1=2 ; U > 3:6 m s1 and Ur13 m s1 > > : 16:39ðU 8:36ÞðSc=600Þ1=2 ; U > 13 m s1

½15 At the transition between the smooth and wavy regime, a sudden artificial jump in the Schmidt number exponent n from 2/3 to 1/2 occurs. This actually causes a discontinuity in the transfer rate for Schmidt number unequal to 600. The empiric parametrization of Wanninkhof simply assumes a quadratic increase of the gas transfer rate with the wind speed: k ¼ 0:861 106 ðs m1 ÞU2 ðSc=600Þ1=2 ½16 Thus, this model has a constant Schmidt number exponent n ¼ 1/2. The two parametrizations differ significantly (see Figure 4). The Wanninkhof parametrization predicts significantly higher values. The discrepancy between the two parametrizations



80

14

60

C

SF6 − 3He 222

k 600 (cm h−1)

where Fw and hw are the surface area and the effective height Vw/Fw of a well-mixed water body, respectively. The time constant tw ¼ hw/k is in the order of days to weeks. It is evident that the transfer velocities obtained in this way provide only values integrated over a large horizontal length scales and timescales in the order of tw. Thus, a parametrization of the transfer velocity is only possible under steady-state conditions over extended periods. Moreover, the mass balance contains many other sources and sinks besides air–sea gas exchange and thus may cause severe systematic errors in the estimation of the transfer velocity. Consequently, mass balance methods are only poorly suited for the study of the mechanisms of air–water gas transfer.

Wanninkhof relationship Liss−Merlivat relationship

70

50

Rn

Heat (CFT)

40 30 20 10 0

0

5

153

10 15 Wind speed u10 (m s−1)

20

Tracer Injection

Figure 4 Summary of gas exchange field data normalized to a Schmidt number of 600 and plotted vs. wind speed together with the empirical relationships of Liss and Merlivat and Wanninkhof. Adapted from Ja¨hne B and HauXecker H (1998) Air–water gas exchange. Annual Review of Fluid Mechanics 30: 443–468.

(up to a factor of 2) mirrors the current uncertainty in estimating the air–sea gas transfer rate.

Experimental Techniques and Results Laboratory Facilities

Laboratory facilities play an important role in the investigation of air–sea gas transfer. Only laboratory studies allow a systematic study of the mechanisms and are thus an indispensible complement to field experiments. Almost all basic knowledge about gas transfer has been gained by laboratory experiments in the past. Among other things this includes the discovery of the influence of waves on air–water gas exchange (Figure 3(a)) and the change in the Schmidt number exponent (Figure 3(b)). Many excellent facilities are available worldwide (Table 3). Some of the early facilities are no longer operational or were demolished. However, some new facilities have also been built recently which offer new experimental opportunities for air–water gas transfer studies. Geochemical Tracer Techniques

The first oceanic gas exchange measurements were performed using geochemical tracer methods such as the 14C, 3He/T, or 222Rn/226Ra methods. The volume and time-average flux density is given by mass balance of the tracer concentration in a volume of water Vw: Vw c_w ¼ Fw j or j ¼ hw c_w

½17

The pioneering lake studies for tracer injection used sulfur hexafluoride (SF6). However, the tracer concentration decreases not only by gas exchange across the interface but also by horizontal dispersion of the tracer. This problem can be overcome by the ‘dual tracer technique’ (Watson and co-workers) simultaneously releasing two tracers with different diffusivities (e.g., SF6 and 3He). When the ratio of the gas transfer velocities of the two tracers is known, the dilution effect by tracer dispersion can be corrected, making it possible to derive gas transfer velocities. But the basic problem of mass balance techniques, that is, their low temporal resolution, remains also with artificial tracer approaches. Eddy Correlation Flux Measurements

Eddy correlation techniques are used on a routine basis in micrometeorology, that is, for tracers controlled by the boundary layer in air (momentum, heat, and water vapor fluxes). Direct measurements of the air–sea fluxes of gas tracers are very attractive because the flux densities are measured directly and have a much better temporal resolution than the mass balance-based techniques. Unfortunately, large experimental difficulties arise when this technique is applied to gas tracers controlled by the aqueous boundary layer. The concentration difference in the air is only a small fraction of the concentration difference across the aqueous mass boundary layer. But after more than 20 years of research has this technique delivered useful results. Some successful measurements under favorable conditions have been reported and it appears that remaining problems can be overcome in the near future. The Controlled Flux Technique

The basic idea of this technique is to determine the concentration difference across the mass boundary


(c) 2011 Elsevier Inc. All Rights Reserved. 2.0 0.8 104 83 15 N Y Y Y Y Y N

1.5 0.3 27 8 25 N N Y N N N N

3.0 0.8 800 768 15 N Y Y N N N N

100 8.0

D

2.4 1.5 96 144 12 Y Y N N N N N

40 2.4

SIO

0.85 0.25 24.8 8 25 N Y 70.6 Y N N Y

33 0.76

C

1.0 0.5 15 10 30 Y Y 70.5 Y N N Y

15 1.0

UM

1.22 0.76 16.7 ? 25 N Y 70.5 Y Y N Y

18.3 0.91

W 119 2.0 40 36 5.6 3.0 239 716 19 Y N N N N N N

SU 1.57 0.10 0.60 0.40 0.50 0.08 0.16 0.01 11 Y N N 5–35 5–35 Y Y

HD1

11.6 0.20 4.0 3.4 0.70 0.25 3.5 0.87 12 Y N N N N N Y

HD2

29.2 0.62 9.92 8.68 2.40 1.20 18.0 20.7 15 Y N o 0.6 5–35 5–35 Y Y

HD3

0.33 0.10 1.44 0.14 8 Y N o 0.1 Y N N Y

3.90 0.37

HD4

HH, Bundesanstalt fu¨r Wasserbau, Hamburg; M, IMST, Univ. Marseille, France; D, Delft Hydraulics, Delft, The Netherlands (no longer operational); SIO, Hydraulic Facility, Scripps Institution of Oceanography, La Jolla, USA; C, Canada Center for Inland Waters (CCIW); UM, University of Miami; W, NASA Air–Sea Interaction Research Facility, Wallops; SU, Storm basin, Marine Hydrophysical Institute, Sevastopol, Ukraine (no longer operational), HD1, Small annular wind/wave flume, Univ. Heidelberg (no longer operational), HD2, Large annular wind/wave flume, Univ. Heidelberg (dismantled); HD3, Aeolotron, Univ. Heidelberg (HD3, in operation since June 2000); HD4, Teflon-coated small Heidelberg linear wind/wave flume (N ¼ No, Y ¼ Yes).

40 2.6

M

15 1.8

HH

Comparison of the features of some major facilities for small-scale air–sea interaction studies (operational facilities are typeset in boldface)

Length (mean perimeter) (m) Width of water channel (m) Outer diameter (m) Inner diameter (m) Total height (m) Max. water depth (m) Water surface area (m2) Water volume (m3) Maximum wind speed (m s 1) Suitable for sea water Wave maker Water current generator (m s 1) Water temperature control (1C) Air temperature control (1C) Air humidity control Gastight air space

Table 3

14

28

13 12

Wind (20 Hz)

26

Wind (60 s average)

24

11 Gas transfer rates 10

(k600, CFT data)

22 20

9

18

8

16

7

14

6

12

5

10

4

8

3

6

2

4

1

2

0 03:00

155

03:30

04:00

k600 (cm h−1)

Wind (u10) (m s−1)


0 04:30

Time (UTC) (hh:mm) Figure 5 Wind speeds and gas transfer velocities computed with the controlled flux technique (CFT) during the 1995 MBL/CoOP West Coast experiment (JD133) for a period of 90 min. The transfer velocities are normalized to Schmidt number 600 and averaged over 4 min each. From Ja¨hne B and HauXecker H (1998) Air–water gas exchange. Annual Review of Fluid Mechanics 30: 443–468.

layer when the flux density j of the tracer across the interface is known. The local transfer velocity can be determined by simply measuring the concentration difference Dc across the aqueous boundary layer (cold surface skin temperature) according to [4] with a time constant t˜ for the transport across the boundary layer [8]. This technique is known as the ‘controlled flux technique’ (CFT). Heat proves to be an ideal tracer for the CFT. The temperature at the water surface can then be measured with high spatial and temporal resolution using IR thermography. A known and controllable flux density can be applied by using infrared radiation. Infrared radiation is absorbed in the first few 10 mm at the water surface. Thus, a heat source is put right at top of the aqueous viscous boundary layer. Then the CFT directly measures the waterside heat transfer velocity. A disadvantage of the CFT is that the transfer velocity of gases must be extrapolated from the transfer velocity of heat. The large difference in the Schmidt number (7 for heat, 600 for CO2) casts some doubt whether the extrapolation to so much higher Schmidt numbers is valid. Two variants of the technique proved to be successful. Active thermography uses a CO2 laser to heat a spot of several centimeters in diameter on the water surface. The heat transfer rates are estimated

from the temporal decay of the heated spot. Passive thermography uses the naturally occurring heat fluxes caused by latent heat flux jl, sensible heat flux js, and long wave emission of radiation jr . The net heat flux jn ¼ jl þ js þ jr results according to [4] in a temperature difference across the interface of DT ¼ jh/(rcpkh). Because of the turbulent nature of the exchange process any mean temperature difference is associated with surface temperature fluctuations which can be observed in thermal images. With this technique the horizontal structure of the boundary layer turbulence can be observed. Surface renewal is directly observable in the IR image sequences, which show patches of fluid being drawn away from the surface. With some knowledge about the statistics of the temperature fluctuations, the ‘temperature difference’ DT across the interface as well as the time constant t˜ of heat transfer can be computed from the temperature distribution at the surface. Results obtained with this technique are shown in Figure 5 and also in the overview graph (Figure 4). Summary of Field Data

A collection of field data is shown in Figure 4. Although the data show a clear increase of the transfer


156


velocity with wind speed, there is substantial scatter in the data that can only partly be attributed to uncertainties and systematic errors in the measurements. Thus, in addition, the field measurements reflect the fact that the gas transfer velocity is not simply a function of the wind speed but depends significantly on other parameters influencing nearsurface turbulence, such as the wind-wave field and the viscoelastic properties of the surface film.

Outlook In the past, progress toward a better understanding of the mechanisms of air–water gas exchange was hindered by inadequate measuring technology. However, new techniques have become available and will continue to become available that will give a direct insight into the mechanisms under both laboratory and field conditions. This progress will be achieved by interdisciplinary research integrating different research areas such as oceanography, micrometeorology, hydrodynamics, physical chemistry, applied optics, and image processing. Optical- and image-processing techniques will play a key role because only imaging techniques give direct insight to the processes in the viscous, heat, and mass boundary layers on both sides of the air– water interface. Eventually all key parameters including flow fields, concentration fields, and waves will be captured by imaging techniques with sufficient spatial and temporal resolution. The experimental data gained with such techniques will stimulate new theoretical and modeling approaches.

Nomenclature 2

D (cm s 1) jc (Mol cm 2 s 1) k (cm s 1) Kc (cm2 s 1) R (cm 1 s) Sc ¼ n/D t˜ ¼ z˜/k (s) u (cm s 1) z˜ (cm) a n (cm2 s 1)

molecular diffusion coefficient concentration flux density transfer velocity turbulent diffusion coefficient transfer resistance Schmidt number boundary layer time constant friction velocity boundary layer thickness dimensionless solubility kinematic viscosity

Sea Transfer: N2O, NO, CH4, CO. Breaking Waves and Near-Surface Turbulence. Bubbles. Carbon Dioxide (CO2) Cycle. Pollution: Effects on Marine Communities. Surface Gravity and Capillary Waves.

Further Reading Borger AV and Wanninkhof R (eds.) (2007) Special Issue: 5th International Symposium on Gas Transfer at Water Surfaces. Journal of Marine Systems 66: 1--308. Businger JA and Kraus EB (1994) Atmosphere–Ocean Interaction. New York: Oxford University Press. Donelan M, Drennan WM, Saltzman ES, and Wanninkhof R (eds.) (2001) Gas Transfer at Water Surfaces. Washington, DC: American Geophysical Union. Duce RA and Liss PS (eds.) (1997) The Sea Surface and Global Change. Cambridge, UK: Cambridge University Press. Garbe C, Handler R, and Ja¨hne B (eds.) (2007) Transport at the Air–Sea Interface, Measurements. Models, and Parameterization. Berlin: Springer. Ja¨hne B (1980) Zur Parameterisierung des Gasaustausches mit Hilfe von Laborexperimenten. Dissertation, University of Heidelberg. Ja¨hne B and HauXecker H (1998) Air–water gas exchange. Annual Review of Fluid Mechanics 30: 443--468. Ja¨hne B, Heinz G, and Dietrich W (1987) Measurement of the diffusion coefficients of sparingly soluble gases in water. Journal of Geophysical Research 92: 10767--10776. Ja¨hne B and Monahan E (eds.) (1995) Air–Water Gas Transfer. Hanau: Aeon. King DB, De Bryun WJ, Zheng M, and Saltzman ES (1995) Uncertainties in the molecular diffusion coefficient of gases in water for use in the estimation of air–sea exchange. In: Ja¨hne B and Monahan E (eds.) Air–Water Gas Transfer, pp. 13--22. Hanau: Aeon. Liss PS and Merlivat L (1986) Air–sea gas exchange rates: Introduction and synthesis. In: Buat-Menard P (ed.) The Role of Air–Sea Exchange in Geochemical Cycles, pp. 113--127. Dordrecht: Reidel. McGilles WR, Asher WE, Wanninkhof R, and Jessup AT (eds.) (2004). Special Issue: Air Sea Exchange. Journal of Geophysical Research 109. Wanninkhof R (1992) Relationship between wind speed and gas exchange over the ocean. Journal of Geophysical Research 97: 7373--7382. Wilhelms SC and Gulliver JS (eds.) (1991) Air–Water Mass Transfer. New York: ASCE.

Relevant Websites See also Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens. Air–

http://www.ifm.zmaw.de – Institute of Oceanography, Universita¨t Hamburg. http://www.solas-int.org – SOLAS.


AIR–SEA TRANSFER: DIMETHYL SULFIDE, COS, CS2, NH4, NON-METHANE HYDROCARBONS, ORGANO-HALOGENS J. W. Dacey, Woods Hole Oceanographic Institution, Woods Hole, MA, USA H. J. Zemmelink, University of Groningen, Haren, The Netherlands Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 131–137, & 2001, Elsevier Ltd.

The oceans, which cover 70% of Earth’s surface to an average depth of 4000 m, have an immense impact on the atmosphere’s dynamics. Exchanges of heat and momentum, water and gases across the sea surface play major roles in global climate and biogeochemical cycling. The ocean can be thought of as a vast biological soup with myriad processes influencing the concentrations of gases dissolved in the surface waters. The quantities of mass flux across the surface interface, though perhaps small on a unit area basis, can be very important because of the extent of the ocean surface and the properties of the gases or their decomposition products in the atmosphere. Gas exchange across the sea–air surface depends, in part, on differences in partial pressures of the gases between the ocean surface and the atmosphere. The partial pressure of a gas in the gas phase can be understood in terms of its contribution to the pressure in the gas mixture. So the partial pressure of O2, for example, at 0.21 atm means that at 1 atmosphere total pressure, O2 is present as 21% of the gas, or mixing, volume. Trace gases are present in the atmosphere at much lower levels, usually expressed as parts per million (106 atm), parts per billion (109 atm) or parts per trillion (1012 atm, pptv). Dimethylsulfide (DMS), when present at 100 pptv, accounts for about 100 molecules per 1012 molecules of mixed gas phase, or about 1010 of the gas volume. In solution, a dissolved trace gas in equilibrium with the atmosphere would have the same partial pressure as the gas in the air. Its absolute concentration in terms of molecules or mass per unit volume of water depends on its solubility. Gas solubility varies over many orders of magnitude depending on the affinity of water for the gas molecules and the volatility of the gas. Gases range widely in their solubility in sea water, from the permanent gases like nitrogen (N2), oxygen (O2), nitrous oxide (N2O) and methane (CH4) that have a low solubility in sea

water to the moderately soluble carbon dioxide (CO2) and dimethylsulfide (CH3)2S, to highly soluble ammonia (NH3 and its ionized form NHþ 4 ) and sulfur dioxide (SO2). Sulfur dioxide is more than 106 times more soluble than O2 or CH4. Using the example above of an atmospheric DMS concentration of 100 pptv, the equilibrium concentration of DMS in surface water would be about 0.07 nmol l1. Generally the solubility of any individual gas increases at cooler water temperatures, and solubility of gases in sea water is somewhat less than for fresh water because of the so-called ‘salting out’ effect of dissolved species in sea water. At any moment the partial pressure difference between surface water and the atmosphere depends on an array of variables. The gases in this article are biogenic, meaning that their mode of formation is the result of one or more immediate or proximate biological processes. These dissolved gases may also be consumed biologically, or removed by chemical processes in sea water, or they may flux across the sea surface to the atmosphere. The rates at which the source and sink processes occur determines the concentration of the dissolved gas in solution as well as the turnover, or residence time, of each compound. Similarly, there can be several source and sink processes for the gases in the atmosphere. Long-lived compounds in the atmosphere will tend to integrate more global processes, whereas short-lived compounds are concentrated near their source and reflect relatively short-term influences of source and sink. In this sense, carbonyl sulfide is a global gas. At it has a residence time of several years in the atmosphere, its concentration does not vary in the troposphere to any appreciable degree. On the other hand, the concentration of DMS varies on a diel basis and with elevation, with higher concentrations at night when atmospheric oxidants (most notable hydroxyl) are relatively depleted. The extent of disequilibrium between the partial pressures of a gas in the surface water and in the atmosphere determines the thermodynamic gradient which drives gas flux. The kinetics of flux ultimately depend on molecular diffusion and larger-scale mixing processes. Molecular diffusivity is generally captured in a dimensionless parameter, the Schmidt number (ratio of viscosity of water to molecular


157

158

AIR–SEA TRANSFER

diffusivity of gas in water), and varies widely between gases depending primarily on the molecular cross-section. From moment to moment, the flux of any particular gas is dependent on interfacial turbulence which is generated by shear between the wind and the sea surface whereby higher wind speed causes increasing turbulence and thus stimulating the onset of waves and eventually the production of bubbles and sea spray. There are considerable uncertainties relating gas exchange to wind speed. These arise due to the various sea-state factors (wave height, swell, breaking waves, bubble entrainment, surfactants, and others) whose individual dependencies on actual wind speed and wind history are not well quantified. The fluxes of gases across the air–sea interface are usually calculated using a wind-speed parameterization. These estimates are considered to be accurate to within a factor of 2 or so. This article summarizes the characteristics of several important trace gases – dimethylsulfide, carbonyl sulfide, carbon disulfide, nonmethane hydrocarbons, ammonia and methylhalides – focusing on their production and fate as it is determined by biological and chemical processes.

Dimethylsulfide Natural and anthropogenic sulfur aerosols play a major role in atmospheric chemistry and potentially in modulating global climate. One theory holds that a negative feedback links the emission of volatile organic sulfur (mostly as DMS) from the ocean with the formation of cloud condensation nuclei, thereby

regulating, in a sense, the albedo and radiation balance of the earth. The direct (backscattering and reflection of solar radiation by sulfate aerosols) and indirect (cloud albedo) effects of sulfate aerosols may reduce the climatic forcing of trace greenhouse gases like CO2, N2O and CH4. The oxidation products of DMS which also contribute to the acidity of rain, particularly in marine areas, result from industrialized and/or well-populated land. Dimethylsulfide (DMS) is the most abundant volatile sulfur compound in sea water and constitutes about half of the global biogenic sulfur flux to the atmosphere. Studies of the concentration of DMS in the ocean have shown that average surface water concentrations may vary by up to a factor of 50 between summer and winter in mid and high latitudes. Furthermore, there are large-scale variations in DMS concentration associated with phytoplankton biomass, although there are generally poor correlations between local oceanic DMS concentrations and the biomass and productivity of phytoplankton (due to differences between plankton species in ability to produce DMS). The nature and rates of the processes involved in the production and consumption of DMS in sea water are important in determining the surface concentrations and the concomitant flux to the atmosphere. The biogeochemical cycle of DMS (Figure 1) begins with its precursor, b-dimethylsulfoniopropionate (DMSP). DMSP is a cellular component in certain species of phytoplankton, notably some prymnesiophytes and dinoflagellates. The function of DMSP is unclear, although there is evidence for an

Photochemistry SO2 Stratosphere COS DMS CS2

SO2 MSA

Troposphere Photochemistry DMS

Ocean

DMSO Photochemistry 2_ 4

DMSP

SO Organic matter

Figure 1 Fate and production of dimethylsulfide (DMS), carbonylsulfide (COS) and carbon disulfide. DMSO, dimethylsulfoxide; DMSP, dimethylsulfoniopropionate; MSA, methane sulfonic acid.


AIR–SEA TRANSFER

osmoregulatory role as its cellular concentrations have been found to vary with salinity. It is generally thought that healthy algal cells do not leak either DMSP or DMS, although mechanical release into the surrounding sea water can lead to DMS production during cell senescence and grazing by zooplankton or as a consequence of viral attack on phytoplankton cells. Oceanic regions dominated by prolific DMSPproducing phytoplankton tend to have high DMS and DMSP concentrations. Breakdown of DMSP, presumably after transfer from the particulate algal (pDMSP) form to a dissolved (dDMSP) form in sea water, can proceed in different ways, mostly depending on microbiological conditions. One major pathway involves cleavage of DMSP to DMS and acrylic acid. Bacterial metabolism of dDMSP may be a major mechanism for DMS production in sea water, with acrylic acid residue acting as a carbon source for heterotrophic growth. Sulfonium compounds are vulnerable to attack by hydroxide ion; the resulting chemical elimination reaction occurs rapidly and quantitatively in strong base but only slowly at the pH of sea water. DMS in sea water has many potential fates. The volatility of DMS and the concentration gradient across the sea–air interface lead to the ocean being the major source of DMS to the atmosphere. Estimates of the annual sulfur release (as DMS) vary from 13–37 Tg S y1 (Kettle and Andreae, 1999). However, whereas the absolute flux of DMS from sea to air may be large on a global scale, sea–air exchange may represent only a minor sink for seawater DMS. It has been estimated that DMS loss to the atmosphere is only a very small percentage of the DMS sink, but this undoubtedly depends on the biogeochemical conditions in the water column at the time. Photochemical oxidation of DMS, either to dimethylsulfoxide (DMSO) or to other products, occurs via photosensitized reactions. The amount of photochemical decomposition depends on the amount of light of appropriate wavelengths and the concentration of colored organic compounds in solution to convert light energy into reactive radicals. Light declines exponentially with depth; the distribution of colored dissolved organic materials exhibits depth and seasonal variability. Microbial consumption of DMS, although extremely variable in both time and space in the ocean, appears to be a significant sink for oceanic DMS. The residence time of DMS is probably of the order of a day or two in most seawater systems. Since the atmospheric residence time of DMS is about a day or two, the atmospheric consequences of DMS flux are mostly confined to the troposphere. In the troposphere, DMS is oxidized primarily by hydroxyl radical. The main atmospheric oxidation

159

products are methane sulphonic acid, SO2 and DMSO.

Carbonyl Sulfide Carbonyl sulfide (COS, OCS) is the major sulfur gas in the atmosphere, present throughout the troposphere at 500 pptv. COS has a long atmospheric residence time (B4 years). Because of its relative inertness COS diffuses into the stratosphere where it oxidizes to sulfate particles and contributes in reactions involving stratospheric ozone chemistry. Unlike DMS which is photochemically oxidized in the troposphere, the major sink for COS is terrestrial vegetation and soils. COS is taken up by plants by passing through the stomata and subsequently hydrolyzing to CO2 and H2S through the action of carbonic anhydrase inside plant cells. There is no apparent physiological significance to the process; it appears to just occur accidentally to the normal physiology of plants. COS is produced in the ocean by photochemical oxidation of organic sulfur compounds whereby dissolved organic matter acts as a photosensitizer. The aqueous concentration of COS manifests a strong diel cycle, with the highest concentrations in daytime (concentration range on the order of 0.03– 0.1 nmol l1). COS hydrolyzes in water to H2S at rates dependent on water temperature and pH. The flux of oceanic COS to the atmosphere may represent about one-third of the global COS flux.

Carbon Disulfide Concentrations in surface water are around 1011 mol 11. Although a number of studies have indicated that the ocean forms an important source for atmospheric CS2, the underlying biochemical cycles still remain poorly understood. CS2 is formed by photochemical reactions (possibly involving precursors such as DMS, DMSP and isothiocyanates). CS2 formation has been observed to occur in bacteria in anoxic aquatic environments and in cultures of some marine algae species. The residence time of CS2 in the atmosphere is relatively short (about one week). Although CS2 might contribute directly to SO2 in the troposphere, its main significance is in the formation of COS via photochemical oxidation which results in the production of one molecule each of SO2 and COS per molecule of CS2 oxidized. The resulting COS may contribute to the stratospheric aerosol formation. Concentrations around 14 pmol l1 of carbon disulfide in the mid-Atlantic Ocean were first observed


160

AIR–SEA TRANSFER

(1974); higher concentrations have been found in coastal waters. More than a decade later CS2 concentrations in the North Atlantic were found to be comparable to the earlier observations. However, in coastal waters CS2 concentrations were found to be a factor 10 lower, respectively 33 and 300 pmol l1. The global CS2 flux has been estimated on 6.7 Gmol S y1, and it has been concluded that the marine emission of CS2 provides a significant indirect source of COS, but it forms an insignificant source of tropospheric SO2.

Nonmethane Hydrocarbons Nonmethane hydrocarbons (NMHCs) are important reactive gases in the atmosphere since they provide a sink for hydroxyl radicals and play key roles in the production and destruction of ozone in the troposphere. NMHCs generally refer to the C2–C4 series, notably ethane, ethene, acetylene, propane, propene, and n-butane, but also the five-carbon compound isoprene. Of these, ethene is generally the most abundant contributing 40% to the total NMHC pool in sea water. Published data of concentrations of NMHCs in sea water vary widely sometimes exceeding a factor 100. For example, in one extensive study, ethene and propane were found to be the most abundant species in the intertropical South Pacific, with mixing ratios of 2.7 to 58 and 6 to 75 pptv, respectively; whereas in the equatorial Atlantic these species showed mixing ratios of 20 pptv and 10 pptv, respectively. The water-column dynamics of NMHCs are poorly understood. NMHCs have been detected in the surface sea and with maxima in the euphotic zone and tend to be present at concentrations in sea water at around 1010 mol l1. Evidence suggests that photochemical oxidation of dissolved organic matter results in the formation of NMHCs. There can be very little doubt that the physiology of planktonic organisms is also involved in NMHC formation. Ethene and isoprene are freely produced by terrestrial plants where the former is a powerful plant hormone but the function of the latter less well understood. It is likely that similar processes occur in planktonic algae. NMHC production tends to correlate with light intensity, dissolved organic carbon and biological production. A simplified scheme of marine NMHC production is shown in Figure 2. The flux of NMHCs to the atmosphere (with estimates ranging from o10 Mt y1 to 50 Mt y1) is minor on a global scale, but has a potential significance in local atmospheric chemistry. Although oceans are known to act as sources of NMHCs, the

NMHC (R-H) + OH

Products (R + H2O) Troposphere

NMHC (R-H) Photosynthetic organisms

Ocean Photolysis and chemical conversion Dissolved organic matter

Zooplankton Bacteria

Figure 2 Simplified scheme of marine nonmethane hydrocarbon (NMHC) production. In the marine troposphere NMHC acts as a sink for hydroxyl (OH) radicals and thereby plays a key role in ozone chemistry.

sources of individual NMHCs in the marine boundary layer are not always clear. Those NMHCs with a life time of more than a week (e.g., ethane, ethyne, propane, cyclopropane) show latitudinal gradients consistent with a continental source, whereas variations of NMHCs with life times shorter than a week (all alkenes and pentane) are more consistent with a marine source.

Ammonia Ammonia is an extremely soluble gas, reacting with water and dissociating into an ammonium ion at ambient pH. At pH 8.2, about one-tenth of dissolved ammonia is present as NH3. Ammonium is also a rapidly cycling biological nutrient; it is taken up by bacteria and phytoplankton as a source of fixed nitrogen, and released by sundry physiological and decompositional processes in the food web. Anthropogenic loading of ammonium (and other nutrients) into the coastal marine environment results in increased phytoplankton growth in a phenomenon called eutrophication. Ammonium is oxidized to nitrate by bacteria in a process known as nitrification (Figure 3). Conversely, in anoxic environments, ammonium can be formed by nitrate-reducing bacteria. Ammonia plays an important role in the acid–base chemistry in the troposphere where the unionized ammonia (NH3) is converted into ionized ammonia (NHþ 4 ) via a reaction that neutralizes atmospheric acids as HNO3 and H2SO4. This leads to the formation of ammonium aerosols such as the stable ammonium sulfate. Eventually the ammonia returns to the surface by dry or wet deposition. Few data exist on the fluxes of NH3 over marine environments. Evidence suggests that most of the ocean surface serves as a source of NH3 to the


AIR–SEA TRANSFER

NOx

Oxidation

NH3(g)

Reduction

161

NH4-aerosol

Dry and wet deposition

Troposphere NH3(g)

Ammonification

NH4+

Water

Nitrification

NO3

Denitrification

Organic nitrogen Nitrogen fixation

N2

Figure 3 Simplified scheme of marine NHx chemistry. In the marine boundary layer NO3 acts as an initiator for the degradation of many organic compounds, in particular dimethylsulfide (DMS).

atmosphere, even in regions of very low nutrients. In the North Sea, an area situated in the middle of densely populated and industrialized countries of Western Europe, air from nearby terrestrial sources may act as a source of NH3 into surface waters. It has been estimated that the annual biogenic emission of ammonia from European seas is around 30 kt N y1, which is comparable to the emissions of smaller North European countries, leading to the conclusion (amongst others) that seas are among the largest sources of imported ammonium for maritime countries. The net emission of ammonia from coastal waters of the north-east Pacific Ocean to the atmosphere has been shown to be in the order of 10 mmol m2 d1.

Organohalogens Halogenated compounds, such as methyl chloride (CH3Cl), methyl bromide (CH3Br) and methyl iodide (CH3I) are a major source of halogens in the atmosphere, and subsequently form sources of reactive species capable of catalytically destroying ozone. Among these CH3I is likely to play an important role in the budget of tropospheric ozone, through production of iodine atoms by photolysis. Due to their higher photochemical stability methyl chloride and

methyl bromide are more important in stratospheric chemistry; it has been suggested that BrO species are responsible for losses of tropospheric ozone in the Arctic (Figure 4). Atmospheric methyl halides, measured over the ocean by several cruise surveys, have been shown to have average atmospheric mixing ratios of: CH3Cl, 550–600 pptv; CH3Br, 10–12 pptv; CH3I, 0.5–1 pptv. Their temporal and spatial variations are not well understood, neither is their production mechanism in the ocean known. Measurements of atmospheric and seawater concentrations of CH3Cl and CH3I have indicated that the oceans form natural sources of these methyl halides. In contrast, CH3Br appears to be undersaturated in the open ocean and exhibits moderate to 100% supersaturation in coastal and upwelling regions, leading to a global atmosphere to ocean flux of 13 Gg y1. Coastal salt marshes, although they constitute a minor area of the global marine environment, may produce roughly 10% of the total fluxes of atmospheric CH3Br and CH3Cl and thus contribute significantly to the global budgets. Macrophytic and phytoplanktonic algae produce a wide range of volatile organohalogens including di- and tri-halomethanes and mixed organohalogens. There is evidence for the involvement of enzymatic synthesis of methyl halides, but the metabolic


162

AIR–SEA TRANSFER

Temporary reservoirs HBr, BrONO2, etc. O3 Hv

Hv

CH3Br

O2

Br Hv

Coupling with CIOx

BrO HO2

O3 HOBr

Stratosphere

O3

Troposphere

NO2 HO2

Ozone loss CH3Br

OH

Photolabile fraction

Br O2 BrO

Hv

_

Br

BrO RCHO

and different dynamics spatially and with depth in the ocean. As conditions change in an apparently warming world, changes in the dynamics of surface ocean gases can be expected. The behavior of these trace gases or even the dynamics of the planktonic community are not understood sufficiently to allow good quantitative predictions about changes in trace gas flux to be made. Changes in flux of some gases could lead to an acceleration of warming, while changes in others could lead to cooling. It is, thus, important to understand the factors controlling trace gas dynamics in the surface ocean.

See also

Soluble fraction (HBr) Dry deposition

Wet deposition

Air–Sea Gas Exchange. Air–Sea Transfer: N2O, NO, CH4, CO. Atmospheric Input of Pollutants. Carbon Cycle. Chlorinated Hydrocarbons. Nitrogen Cycle.

Ocean CH3Br

Photochemistry

Dissolved organic matter

Further Reading

Phytoplankton/Bacteria

Figure 4 Schematic illustration of the circulation of methyl bromide.

production pathways are not well known. In free sea water, photochemical processes, ion substitution and, possibly the alkylation of halide ions (during the oxidation of organic matter by an electron acceptor such as Fe(III)) are also potential formation mechanisms. Sunlight or microbial mediation are not required for these reactions. In the ocean, chemical degradation of CH3Br occurs by nucleophilic substitution by chloride and hydrolysis. Microbial consumption is also a likely sink for halogenated compounds.

Conclusions The biogenic trace gases are influenced by the complete range of biological processes – from the biochemical and physiological to the ecological level of food web dynamics. The gases that are influenced directly by plant physiology (probably the light NMHCs and isoprene, for example) tend to be most closely related to phytoplankton biomass or primary productivity. Other gases produced during grazing and decomposition (e.g., DMS, NH3), or gases formed by photochemical reactions in dissolved organic material show differing temporal dynamics

Andreae MO (1990) Ocean–atmosphere interactions in the global biogeochemical sulfur cycle. Marine Chemistry 10: 1--29. Andreae MO and Crutzen PJ (1997) Atmospheric aerosols: biogeochemical sources and role in atmospheric chemistry. Science 276: 1052--1058. Barrett K (1998) Oceanic ammonia emissions in Europe and their transboundary fluxes. Atmospheric Environment 32(3): 381--391. Chin M and Davis DD (1993) Global sources and sinks of OCS and CS2 and their distributions. Global Biogeochemical Cycles 7: 321--337. Cox RA, Rattigana OV, and Jones RL (1995) Laboratory studies of BrO reactions of interest for the atmospheric ozone balance. In: Bandy RA (ed.) The Chemistry of the Atmosphere; Oxidants and Oxidation in the Earth’s Atmosphere. Cambridge: The Royal Society of Chemistry. Crutzen PJ (1976) The possible importance of COS for the sulfate layer of the stratosphere. Geosphysical Research Letters 3: 73--76. Graedel TE (1995) Tropospheric budget of reactive chlorine. Global Biogeochemical Cycles 9: 47--77. Kettle AJ and Andreae MO (1999) Flux of dimethylsulfide from the oceans: a comparison of updated datasets and flux models. Journal of Geophysical Research 105: 26793--26808. Lovelock JE (1974) CS2 and the natural sulfur cycle. Nature 248: 625--626. Turner SM and Liss PS (1985) Measurements of various sulfur gases in a coastal marine environment. Journal of Atmospheric Chemistry 2(3): 223--232.


AIR–SEA TRANSFER: N2O, NO, CH4, CO C. S. Law, Plymouth Marine Laboratory, The Hoe, Plymouth, UK Copyright & 2001 Elsevier Ltd.

identifies the marine contribution to total atmosphere budgets. There is also a brief examination of the approaches used for determination of marine trace gas fluxes and the variability in current estimates.

This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 137–144, & 2001, Elsevier Ltd.

Nitrous Oxide (N2O) Introduction The atmospheric composition is maintained by abiotic and biotic processes in the terrestrial and marine ecosystems. The biogenic trace gases nitrous oxide (N2O), nitric oxide (NO), methane (CH4) and carbon monoxide (CO) are present in the surface mixed layer over most of the ocean, at concentrations which exceed those expected from equilibration with the atmosphere. As the oceans occupy 70% of the global surface area, exchange of these trace gases across the air–sea interface represents a source/sink for global atmospheric budgets and oceanic biogeochemical budgets, although marine emissions of NO are poorly characterized. These trace gases contribute to global change directly and indirectly, by influencing the atmospheric oxidation and radiative capacity (the ‘greenhouse effect’) and, together with their reaction products, impact stratospheric ozone chemistry (Table 1). The resultant changes in atmospheric forcing subsequently influence ocean circulation and biogeochemistry via feedback processes on a range of timescales. This article describes the marine sources, sinks, and spatial distribution of each trace gas and

The N2O molecule is effective at retaining long-wave radiation with a relative radiative forcing 280 times that of a CO2 molecule. Despite this the relatively low atmospheric N2O concentration results in a contribution of only 5–6% of the present day ‘greenhouse effect’ with a direct radiative forcing of about 0.1 Wm2. In the stratosphere N2O reacts with oxygen to produce NO radicals, which contribute to ozone depletion. N2O is a reduced gas which is produced in the ocean primarily by microbial nitrification and denitrification. N2O is released during ammonium (NHþ 4 ) oxidation to nitrite (NO2 ) (Figure 1), although the exact mechanism has yet to be confirmed. N2O may be an intermediate of nitrification, or a byproduct of the decomposition of other intermediates, such as nitrite or hydroxylamine. Nitrification is an aerobic process, and the N2O yield under oxic conditions is low. However, as the nitrification rate decreases under low oxygen, the relative yield of N2O to nitrate production increases and reaches a maximum at 10–20 mmol dm3 oxygen (mmol ¼ 1 106 mol). Conversely, denitrification is an anaerobic process in which soluble oxidized nitrogen

Table 1 The oceanic contribution and atmospheric increase and impact for methane, nitrous oxide, nitric oxide, and carbon monoxidea Trace gas

Atmospheric concentration (ppbv)

Atmospheric lifetime (years)

Major impact in atmosphere

Increase in atmosphere (1980– 90)

Oceanic emission as % of total global emissions

Nitrous oxide (N2O) Nitric oxide (NO)

315

110–180

0.25% (0.8 ppbv y1)

7–34%

0.01

o0.2

Not known

Not known

Methane (CH4)

1760

10

0.8% (0.6 ppbv y1)

1–10%

Carbon monoxide (CO)

120

0.2–0.8

Infrared active Ozone sink/source Ozone sink/source OH sink/oxidation capacity Infrared active OH sink/oxidation capacity Ozone sink/source OH sink/oxidation capacity Ozone sink/source Infrared active

13 to 0.6%

0.9–9%

a

ppbv, parts per billion by volume. (Adapted from Houghton et al., 1995.)


163

164

AIR–SEA TRANSFER: N2O, NO, CH4, CO

Nitrification +

NH 4 H2O O2

_

NO2

_

N2O

NO

NO3

N2 Denitrification Figure 1 ‘Leaky Pipe’ flow diagram of nitrification and denitrification indicating the potential exchange and intermediate role of NO and N2O (Reprinted by permission from Nature copyright (1990), Macmillan Magazines Ltd.)

compounds, such as nitrate and nitrite, are converted to volatile reduced compounds (N2O and N2) in the absence of oxygen. Oxygen availability inhibits denitrification at ambient levels, and also determines the products of denitrification. An enzymatic gradient of sensitivity to oxygen results in the accumulation of N2O under sub-oxia (3–10 mmol dm3) due to the inhibition of the enzyme nitrous oxide reductase. At lower oxygen (o3 mmol dm3) the reaction continues through to N2 and so anoxic environments are sinks for N2O. N2O yields from nitrification are 0.2–0.5%, whereas denitrification yields may be as high as 5% at optimal levels of sub-oxia. An inverse correlation between N2O and oxygen, and associated linear relationship between nitrate and N2O, suggest that N2O in the ocean originates primarily from nitrification. This may not be the case for sediments, in which denitrification is the dominant source of N2O under variable oxygen tension, with nitrification only contributing in a narrow suboxic band. Attribution of source is difficult as nitrification and denitrification may occur simultaneously and interact, with exchange of products and intermediates (Figure 1). This is further complicated, as denitrification will be limited to some extent by nitrate supply from nitrification. Isotopic data from the surface ocean in oligotrophic regions imply that N2O originates primarily from nitrification. However, recent evidence from waters overlying oxygen-deficient intermediate layers suggests that the elevated surface mixed-layer N2O arises from coupling between the two processes, as the observed isotope signatures cannot be explained by nitrification or denitrification alone. An additional N2O source from the dissimilatory reduction of nitrate to ammonium is restricted to highly anoxic environments such as sediments. The oceanic N2O distribution is determined primarily by the oxygen and nutrient status of the water

column. Estuaries and coastal waters show elevated supersaturation in response to high carbon and nitrogen loading, and the proximity of sub-oxic zones in sediment and the water column. As a result the total marine N2O source tends to be dominated by the coastal region. The N2O flux from shelf sea sediments is generally an order of magnitude lower than estuarine sediments, although the former have a greater spatial extent. A N2O maximum at the base of the euphotic zone is apparent in shelf seas and the open ocean, and is attributed to production in suboxic microzones within detrital material. Oceanic surface waters generally exhibit low supersaturations (o105%), although N2O supersaturations may exceed 300% in surface waters overlying low oxygen intermediate waters and upwelling regions, such as the Arabian Sea and eastern tropical North Pacific. These ‘natural chimney’ regions dominate the open ocean N2O source, despite their limited surface area (Table 2). The surface N2O in upwelling regions such as the Arabian Sea originates in part from the underlying low-oxygen water column at 100– 1000 m, where favorable conditions result in the accumulation of N2O to supersaturations exceeding 1200%. N2O transfer into the surface mixed layer will be limited by vertical transport processes and a significant proportion of N2O produced at these depths will be further reduced to N2. The oceans account for 1–5 Tg N-N2O per annum (Tg ¼ 1 1012 g) or 6–30% of total global N2O emissions, although there is considerable uncertainty attached to this estimate (Figure 2). A recent estimate with greater representation of coastal sources has resulted in upward revision of the marine N2O source to 7–10.8 Tg N-N2O per annum; although this may represent an upper limit due to some bias from inclusion of estuaries with high N2O supersaturation. However, this estimate is in agreement with a total



165

Table 2 N2O and CH4 regional surface water supersaturations (from Bange et al., 1996; 1998) (supersaturation is >100%, undersaturation is o100% with equilibrium between atmosphere and water at 100%)

Estuaries Coastal/shelf Oligotrophic/transitional ocean Upwelling ocean

Surface % N2O saturation mean (range)

Surface % CH4 saturation mean (range)

607 (101–2500) 109 (102–118) 102.5 (102–104) 176 (108–442)

1230 (146–29 000) 395 (85–42 000) 120 (80–200) 200 (86–440)

N2O sources _1 (Tg N yr ) Oceans 3 (1_ 5) Soil natural 6 (3.3 _ 9.7)

Soil anthropogenic 3.3 (0.6 _ 14.8)

Biomass burning 0.5 (0.2 _1) Industrial 1.3 (0.7 _1.8)

½2

NO2 þ hn-NO þ O 3 P

½3

O þ O2 þ M-O3 þ M

½4

At high concentrations (>50 ppbv; ppbv ¼ parts per billion by volume), O3 in the atmospheric boundary layer becomes a toxic pollutant that also has important radiative transfer properties. The production of nitric acid from NO influences atmospheric pH, and contributes to acid rain formation. In addition, the oxidation of NO to the nitrate (NO3) radical at night influences the oxidizing capacity of the lower troposphere. Determination of the magnitude and location of NO sources is critical to modeling boundary layer and free tropospheric chemistry. NO cycling in the ocean has received limited attention, as a result of its thermodynamic instability and high reactivity. Photolysis of nitrite in surface waters occurs via the formation of a nitrite radical with the production of NO:

Livestock feed 2.1 (0.6 _ 3.1)

N2O sinks _1 (Tg N yr ) Atmospheric increase 3.9 (3.1_ 4.7)

Stratospheric photolysis 12.3 (9 _ 16) Figure 2 Atmospheric nitrous oxide sources and sinks (adapted from Houghton et al., 1995). Units: Tg ¼ 1 1012 g.

oceanic production rate of 11 Tg N-N2O per annum calculated from new production and nitrification.

Nitric Oxide (NO) Nitric oxide (NO) plays a central role in atmospheric chemistry, influencing both ozone cycling and the tropospheric oxidation capacity through reactions with hydroperoxy- and organic peroxy-radicals. When the NO concentration exceeds B40 pptv (pptv ¼ parts per trillion by volume) it catalyzes the production of ozone (O3): CO þ OH * þ O2 -HO2* þ CO2

HO2* þ NO-OH * þ NO2

½1

* þ HOH-NO þ OH þ OH NO 2 þ hn-NO2

This reaction may account for 10% of nitrite loss in surface waters of the Central Equatorial Pacific, resulting in a 1000-fold increase in dissolved NO at a steady-state surface concentration of 5 pmol dm3 during light periods (pmol ¼ 1 1012 mol). This photolytic production is balanced by a sink reaction with the superoxide radical (O 2 ) to produce peroxynitrite: O 2 þ NO- OONO

This reaction will be dependent upon steady-state concentration of the superoxide radical; however, as the reaction has a high rate constant, NO is rapidly turned over with a half-life on the order of 10–100 seconds.


166

AIR–SEA TRANSFER: N2O, NO, CH4, CO NO sources _1 (Tg N yr )

Soils 15 (10 _ 20) Aircraft emmisions 0.6

Biomass burning 5.5 (2.5 _ 8.5)

Lightning 6 (2 _ 8) Transport from stratosphere 0.5

Fossil fuel combustion 21

Figure 3 Atmospheric nitric oxide sources (from Graedel and Crutzen, 1992). Units: Tg ¼ 1 1012 g.

As with N2O, NO may also be produced as a byproduct or intermediate of denitrification and nitrification (Figure 1). NO production by soils is better characterized than in marine systems, and is significant both in terms of nitrogen loss and the global NO budget (Figure 3). The greater oxygen availability in soils limits reduction of NO via denitrification and so enhances NO efflux. Sediment pore water NO maxima have been attributed to denitrification, although, as this process also represents a sink for NO (Figure 1), this may reflect poising at an optimal redox potential for NO production. Conversely, the NO maximum in low oxygen intermediate waters in the east tropical North Pacific derives from nitrification. Current understanding of the oceanic NO distribution is that it is limited to the surface ocean and intermediate low oxygen water column. There is potential for higher NO concentrations in coastal and estuarine waters from sediment and photolytic sources, and nitrite photolysis to NO may also be significant in upwelling regions. Despite the short half-life of NO in surface waters, the maintenance of steady-state NO concentration suggests that photolytic production may support an, as yet unquantified, source of atmospheric NO. Surface concentrations in the Central Equatorial Pacific suggest that the oceanic NO source would not exceed 0.5 Tg N per annum, which is relatively insignificant when compared with other sources (Figure 3).

Methane (CH4) CH4 is the most abundant organic volatile in the atmosphere and, next to CO2, is responsible for 15% of the current greenhouse radiative forcing, with a direct radiative forcing of 0. 5 Wm2. CH4 reacts with OH and so limits the tropospheric oxidation capacity and influences ozone and other greenhouse

gases. The reaction with OH generates a feedback that leads to a reduction in the rate of CH4 removal. CH4 is a reduced gas which, paradoxically, is supersaturated in the oxidized surface waters of the ocean (see Table 2). CH4 is produced biotically and abiotically, although its oceanic distribution is controlled primarily by biological processes. Methanogenesis is classically defined as the formation of CH4 from the fermentation and remineralization of organic carbon under anoxic conditions. Methanogens require a very low reducing potential and are generally obligate anaerobes, although there is evidence that they can tolerate some exposure to oxygen. However, methanogens cannot utilize complex organic molecules and often coexist with aerobic consortia to ensure a supply of simple C1 substrates. Methanogens utilize formate, acetic acid, CO2, and hydrogen in sulfate-rich anoxic environments, although they are generally out-competed by sulfatereducing bacteria which have a greater substrate affinity. However, the methanogens can also utilize other noncompetitive substrates such as methanol, methylamines, and reduced methylated compounds, when out-competed for the C1 compounds. A significant fraction of CH4 is oxidized before exiting the marine system and so the oxidation rate is critical in determining the air–sea flux. This is accomplished by methanotrophs that obtain their carbon and energy requirements from CH4 oxidation under aerobic conditions via the following reactions: CH4 -CH3 OH -HCHO-HCOOC -CO2 methane - methanol - formaldehyde - formate - carbon dioxide

Methanotrophs are found in greater numbers in sediments than in oxic sea water, and consequently the oceanic water column CH4 oxidation is an order of magnitude lower than in sediments. Methanotrophs have a high inorganic nitrogen requirement and so methanotrophy is highest at the oxic–anoxic interface where ammonium is available. Anaerobic CH4 oxidation also occurs but is less well characterized. It is generally restricted to anaerobic marine sediments, utilizing sulfate as the only oxidant available, and is absent from anaerobic freshwater sediments which lack sulfate. A significant proportion of CH4 produced in anaerobic subsurface layers in sediments is oxidized during diffusive transport through the sulfate-CH4 transition zone by anaerobic oxidation and subsequently by aerobic oxidation in the overlying oxic layers. Anaerobic oxidation represents the main sink for CH4 in marine sediments, where it may account for 97% of CH4 production.



CH4 production is characteristic of regions with high input of labile organic carbon such as wetlands and sediments, but is usually restricted to below the zone of sulfate depletion. The oceanic CH4 source is dominated by coastal regions, which exhibit high CH4 fluxes as a result of bubble ebullition from anoxic carbon-rich sediments, and also riverine and estuarine input. Some seasonality may result in temperate regions due to increased methanogenesis at higher temperatures. The predominant water column source in shelf seas and the open ocean is CH4 production at the base of the euphotic zone. This may arise from lateral advection from sedimentary sources, and in situ CH4 production. The latter is accomplished by oxygen-tolerant methanogens that utilize methylamines or methylated sulfur compounds in anoxic microsites within detrital particles and the guts of zooplankton and fish. Lateral advection and in situ production may be greater in upwelling regions, as suggested by the increased CH4 supersaturation in surface waters in these regions. Oceanic CH4 concentration profiles generally exhibit a decrease below 250 m due to oxidation. Methanogenesis is elevated in anoxic water columns, although these are not significant sources of atmospheric CH4 due to limited ventilation and high oxidation rates. Other sources include CH4 seeps in shelf regions from which CH4 is transferred directly to the atmosphere by bubble ebullition, although their contribution is difficult to quantify. Abiotic CH4 originating from high-temperature fluids at hydrothermal vents also elevates CH4 in the deep and intermediate waters in the locality of oceanic ridges. A significant proportion is oxidized and although the contribution to the atmospheric CH4 pool may be significant in localized regions this has yet to be constrained. Hydrates are crystalline solids in which methane gas is trapped within a cage of water molecules. These form at high pressures and low temperatures in seafloor sediments generally at depths below 500 m. Although CH4 release from hydrates is only considered from anthropogenic activities in current budgets, there is evidence of catastrophic releases in the geological past due to temperature-induced hydrate dissociation. Although oceanic hydrate reservoirs contain 14 000 Gt CH4, there is currently no evidence of significant warming of deep waters which would preempt release. Other aquatic systems such as rivers and wetlands are more important sources than the marine environment. Shelf regions are the dominant source of CH4 from the ocean (14(11–18) Tg CH4 per annum), accounting for 75% of the ocean flux (Table 2). The ocean is not a major contributor to the atmospheric

167

CH4 sources _1 (Tg CH4 yr )

Other natural 35 (20 _ 90) Wetland 115 (55 _ 150) Anthropogenic (other) 275 (200 _ 350)

Oceans 14 (11_ 18) Anthropogenic (fossil fuel) 100 (70 _ 120)

CH4 sinks _1 (Tg CH4 yr )

Tropospheric OH oxidation 445 (360 _ 530)

Loss to stratosphere 40 (32 _ 48) Soils 30 (15 _ 45) Atmospheric increase 37 (35 _ 40)

Figure 4 Atmospheric methane sources and sinks (adapted from Houghton et al. (1995)). Units: Tg ¼ 1 1012 g.

CH4 budget, as confirmed by estimates of the oceanic CH4 source (Figure 4).

Carbon Monoxide (CO) The oxidation of CO provides the major control of hydroxyl radical content in the troposphere and limits the atmospheric oxidation capacity. This results in an increase in the atmospheric lifetime of species such as CH4, N2O, and halocarbons, and enhances their transfer to the stratosphere and the potential for subsequent ozone destruction. It has been suggested that decreasing stratospheric ozone and the resultant increase in incident ultraviolet (UV) radiation may increase marine production and efflux of CO, thereby generating a positive feedback loop. However, this may be compensated by a negative feedback in which increased UV reduces biological production and dissolved organic matter, so reducing the CO source. CO also influences tropospheric ozone by its interaction with NOx, and is a minor greenhouse gas with a radiative forcing of 0.06 Wm2 at current atmospheric concentrations.


168


The principal source of dissolved CO is the abiotic photodegradation of dissolved organic matter (DOM) by UV-R, and CO represents one of the major photoproducts of DOM in the ocean. Quantum yields for CO are highest in the UV-B range (280–315 nm) and decrease with increasing wavelengths. However, the UV-A (315–390 nm) and blue portion of the visible spectrum contribute to marine CO production as a greater proportion of radiation at these wavelengths reaches the Earth’s surface. Humics represent approximately half of the DOM and account for the majority of the chromophoric dissolved organic matter (CDOM), the colored portion of dissolved organic matter that absorbs light energy. The CO photoproduction potential of humics is dependent upon the degree of aromaticity. Terrestrial humics are characterized by an increased prevalence of phenolic groups, and addition of precursor compounds containing phenolic moieties to natural samples stimulates CO production. Direct photo-oxidation of humics and compounds containing carbonyl groups, such as aldehydes, ketones, and quinones, occurs via the production of a carbonyl radical during a-cleavage of an adjacent bond:R þ ðCOR0 Þ * -R0 þ CO RCOR0 þ hv-hskip200pt ðRCOÞ * þR0 -R0 þ CO CO production may also occur indirectly by a photosensitized reaction in which light energy is transferred via an excited oxygen atom to a carbonyl compound. This may occur with ketonic groups via the photosensitized production of an acetyl radical. Whereas light and CDOM are the primary factors controlling CO production there may also be additional influence from secondary factors. For example, organo-metal complexes have increased light absorption coefficients and their photo-decomposition will enhance radical formation and CO production at higher levels of dissolved metals such as iron. There is also minor biotic production of CO by methanogens, but this does not appear to be significant. CO can be oxidized to carbon dioxide by selected microbial groups including ammonia oxidizers and methylotrophs that have a broad substrate specificity and high affinity for CO. However, only the carboxidotrophs obtain energy from this reaction, and these may be unable to assimilate CO efficiently at in situ concentrations. CO turnover times of 4 hours are typical for coastal waters, whereas this varies between 1 and 17 days in the open ocean. The lower oxidation rate in the open ocean may be due to light inhibition of CO oxidation. Extrapolation from laboratory

measurements suggests that only 10% of photochemically produced CO is microbially oxidized. Dissolved CO exhibits diurnal variability in the surface ocean in response to its photolytic source, although this is also indicative of a strong sink term. The decline in the surface mixed-layer CO concentration in the dark results from a combination of CO oxidation, vertical mixing, and air–sea exchange. As the equilibration time between atmosphere and oceanic surface mixed layer is on the order of a month, this suggests that the former two processes dominate. Superimposed upon the diurnal cycle of CO in the surface ocean are spatial and seasonal gradients that result from the interaction of photoproduction and the sink processes. Below the euphotic zone CO is uniformly low throughout the intermediate water column. CO production potential is highest in wetland regions, which are characterized by high CDOM and enhanced light attenuation. Photochemical production of CO represents a potential sink for terrestrial dissolved organic carbon (DOC) in estuaries and coastal waters. This pathway may account for some of the discrepancy between the total terrestrial DOC exported and the low proportion of terrestrial DOC observed in the marine pool. Although a strong lateral gradient in CDOM exists between rivers and the open ocean, estuarine CO production may be limited by reduced UV light penetration. CO photoproduction may occur down to 80 m in the open ocean, and 20 m in the coastal zone, but is restricted to the upper 1 m in wetlands and estuaries. In addition, estuarine and coastal CO flux may also be restricted by the higher CO oxidation rates. There is evidence that upwelling regions may support enhanced CO production, in response to upwelled CDOM that is biologically refractory but photolabile. The presence of a CO gradient in the 10 m overlying the surface ocean suggests that the photolytic source of CO may influence the marine boundary layer. The marine source of CO is poorly constrained, with estimates varying from 10 to 220 Tg CO per annum. A flux of 1200 Tg CO per annum was estimated on the assumption that low rates of oceanic CO oxidation would only remove a small proportion of photoproduced CO, and that the residual would be ventilated to the atmosphere. The discrepancy between this and other flux estimates implies that a significant CO sink has been overlooked, although this may reflect shortcomings of different techniques. The oceanic contribution to the global source is between 1 and 20%, although Extrapolation of photochemical production rates from wetlands, estuaries, and coasts suggests that these



Table 3

169

Atmospheric CO sources and sinksa (adapted from Zuo et al., 1998)

CO sources (Tg CO y1) Industrial/fuel combustion Biomass burning Vegetation and soils Methane oxidation NMHC oxidation Ocean (Coast/Shelf Total sources

CO sinks (Tg CO y1) 400–1000 300–2200 50–200 300–1300 200–1800 10–220 300–400) 1260–6720

Tropospheric hydroxyl oxidation Soils Flux to stratosphere

1400–2600 250–530 80–140

Total sinks

1730–3270

a

Note that a separate estimate of the coastal/shelf CO source is shown for comparison, but does not contribute to the total source. Tg ¼ 1 1012 g.

alone may account for 20% of the total global CO flux. Although the marine source is responsible for o10% of the total global flux (Table 3), it may still dominate atmospheric oxidation conditions in remote regions at distance from land.

Air–Sea Exchange of Trace Gases The flux of these trace gases across the air–sea interface is driven by physical transfer processes and the surface concentration anomaly, which represents the difference between the partial pressure observed in surface water and that expected from equilibrium with the atmosphere. Direct determination of the oceanic emission of a trace gas is difficult under field conditions. Atmospheric gradient measurements above the ocean surface require enhanced analytical resolution, whereas more advanced micrometeorological techniques have yet to be applied to these trace gases. Determination of the accumulation rate in a floating surface flux chamber is a simpler approach, but may generate artefactual results from the damping of wave- and wind-driven exchange, and enhanced transfer on the inner chamber surfaces. Consequently the majority of flux estimates are calculated indirectly rather than measured. The surface anomaly is derived from the difference between the measured surface concentration (Cw), and an equilibrium concentration calculated from the measured atmospheric concentration (Cg) and solubility coefficient (p) at ambient temperature and salinity. This is then converted to a flux by the application of a dynamic term, the gas transfer velocity, k: F ¼ kðCw aCgÞ The transfer velocity k is the net result of a variety of molecular and turbulent processes that operate at different time and space scales. Wind is the primary driving force for most of these turbulent processes,

and it is also relatively straightforward to obtain accurate measurements of wind speed. Consequently, k is generally parameterized in terms of wind speed, with the favored approaches assuming tri-linear and quadratic relationships between the two. These relationships are defined for CO2 at 201C in fresh water and sea water and referenced to other gases by a Schmidt number (Sc) relationship: k gas ¼ k ref ðSc gas=Sc ref Þn where n is considered to be 1/2 at most wind speeds. This dependency of k is a function of the molecular diffusivity (D) of the gas and the kinematic viscosity of the water (m), and is expressed in terms of the Schmidt number (Sc ¼ m/D). Determination of marine trace gas fluxes using different wind speed–transfer velocity relationships introduces uncertainty, which increases at mediumhigh wind speeds to a factor of two. Furthermore, additional uncertainty is introduced by the extrapolation of surface concentration gradient measurements to long-term climatological wind speeds. Current estimates of oceanic fluxes are also subject to significant spatial and temporal bias resulting from the fact that most studies focus on more productive regions and seasons. This uncertainty is compounded by the extrapolation of observational data sets to unchartered regions. With the exception of N2O, the ocean does not represent a major source for these atmospheric trace gases, although spatial variability in oceanic source strength may result in localized impact, particularly in remote regions. In the near future, advances in micrometeorological techniques, improved transfer velocity parameterizations and the development of algorithms for prediction of surface ocean concentrations by remote sensing should provide further constraint in determination of the oceanic source of N2O, NO, CH4, and CO.


170


See also Air–Sea Gas Exchange. Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens. Carbon Dioxide (CO2) Cycle. Gas Exchange in Estuaries. Photochemical Processes. Plankton and Climate. Surface Films. Upwelling Ecosystems.

Further Reading Bange HW, Bartel UH, et al. (1994) Methane in the Baltic and North Seas and a reassessment of the marine emissions of methane. Global Biogeochemical Cycles 8: 465--480. Bange HW, Rapsomanikis S, and Andreae MO (1996) Nitrous oxide in coastal waters. Global Biogeochemical Cycles 10: 197--207.

Carpenter EJ and Capone DG (eds.) (1983) Nitrogen in the Marine Environment. London: Academic Press. Graedel TE and Crutzen PJ (eds.) (1992) Atmospheric Change: An Earth System Perspective. London: W. H. Freeman and Co. Houghton JT, Meira Filho M, Bruce J, et al. (1995) Climate Change 1994. Radiative Forcing of Climate Change and an Evaluation of the IPCC IS92 Emission Scenarios, Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press. Liss PS and Duce RA (eds.) (1997) The Sea Surface and Global Change. Cambridge: Cambridge University Press. Zuo Y, Guerrero MA, and Jones RD (1998) Reassessment of the ocean to atmosphere flux of carbon monoxide. Chemistry and Ecology 14: 241--257.


ALCIDAE T. Gaston, National Wildlife Research Centre, Quebec, Canada Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 145–151, & 2001, Elsevier Ltd.

Introduction The auks are seabirds of the northern hemisphere. They form a family (Alcidae) or subfamily (Alcinae) of the order Charadriiformes, which includes the gulls, terns, jaegers, sandpipers, and plovers. They are often considered the northern hemisphere equivalents of the penguins, being well-adapted to underwater swimming. However, all extant species retain the power of flight. They are almost entirely confined to Arctic, Subarctic and Boreal waters and are marine throughout the year, being most common in waters of the continental shelf and slope. Twentythree species are currently recognized and one other became extinct in historic times (great auk Pinguinus impennis). Six species of five genera occur in the North Atlantic and 20, of 10 genera in the North Pacific, south of the Chukchi Sea. Two murres (Uria spp.), the dovekie (Alle alle) and the black guillemot (Cepphus grylle) are circumpolar in distribution. Many species of auks are extrememly abundant, with common and thick-billed murres (Uria aalge, U. lomvia), least and crested auklets (Aethia pusilla, A. cristatella), dovekies and Atlantic puffins (Fratercula arctica) all having world populations of 410 million individuals. They form the dominant avian biomass over large areas of Arctic and Subarctic waters throughout the year and may be significant predators on large zooplankton and small fishes in areas around their breeding colonies.

Miocene, about 15 million years ago (Ma), although putative proto-auks have been described from as far back as the Eocene. A subfamily of flightless auks, the Mancallinae, was present in the Pacific during the late Miocene (7–4 Ma). Deposits from California have yielded at least five species of the flightless Mancalla, ranging in size from about 1 to 4 kg. Auks probably originated in the Pacific, where all but two (Alca, Alle) of the extant genera are widely distributed, colonizing the Atlantic soon after. Since then, Pacific and Atlantic auks evolved largely in isolation. The Pliocene auk fauna from California was similar in diversity to the recent fauna, but the diversity of auks represented in the Pliocene sediments of eastern North America suggests that the current, relatively small, community of auks in the Atlantic (six species, including the great auk) is not a historical consequence of the lack of connections with the Pacific. Instead, it is a consequence of extinctions during the Pleistocene, presumably as a result, directly or indirectly, of the ice ages.

Characteristics and Adaptations Auks are well-adapted to underwater swimming. They have compact, streamlined bodies, short wings and very short tails. The feet are placed far back on the body, webbed, with no hind toe; claws are narrow and the tarsus is laterally compressed. The bill is variable in shape and may be highly ornamented in the breeding season. There are 11 primary wing feathers and 16–21 secondaries, the outermost primary being very small and the longest being, usually, the tenth. The feather tracts of the back and belly are Table 1 Tribes and genera of the Alcidae (after Gaston and Jones 1998)

Evolution and Systematics

Tribe

Genus

Recent work on mitochondrial DNA and allozymes divides the family Alcidae into 5 tribes and 12 genera (Table 1), with the puffins (Fraterculini) and auklets (Aethiini) being sister tribes, more closely related to one another than to the other auks. Otherwise, the molecular data suggest that the tribes originated from an initial rapid divergence among early members of the family, possibly about 10–12 million years ago. There is extensive fossil material on the family. The first definite records of auks are from the middle

Fraterculini

Fratercula Cerorhinca Aethia Cyclorhynchus Ptychoramphus Brachyramphus Cepphus Synthliboramphus Alle Pinguinus Alca Uria

Aethiini

Brachyramphini Cepphini Alcini


171

172

ALCIDAE

continuous and, beneath the contour feathers, down feathers are present and dense all over the body. There are 6–8, occasionally 9, pairs of tail feathers. The fusiform body, with the feet set far back, is common to all underwater swimming birds. Other characters that are common to underwater swimmers, but absent in Charadriiformes that do not dive, are the presence of strongly developed vertebral hypophyses on the last cervical vertebra, and the enlarged number of thoracic vertebrae (8–10 compared to 5–7 in other Charadriiformes). An increase in the number of vertebrae allows for a longer body, while retaining flexibility of movement. Auks fly with rapid wing-beats and without gliding or soaring, using the spread feet for steering and braking. They generally take off from land with difficulty and use the feet to taxi when taking off from water. Underwater, they swim using the wings as paddles, like penguins. Maneuvering underwater is achieved mainly by asymmetrical strokes of the wings. Air is released from plumage before diving by forcing it from the breast feathers using subcutaneous muscles. On land, the Alcini and Brachyramphini rest on the belly, or on the tarsi, rising to their feet only in walking, whereas Aethiini and Fraterculini are more agile and normally stand with the tarsus erect, rather than horizontal. The anatomy of the auks is similar to that of other Charadriiformes, except for their specializations towards underwater swimming. They have shorter wings and legs than gulls and shore birds, and the relative length of the humerus is reduced. The covert feathers are stiffer and more extensive, both above and below the wing, an arrangement that reinforces the trailing edge of the wing and closes gaps between adjacent flight feathers. Primaries 6–10 form a closely knit unit with little independent movement, making the wing more effective as an underwater paddle. The rigidity is caused by a greater development of connective tissue, compared to other Charadriiformes. Despite the fact that the primary feathers are very stiff, the leading edge of the wing is very curved on the downstroke, presumably because of water resistance. The articulating surface of the humerus is much larger in auks than in other Charadriiformes, allowing greater force to be transmitted in swimming. However, like other flighted birds, but unlike penguins, the auks retain considerable flexibility of movement at the articulation of the humerus. The supracoracoideus muscle, which raises the wing during the recovery stroke (2–3% of body mass) is larger than in shore birds and gulls (1.5%) and the pectoralis muscles, the main source of power on the downstroke, although similar in size to those of gulls

and shorebirds, are more elongated, with the supracoracoideus lying directly below them. This arrangement improves streamlining, and also improves insulation by distributing stored fat in a thin subcutaneous layer over most of the body, rather than depositing it in discrete pockets, as occurs in many birds. Underwater, auks swim very jerkily, with a rapid acceleration at each downstroke. Hence, forward propulsion on the upstroke, if any, is relatively small. The wings are held sharply bent at the wrist, reducing the functional surface area to slightly less than half that when spread in flight. The fact that the whole wing area is unnecessary for effective underwater propulsion is emphasized by the fact that all genera except Aethia and Cyclorhynchus lose their flight feathers simultaneously during the annual postbreeding moult, becoming flightless for a month or more. The auks have developed many physiological adaptations for prolonged diving, including high blood volume and high levels of myoglobin (41% body mass), both of which enhance the bird’s ability to store oxygen. Despite these enhancements, many long dives undertaken by thick-billed murres, and probably other auks, exceed the estimated limit for aerobic respiration, requiring the birds to respire anaerobically for a portion of the time underwater. Heart rate and peripheral blood flow also may be restricted. The digestive system includes a well-developed proventriculus, a muscular stomach and a relatively short intestine. A functional crop is present only in auklets, although there is some croplike development of the lower proventriculus in the puffins. Auklets and dovekies develop diverticulae in the throat while breeding, in which they transport food for their young. Nasal glands, which are used to excrete salt and maintain ion balance, are very well developed. The dominant plumage color of auks is black above and white below, but Cepphus species are mainly black in summer, mainly white in winter, and Brachyramphus species are mottled brown in summer (Figure 1). Temporary nuptial ornaments, including ornamental plumes on the head and deciduous, brightly colored, horny plates on the bill, are present in the Aethiini and Fraterculini. Sexual dimorphism is very small: males of most species are slightly larger in some dimensions, especially bill depth. There is no plumage dimorphism and no distinctive immature plumage. First-winter birds generally appear similar to winter-plumage adults, although young Cepphus are readily distinguished by their greater mottling.


ALCIDAE

173

Figure 1 Examples of the Alcidae. (1) Razorbill (Alca torda). Adult breeding plumage. Length: 42 cm; approximate body mass: 710 g. Range: North Atlantic Ocean. (2) Brunnich’s guillemot (Uria lomvia). Other name: thick-billed murre. (a) Adult breeding plumage; (b) adult non-breeding plumage; (c) transitional plumage. Length: 40 cm; approximate body mass: 950 g. Range: North Pacific, North Atlantic and Arctic Oceans. (3) Least auklet (Aethia pusilla). Length: 15 cm; approximate body mass: 85 g. Range: North Pacific Ocean. (4) Marbled murrelet (Brachyramphus marmoratus). Adult breeding plumage. Length: 25 cm; approximate body mass: 235 g. Range: eastern North Pacific Ocean. (5) Crested auklet (Aethia cristatella). (a) adult breeding plumage; (b) adult, non-breeding plumage. Length: 25 cm; approximate body mass: 260 g. Range: North Pacific Ocean. (6) Atlantic puffin (Fratercula arctica). Other name: puffin. Adult, breeding plumage. Length: 35 cm; approximate body mass: 470 g. Range: North Atlantic Ocean. (7) Little auk (Alle alle). Other name: dovekie. Adult breeding plumage. Length: 20 cm; approximate body mass: 165 g. Range: North Pacific and North Atlantic Oceans, Arctic Ocean. Illustrations from Harrison P (1985) Seabirds, an identification guide. Revised edition. Boston, Massachusetts: Houghton Mifflin.


174

ALCIDAE

Diet and Foraging Ecology The diet of the smaller auks consists largely of zooplankton, especially copepods and euphausiids and larval fishes, although the parakeet auklet Cyclorhynchus psittacula takes substantial quantities of jellyfish medusae. The larger auks take a mixture of zooplankton and small fishes and squid, with common murres and Cepphus spp., and possibly Brachyramphus spp., being mainly piscivorous. Nestling Uria, Alca, Brachyramphus, Cepphus, Cerorhinca, and Fratercula species are all fed predominantly on fish. Auks that feed mainly on fish tend to have narrower, more pointed bills (low width/gape ratio), and narrower, more cornified, tongues than those that feed on plankton. Planktivorous auks, apart from having broader, shorter bills, have large numbers of fleshy denticles on the palate and the upper surface of the tongue. The convergence of these characters for auks feeding exclusively on plankton is dramatically demonstrated by the parallels between the auklets and the dovekie. These adaptations are also shown within the genus Uria, with the predominantly fish-eating common murre having a narrower tongue and fewer palatal denticles than the thick-billed murre, which feeds on a greater variety of prey. Bill depth seems to be unrelated to diet, suggesting that the very deep bills of puffins and auklets may have evolved for secondary sexual purposes, rather than as feeding adaptations. Auks feed predominantly in continental shelf waters, with species of Brachyramphus and Cepphus feeding entirely in inshore waters and most other genera occurring mainly within 50–100 km of land. However, Fratercula species are found far offshore in winter, including waters beyond the continental shelf. All species feed entirely underwater, with the smaller species diving to maximum depths of 30–40 m and the larger species to over 100 m. The deepest dives are achieved by Uria species, which may reach 200 m on occasions. Normal foraging depths for smaller auks are 10–30 m and for larger auks 30–60 m. Many species take advantage of prey aggregations caused by oceanographic fronts and tidally induced upwellings. Striking examples occur in the passes among Aleutian Islands and other complex archipelagos, where feeding auklets may reach great densities. The prey involved are usually slow-swimming zooplankton such as pteropod mollusks, copepods, euphausiids, and amphipods.

Reproduction Breeding sites are coastal, except in the marbled and Kittlitz’s murrelets, which are also the only solitary

nesters, and in the dovekie. Most auks breed on islands, and some species breed exclusively on remote islands well offshore, because islands offer refuge from terrestrial predators and at the same time may be close to foraging areas. Hence, the distribution of breeding auks is much influenced by the distribution of suitable breeding islands. Most species are highly social while breeding, with a prolonged prebreeding period and extensive courtship activity in which both vocal and visual displays are prominent. All species are socially monogamous and the sexes play equal roles in incubation and in rearing chicks to the age when they leave the colony. Most show a strong tendency to return to the colony where they were reared. Auks are highly variable in the type of nesting site that they use. None builds a nest: the Aethia auklets and some Synthliboramphus murrelets use crevices under boulders or among scree, and murres lay their eggs on open cliff ledges or on flat rocky islets (common murre), while puffins, ancient murrelet, and Cassin’s auklet often dig extensive burrows in soil. Most of the smaller auks make use of sites that protect them from surface predators, such as gulls and crows: only the largest, the murres, nest in the open in colonies, where their densely packed ranks form a defense against nest predators. Cliffs are frequently used, especially in the Arctic. The other open nesters, the Brachyramphus murrelets, are solitary nesters, either on the horizontal limbs of mature trees (marbled murrelet Brachyramphus marmoratus, commonly) or on the ground on remote mountain tops (Kittlitz’s murrelet B. brevirostris). Like many other small seabirds, some of the auks are nocturnal in their coming and goings to their breeding sites. The Synthliboramphus murrelets and whiskered and Cassin’s auklets (Aethia pygmaea, Ptychoramphus aleuticus) are invariably nocturnal, and the rhinoceros auklet (Cerorhinca monocerata) is largely nocturnal, but diurnal or crepuscular in some parts of its range. Brachyramphus species are normally crepuscular. Everything about the breeding strategies of the auks suggest the dominant influences of predation and kleptoparasitism. Egg formation is a lengthy process and females spend most of their time away from the colony during the last 10 days before they lay. Clutches consist of one or two eggs and only one brood is reared annually (California populations of Cassin’s auklet sometimes rear two). Eggs are thick-shelled and either white, buff, tan, or bluish in ground colour, sometimes with prominent black or brown markings. After clutch completion, the sexes alternate incubation duty, taking equal shares. Laying, in practically all populations, is confined to a span of


ALCIDAE

about 6 weeks, with most eggs being laid within 2–3 weeks. Incubation periods range from 29 to 45 days. The time from the start of incubation to chick departure varies from a minimum of 35 days in the precocial Synthliboramphus species to a maximum of more than 80 days in the Cerorhinca. Most species, including the puffins, auklets and razorbill (Alca torda) that lay only a single egg, have two lateral brood patches. The murres and the dovekie have a single, central patch, set far back on the belly. Brood patches are a potential site of heat loss. They defeather rapidly at the start of incubation, refeather as soon as hatching occurs, and are small relative to the size of the eggs, so that during incubation only part of the egg’s surface is in contact with the patch. Nonbreeders either do not form brood patches or develop only partial patches. Notwithstanding these adaptations, the insulation of the auks appears to be relatively poor, considering the climates in which they live. To maintain normal avian body temperature, they rely principally on a very high rate of metabolism: basal metabolic rates (the metabolic rate of resting, nondigesting birds) among auks are exceptionally high, compared to those of other birds of similar size. Hatchlings are covered with a dense, woolly, down plumage. They are active within 1–2 days after hatching and capable of thermoregulation either immediately (Synthliboramphus) or within a few days (the rest). All genera except Synthliboramphus are semiprecocial and chicks are reared at the nest site for a minimum of 2 weeks. There is no postfledging parental care except in the dovekie, although in Synthliboramphus, Uria, and Alca the young are cared for until some time after departing the colony, which they leave when only partly grown and incapable of sustained flight. Most breeders leave the colony either with their chicks (Alcini and Synthliboramphus spp., where there is parental care after departure), or within 1–2 weeks following chick departure. Juveniles, family parties, and postbreeding adults disperse rapidly from the colony area (except young whiskered auklets).

Postbreeding and Wintering A complete prebasic molt (except in the puffins, where it involves the body plumage only) follows rapidly on the termination of breeding and involves the shedding of nuptial ornaments and the adoption of a generally distinctive nonbreeding (winter) plumage (no change in Craveri’s and Xantus’ murrelets). In the auklets, primary molt is sequential and begins during chick-rearing, but in other species the primaries are dropped simultaneously, making flight

175

impossible for a period. Most species remain scattered offshore while molting. Following the annual prebasic molt, most species disperse or shift toward distant wintering areas. However, some populations of common murres and black guillemots return to the area of the breeding colony and commence periodic attendance at the breeding site. This also applies to some populations of marbled murrelets. These birds feed in the same areas practially year-round. However, among most species breeding in Arctic and Subarctic waters, the bulk of the population moves substantially farther south. Young birds tend to disperse farthest and there is usually a disproportionately high representation of first-year birds in ‘wrecks’; – the periodic casting ashore of large numbers of weakened birds, often during prolonged storms. Wintering areas and the behavior of auks on the wintering grounds have been less studied than their activities during the breeding season. Detailed studies of feeding ecology in winter have mostly been carried out on species and populations that occur in inshore waters and the winter ecology of many is essentially unknown. Movement toward the breeding colonies begins in February– April, depending on latitude.

Population Dynamics Annual survival of adult auks is generally greater than 85%, and greater than 95% in some populations of common murres and Atlantic puffins (Fratercula arctica), making them among the longestlived birds. In the longer-lived species, average age at first breeding is 5 years or more, whereas Cassin’s auklets, ancient murrelets, and probably some Aethia spp. may begin to breed at 2, many at 3 years. Populations contain substantial numbers of nonbreeders. Those in their second summer or older often attend the breeding colony to select mates and breeding sites. Reproductive success increases with age for the first few years of breeding. Comparisons of reproductive success are complicated by the fact that different species depart from the colony at different stages of development. The maximum productivity is about 1.5 young/pair per year and most average o1.

Auks and People Auks and their eggs have been harvested by people from the earliest times and their bones are frequent constituents of middens throughout the coastal areas of the northern hemisphere. The remains of the great auk have been discovered in 40 middens in Norway alone, as well as some in Britain, Iceland, Greenland, and the United States. Excavations in Newfoundland


176

ALCIDAE

dating to 4000 BP contain many Great Auk bones and they are also found in middens in Florida dating to 3000 BP. All major colonies of thick-billed murres in the eastern Canadian Arctic show traces of Eskimo occupation nearby, usually with associated remains of thick-billed murres. In areas surrounding the Straits of Georgia, British Columbia, common murre remains are widespread in Indian middens dating from the pre-European period, while middens in the Aleutians, some dating to 4000 BP, contain the remains of auklets and puffins, as well as murres. A variety of techniques were developed for catching auks. One of the most widespread, used by Bering Sea Inuit, Icelanders, and Faeroese to catch puffins, and by the Inuit of north-west Greenland to catch dovekies, was a net at the end of a long pole. The hunter sheltered behind a stone wall, or depression in the ground, and suddenly raised the net in the path of low-flying birds, which were unable to turn in time to avoid it. In Iceland, snares placed on rafts floating offshore were also used to good effect, trapping murres, puffins, and razorbills. The technique of netting birds flying over the colony is a very efficient way to harvest auks, as the prebreeding component of the population often circles constantly, making them much more vulnerable than the breeders. Similarly, snares placed on boulders or floating rafts used by displaying birds, a method used on St. Lawrence Island to capture least and crested auklets, and in Thule, Greenland, to capture dovekies, select mainly prebreeders. The removal of these birds has much less effect on the population than the killing of breeders and may partly explain how early societies managed to coexist successfully with their prey. It has been estimated that 150 000–200 000 Atlantic puffins were taken annually in Iceland, without any apparent effect on population levels. In contrast, the shooting of breeding birds at colonies that became widespread in Greenland in this century has been the main cause of the drastic decline suffered by thick-billed murre populations there over the past 50 years. The harvesting of auk eggs has been very common throughout their range and may well have been a factor controlling their distribution on inshore islands accessible to permanent human settlements. On St. Kilda, harvesting of puffin and murre eggs was a regular activity, while in the Queen Charlotte Islands, and throughout the Alaskan islands, the excavation of ancient murrelet and auklet burrows for eggs was a routine spring harvest. Thick-billed murre eggs are harvested in Greenland, Canada, the Pribilof Islands, and Russia. In addition to their use as food, the skins and ornaments of certain auks were valued for clothing and

decoration. Inuit on St. Lawrence Island and Aleuts in the Aleutian chain sewed parkas out of auk skins, especially crested auklets and horned puffins. Elsewhere, puffin and dovekie skins were sewn into inner garments, to be worn under furs. In north-west Greenland, dovekie skins were made into undershirts, while farther south, in Upernavik District, murre skins were made into capes. Dovekie skins had to be softened by chewing; only elderly women did this, as their teeth were worn smooth enough not to damage the delicate skins. The spectacular beaks of puffins and auklets were also used as ornaments by the Aleuts and Inuit of the Bering Sea region, hundreds sometimes being sewn on the outside of a garment, along with the golden crests of tufted puffins. Commercial exploitation of auk colonies by postindustrial societies resulted in substantial declines. In the Gulf of St. Lawrence, the huge auk colonies visited by Audubon in 1827 were reduced to a mere remnant by the late nineteenth century, while at Funk Island, Newfoundland, the great auk was exterminated largely for its feathers. The common murre population at the Farallon Islands, California, and several large thick-billed murre populations in the Russian Far East and Novaya Zemlya were decimated by egg harvesting. Substantial harvesting of auks still occurs in several areas. Traditional harvests of murres and puffins continue in Iceland and the Faroes, although reduced from former levels. Relatively small numbers of thick-billed murres are taken by Inuit in the Canadian Arctic, although they form important components of the summer diet at a few settlements. Much larger numbers are taken in West Greenland, where regulations prohibiting the shooting of birds at their colonies were introduced only in 1978 and were still more or less unenforced in 1987. Shooting away from the colonies is still permitted throughout the year in some districts, although subject to seasonal limits in the more populated areas. The same populations are hunted more heavily off Newfoundland and Labrador in winter, with the annual kill estimated at about 200 000 since 1993. Although it is only legal to kill murres, some razorbills and dovekies, and a few puffins, are also shot. Although direct harvests have affected several auk species and were responsible for the extermination of the great auk, the effects of mammalian predators, introduced either deliberately, or accidentally, have probably had a much greater impact on auk populations worldwide. The main agents of destruction were foxes, introduced throughout the Alaskan islands for fur farming. Rats have also caused many declines and extirpations. Raccoons and mink have an important impact in some areas, and rabbits,


ALCIDAE

through their effects on vegetation and soil, may also have caused problems for some burrowing species. Japanese, Craveri’s, Xantus’ and marbled murrelets are all considered endangered or threatened in one way or another. It is certain that the majority of auk populations are smaller, in many cases much smaller, than they would have been a few centuries ago. Probably, we will see little change in that situation, although programs to eliminate introduced predators from certain important Pacific islands may improve the situation for some species. All auks are very susceptible to contamination by oil and they have formed the majority of seabirds killed in oil spills off Europe and North America. Unlike gulls, they have not profited at all from fisheries wastes. Protection from egging has led to increases of some species in the twentieth century. However, overall, the auks remain precariously dependent on human goodwill for their future survival.

177

Dynamics. Seabirds and Fisheries Interactions. Seabirds as Indicators of Ocean Pollution.

Further Reading Gaston AJ and Jones IL (1998) The Auks:: Family Alcidae. Oxford: Oxford University Press. Birkhead TR and Nettleship DN (1985) The Atlantic Acidae. London: Academic Press. Harris MP (1989) The Puffin. London: T & AD Poyser. Gaston AJ and Elliot RD (eds.) (1991) Conservation Biology of the Thick-billed Murre in the Northwest Atlantic. Ottawa: Canadian Wildlife Service. Sealy SG (ed.) (1990) Auks at Sea. Studies in Avian Biology 14, Cooper Ornithological Society. Friesen VL, Baker AJ, and Piatt JF (1996) Phylogenetic relationships within the Alcidae (Aves: Charadriiformes) inferred from total molecular evidence. Molecular Biology and Evolution 13: 359--367.

See also Fish Predation and Mortality. Laridae, Sternidae and Rynchopidae. Network Analysis of Food Webs. Plankton. Seabird Conservation. Seabird Foraging Ecology. Seabird Migration. Seabird Population


ANTARCTIC CIRCUMPOLAR CURRENT S. R. Rintoul, CSIRO Antarctic Climate and Ecosystems Cooperative Research Centre, Hobart, TAS, Australia & 2009 Elsevier Ltd. All rights reserved.

Introduction The Drake Passage between the South American and Antarctic continents is the only band of latitudes where the ocean circles the Earth, unblocked by land. The existence of this oceanic channel has profound implications for the global ocean circulation and climate. The Drake Passage gap permits the Antarctic Circumpolar Current (ACC) to exist, a system of ocean currents which flows from west to east along a roughly 25 000-km-long path circling Antarctica. In terms of transport, the ACC is the largest current in the world ocean, carrying about 13778 106 m3 s 1 through the Drake Passage. The wind-driven ocean circulation theories that explain much of the ocean current patterns observed at other latitudes do not apply in an unbounded channel and the unique dynamics of the ACC have long been a puzzle for oceanographers. The threedimensional circulation in the ACC belt is now understood to reflect the interplay of wind and buoyancy exchange with the atmosphere, water mass modification, eddy fluxes of heat and momentum, and strong interactions between the flow and bathymetry. The strong eastward flow of the ACC has several important implications for the global ocean circulation and its influence on regional and global climate. By transporting water between the major ocean basins, the ACC tends to smooth out differences in water properties between the basins. The interbasin connection allows a global-scale pattern of ocean currents to be established, known as the thermohaline circulation (see Ocean Circulation: Meridional Overturning Circulation), which transports heat, moisture, and carbon dioxide around the globe and strongly influences the Earth’s climate. The strong flow of the ACC is associated with steeply sloping density surfaces, which shoal to the south across the current and bring dense waters to the surface in the high-latitude Southern Ocean. Where the dense waters are exposed at the sea surface, they exchange heat, moisture, and gases like oxygen and carbon dioxide with the atmosphere. In this sense, the Southern Ocean provides a window to the deep sea. The ecology and biogeography of the Southern Ocean

178

are influenced strongly by the ACC and the overturning circulation plays an important role in the marine carbon and nutrient cycles. The west-to-east flow of the ACC inhibits north–south exchanges across the current and isolates Antarctica from the warm waters to the north; the present glacial climate of Antarctica was not established until the South American and Antarctic continents began to drift apart about 30 Ma, opening a circumpolar channel.

Structure of the Antarctic Circumpolar Current A schematic view of the major currents of the Southern Hemisphere oceans south of 201 S is shown in Figure 1. The flow of the ACC is focused in several jets, associated with sharp cross-stream gradients (or fronts) in temperature, salinity, and other properties. The three main fronts of the ACC – the Subantarctic Front, Polar Front, and southern ACC front – are indicated by the arrows circling Antarctica. To the south of the circumpolar flow of the ACC, clockwise gyres are found in the Weddell Sea, Ross Sea, and the Australian–Antarctic Basin (see Current Systems in the Southern Ocean). A westward flow associated with the Antarctic Slope Front and Antarctic Coastal Current is found near the continental shelf break around much of Antarctica. To the north of the ACC, water flows to the east in the southern limb of the large anticlockwise subtropical gyres in each basin. Exchanges of water masses between the ACC and the gyre circulations to the north and south are important components of the global circulation. The distribution of water properties on a transect crossing the ACC is illustrated in Figure 2. The temperature and salinity of water masses are largely set at the sea surface, where there is active exchange with the atmosphere; nutrient and oxygen concentrations are also influenced by biological processes. Water masses carry these characteristics with them as they sink from the surface into the ocean interior. The major water masses of the Southern Ocean are associated with various property extrema that reflect their circulation and formation history. For example, the Subantarctic Mode Water is formed by deep convection in winter on the northern flank of the ACC, producing deep well-mixed layers that are rich in oxygen. The Antarctic Intermediate Water is the name given to the prominent salinity minimum layer north of the ACC. The Circumpolar Deep Water


ANTARCTIC CIRCUMPOLAR CURRENT

179

90° E

12

°E

Leeuwin

ulhas

Ea s C st t

30

Ag

S

e la

CC rn A

de ed W

40° S

Ross G

60° S

ll

Au

180° W

0°

n gu

Su ba nta rc Po ti lar F

C AC the ou

E

c

Be

0°

°E 150

°E

60

W

s ina

120

30

°W

ACC

0°

15

Mal v

Brazil

°W

°W

60

90° W Figure 1 A schematic view of the major ocean currents of the Southern Hemisphere oceans south of 201 S. Depths shallower than 3500 m are shaded. C, current; G, gyre; F, front; ACC, Antarctic Circumpolar Current.

(CDW) is often divided into two layers: Upper CDW corresponds to the oxygen minimum layer, and Lower CDW corresponds to the deep salinity maximum layer. The relatively fresh layer near the seafloor is Antarctic Bottom Water, which sinks near Antarctica and carries water rich in oxygen and chlorofluorocarbons into the deep ocean (see Bottom Water Formation). Water properties at a given depth change dramatically as the Southern Ocean is crossed from north to south (Figure 2). Surfaces of constant temperature, salinity, density, and other properties slope upward to the south. As a result, density layers found at 3000-m depth in subtropical latitudes approach the sea surface near Antarctica. The shoaling of density surfaces to the south is associated with the strong eastward flow of the ACC. Tilted density surfaces create pressure forces which drive ocean currents and an accompanying Coriolis force to balance the pressure force. (This balance of forces,

known as geostrophy, describes the dynamics of all large-scale ocean currents (see Ocean Circulation).) In the Southern Hemisphere, an increase in density to the south supports an eastward flow (relative to the seafloor), as observed in the ACC. The rise of temperature, salinity, and density surfaces to the south occurs in a series of steps, or rapid transitions, rather than as a uniform slope across the Southern Ocean (Figure 2). These rapid transitions are known as fronts. Because the strength of an ocean current is proportional to the magnitude of the horizontal density gradient, each of the fronts is associated with a maximum in velocity. Most of the flow of the ACC is concentrated in the fronts, with smaller transports observed between the fronts (Figure 2(f)). The zones between the fronts also coincide with regions of relatively uniform water properties at each depth. Unlike many fronts at lower latitudes, the ACC fronts extend from the sea surface to the seafloor. The current jets associated


180


FN

FS

SA

SA

PF -N

AS SB F sA C

PF -S

C sA f-S C C f-N

(a)

0m -0.8

3

-1.4

13

10

8

12

1

250

0

2

1.8

9

6 2.6

500

2.3

2.2 1.6

750

SAMW

5

0.8

8

4

2.4

0.6 0.4

2.3

1.4

6

1000 65° S

0m

60° S

50° S

55° S

45° S

-1.4

10 2

500

12

9

6

SAMW

2.3

8

1000

5

1500

3

2.2

2

1.6

4

0 2.4

2.6

1.4

2000

0.8

0.4

2 1

2500 3000

1.8 −0.2

0.6

0.2

0.8

3500 0

4000

−0.4

4500 5000 0 km

500

1000

1500

2000

Figure 2 Property distributions along a roughly north–south section across the Southern Ocean at 1401 E south of Australia (World Ocean Circulation Experiment section SR3): (a) potential temperature (1C), (b) salinity (PSS78), (c) neutral density (kg m 3), (d) oxygen (mmol kg 1), (e) chlorofluorocarbon 11 (CFC-11) (pM kg 1), and (f) transport at each station pair (solid line, left axis) and cumulative transport from south to north (dashed line, right axis). Transport is in units of sverdrups (1 Sv ¼ 106 m3 s 1). Contours slope upward from north to south; regions where the slope of the contours is steep correspond to the ACC fronts and to transport maxima. SAMW, Subantarctic Mode Water; LCDW, Lower Circumpolar Deep Water; AAIW, Antarctic Intermediate Water; UCDW, Upper Circumpolar Deep Water; CDW, Circumpolar Deep Water CDW; AABW, Antarctic Bottom Water. SAF, Subantarctic Front; PF, Polar Front; sACCf, southern ACC front; SB, southern boundary; ASF, Antarctic Slope Front; N and S indicate northern and southern branches of the primary fronts. The positive and negative peaks in transport labeled SAZ indicate a strong recirculation in the Subantarctic Zone north of the ACC at the time this section was occupied. Sections are adapted from Orsi AH and Whitworth T, III (2005) In: Sparrow M, Chapman P, and Gould J (eds.) Hydrographic Atlas of the World Ocean Circulation Experiment (WOCE), Vol. 1: Southern Ocean. Southampton: International WOCE Project Office (ISBN 0-904175-49-90), with permission (http://www.soc.soton.ac.uk).

with the fronts therefore also extend throughout the water column. The deep-reaching nature of the ACC fronts reflects the weak stratification of the Southern Ocean, where the change in density from surface

waters to deep waters is small compared to lower latitudes. The mean current speeds of the ACC jets are relatively modest, typically less than 0.5 m s 1 (about 1 knot), with much weaker flow between the



181

0m

S

N

F-

F-

SA

SA

-N PF

PF

-S

AS F SB sA C sA Cf-S C C f-N

(b)

33.8

35.2

33.9

34.74

34.4 34.4

34

250

34.68

34.3

34.2

35

34.3 34.66

34.6 34.71

34.4

34.66

500

34.66

34.6

34.5

34.68 34.69

34.73 34.7

34.5

34.72

750

34.74

34.67

1000 65° S

60° S

55° S

0m 34.4

34.6

45° S

50° S

35.2

33.9

34

34.2

35 34.74

34.66

500

34.3

34.6 34.5

34.5

34.71 34.66 34.68

1000

34.74

AAIW

34.73

34.73

1500

34.68

34.4

34.75

LCDW 34.72

2000 2500

34.6

34.7

34.74

34.7 34.71

3000

34.75

34.74

34.68 34.69

34.7 34.73

3500 28.36

4500 5000 0 km

500

1000

1500

2000

Figure 2 Continued.

nutrient concentrations (close to zero for silicic acid and somewhat higher concentrations of nitrate and phosphate). The Antarctic Zone is characterized by fresh surface waters, shallow summer mixed layers, and high concentrations of major nutrients like nitrate and silicic acid, but low concentrations of micronutrients like iron. The zones delimited by the fronts of the ACC also define biogeographic zones populated by distinct species assemblages. For example, waters south of the Polar Front tend to be dominated by large phytoplankton such as diatoms (who need silicic acid) and large zooplankton, while coccolithophores

and small zooplankton dominate north of the Subantarctic Front. The distribution and foraging patterns of larger animals (e.g., fish, seabirds, and marine mammals) are also influenced by the frontal structure of the ACC. In some cases, the fronts themselves tend to be associated with higher primary productivity. The higher productivity near fronts can be caused by advection of micronutrients by the current or by upwelling caused by eddies or by topographic interactions. The currents of the ACC can also play a direct role in ecosystem dynamics, for example, by carrying krill and larvae from the Antarctic Peninsula to South Georgia.



183

SA F SA -S FN

PF -N

PF -S

AS F SB sA C sA Cf-S C C f-N

(d)

0m 220

280

250

270

260

240

UCDW

200

500

260

205 230

190

180 175

220

205

240

180

750

250 210

180

215

200

180

210

1000 65° S

55° S

60° S

50° S

45° S

0m

280

280 250

UC

500

180

DW

180

230 190

175

220 240

200

190

1500

210

180

1000

UC

DW

195

210

200

170

205

175

2000

200

205

2500

180

220 210

220 190

3000 215

3500

240

205

230

200

4000 210

4500 5000 0 km Figure 2 Continued.

240

500

1000

Circulation of the Antarctic Circumpolar Current A map of the elevation of the sea surface (dynamic height) shows how the position and intensity of the current varies along its circumpolar path (Figure 3). Contours of dynamic height are approximate streamlines for the flow, so the current flow is rapid in regions where the contours are closely spaced and weak in regions where the contours are widely separated. The steering of the fronts by large bathymetric features is clearly illustrated in Figure 3. Recent studies using high-resolution sampling from ships and satellites have revealed a more complex structure to the ACC than previously appreciated.

1500

2000

The ACC fronts consist of multiple branches, which merge and diverge in different regions and at different times along the circumpolar path of the current system (Figures 3 and 4). The multiple jets in the ACC reflect the tendency for geophysical flows on a sphere to self-organize into narrow, elongated, persistent zonal flows; similar circulation patterns form in the Earth’s atmosphere and on other planets. The position of the fronts varies with time, but generally over a relatively small latitude range at any given longitude (typically 711 of latitude). The variability is larger downstream of major bathymetric features and in regions where the ACC fronts interact with the strong boundary currents of the subtropical gyres to the north (e.g., south of Africa).


184


S

N

F-

F-

SA

SA

-N PF

PF

-S

AS F SB sA C C sA f-S C C f-N

(e)

0m 6

4

250

2.5 0.6

3

2 1

500

1.5 0.8

0.4

3.5

1

0.2 0.3

750 0.25

0.15

0.1

2

0.4

1000 65° S

60° S

55° S

45° S

50° S

0m 4 1.5

500

0.15

1000

0.25

1

3 2

0.2 0.3

0.2

0.6

AAIW

0.1

0.08

1500

3.5

SAMW

2.5

0.8

0.4

0.4 0.05

2000 1

2500 3000

0.05

0.1

CDW 0.08

0.6

0.01

0.15

3500

1.6 2

AABW 1

0.3

0.25 0.2

0.4

4000

>2 0.8

4500 5000 0 km

500

1000

1500

2000

Figure 2 Continued.

Transport of the Antarctic Circumpolar Current During the World Ocean Circulation Experiment (WOCE) in the 1990s, the transport through Drake Passage was estimated to be 13778 106 m3 s 1, similar to the estimates made in the late 1970s. Because the Atlantic basin is nearly closed to the north of the Southern Ocean, the net transport between Africa and Antarctica must be very close to the Drake Passage transport (to within c. 1 106 m3 s 1). The transport between Australia and Antarctica must be somewhat greater, to compensate for the flow from the Pacific to the Indian Oceans through the

Indonesian archipelago. Repeat transects during WOCE showed that the baroclinic transport south of Australia is 147710 106 m3 s 1, consistent with estimates that about 10–15 106 m3 s 1 flows through the Indonesian passages. The fronts, in particular the Subantarctic and Polar Fronts, carry most of the ACC transport. The relative contribution of these two fronts to the total transport varies around the circumpolar path. For example, south of Australia the Subantarctic Front carries 4 times more water to the east than the Polar Front, while in Drake Passage the transport carried by the two fronts is roughly equal in magnitude.



(f)

185

200

30

150

Transport (Sv)

20

100 SAF-N

sACCf-S SB

SAZ

PF-S

sACCf-N

10

50

0

0

Cumulative transport (Sv)

SAF-S

PF-N −10

−50

−20 64

62

60

58

56

54

52

50

48

46

44

Latitude (°S) Figure 2 Continued.

The transport of water masses by the ACC also changes with longitude. For example, the ACC carries an excess of intermediate density water into the Atlantic through Drake Passage, which is compensated by an excess of deep water leaving the basin south of Africa. These changes in water mass transports by the ACC reflect water mass transformations in the Atlantic basin, where relatively light Antarctic Intermediate Water is converted to denser North Atlantic Deep Water. The ACC also transports vast amounts of heat, fresh water, nutrients, carbon, and other properties between the ocean basins. The transport of the ACC varies over a range of timescales. Multiyear deployments of bottom pressure recorders in Drake Passage during the late 1970s and 1990s suggest a standard deviation in net transport of about 8–10 106 m3 s 1. For periods shorter than about 6 months, most of the variability is due to changes in sea level (i.e., changes in the barotropic, or depth-independent, flow). Models and sea level measurements suggest these barotropic motions are highly correlated with changes in wind stress and tend to follow bathymetric contours (more precisely, the flow is along lines of constant planetary vorticity, where planetary vorticity is given by the Coriolis parameter (equal to twice the rotation rate of the Earth multiplied by the sine of the latitude) divided by the ocean depth). For longer periods, variations in the density field (and hence the

baroclinic, or depth-varying, flow) also become important.

Dynamics of the Antarctic Circumpolar Current The absence of continental barriers in the latitude band of Drake Passage makes the dynamics of the ACC distinctly different in character from those of currents at other latitudes. Simple wind-driven ocean circulation theory (the Sverdrup balance), which generally does a good job of describing the circulation of the upper ocean in basins bounded by continents, cannot be applied in the usual way in a continuous ocean channel. The dynamical balance of the ACC has therefore been a topic of great interest for many years. The strong westerly winds over the Southern Ocean have long been recognized to help drive the ACC. The winds drive surface waters to the left of the wind (see Ekman Transport and Pumping), causing upwelling to the south of the wind stress maximum and downwelling to the north. This pattern of upwelling and downwelling helps to establish the tilt of density surfaces across the ACC and therefore its geostrophic flow. Despite the absence of continental barriers, the Sverdrup theory of wind-driven currents has been applied in the Southern Ocean by assuming


ANTARCTIC CIRCUMPOLAR CURRENT ∇ SSH , 129−130° E 1994

1994.5

1995

1995.5

1996

1996.5

1997 40

50

60

70

Latitude (° S)

−0.1

0

0.1

0.2

0.3

Figure 4 The north–south gradient of sea surface height (in m per 100 km) at 1301 E, south of Australia. Large height gradients indicate strong currents. The multiple bands of large gradient correspond to the jets of the ACC. Note that the jets merge and diverge, and change in intensity with time, and also persist for many months at the same latitude. Reproduced with permission from Sokolov S and Rintoul SR (2007) Multiple jets of the Antarctic Circumpolar Current. Journal of Physical Oceanography 37: 1394–1412. & Copyright [2007] AMS.

that relatively shallow bathymetric features act as ‘effective continents’. However, such calculations assume no interaction between the current and the bathymetry, which we know to be a poor

187

assumption for the deep-reaching ACC. In addition, wind is not the only factor driving the ACC. The atmosphere also drives ocean currents by exchanging heat and fresh water with the ocean, causing the density of seawater to change. Exchange of fresh water can result from precipitation, evaporation, and the freezing and melting of ice, both sea ice and glacial ice in the form of icebergs and ice shelves. Because the speed of ocean currents is proportional to the horizontal gradient of density, any process that produces horizontal density gradients will drive ocean currents. In the case of the ACC, both the strong westerly winds and the air– sea exchange of buoyancy play a part in driving the current. The momentum supplied to the Southern Ocean by the wind needs to be compensated in some way. The question of what balances the wind forcing has been a topic of debate for many decades. Recent studies have confirmed the early hypothesis by W. Munk and E. Palmeń that interaction of the ACC with seafloor topography provides a force to balance the wind. This force, known as the bottom form stress, results when the ocean currents are organized such that there is higher pressure on one side of a ridge on the seafloor than is found on the other side. In the case of the ACC, higher pressure is generally found on the west side of topographic ridges or hills, providing a force from the solid Earth to the ocean that balances the wind stress at the sea surface. While these pressure differences are too small to observe directly, realistic numerical simulations clearly show this force balance in action. Eddies produced by dynamical instabilities of the ACC fronts play a crucial role in establishing the momentum and heat balance of the Southern Ocean. The ACC has some of the most vigorous eddy activity observed in the ocean (Figure 5). Eddies are produced when dynamical processes release some of the energy stored in the sloping of density surfaces across the ACC, converting some of the energy in the mean flow into motions that vary with time, or eddies. This process is called baroclinic instability. The eddies transfer momentum vertically from the sea surface to the deep ocean, helping to set up the system of deep currents that interact with bathymetry to provide the bottom form stress. Both transient eddies (motions that vary with time) and standing eddies (deviations of the flow from the east–west average) contribute to the momentum transport. At the same time, the eddies carry heat poleward across the ACC, to compensate the heat lost to the cold atmosphere near Antarctica. Attempts to relate ACC transport variability to variations in wind have been inconclusive. It is now


188


1/6° HIM

Day 365 Africa

S Am outh eri ca

The Antarctic Circumpolar Current and the Overturning Circulation

Au st ra

lia

Antarctica

−5

−3

ACC starts to speed up, but this causes more vigorous eddy fluxes that dissipate the extra energy in the mean flow.

−2 −1.8 −1.6 −1.4 −1.2 −1 −0.9 −0.8 −0.7 −0.6 −0.5 −0.4 −0.3 −0.2 −0.1 0

0.2 0.4 0.6

log10 of magnitude of velocity averaged over top 100m in ms−1

Figure 5 A snapshot of surface speed from a high-resolution numerical simulation of the ACC. The filamented, eddy-rich structure of the ACC stands out clearly. Adapted with permission from Hallberg R and Gnanadesikan A (2006) The role of eddies in determining the structure and response of the Southern Ocean overturning: Results from the Modelling Eddies in the Southern Ocean project. Journal of Physical Oceanography 36: 2232–2252. & Copyright [2007] AMS.

understood that the dynamical balance of the ACC depends on a number of factors, including wind and buoyancy forcing, eddy–mean flow interaction, topographic form stress and the ocean stratification, so a simple relationship between transport and wind should not be expected. Recent improvements in ocean observing systems have allowed changes in the ACC to be assessed for the first time. Comparison of temperature profiles collected since the 1950s suggests that much of the Southern Ocean has warmed. The warming is largest in the ACC belt and is consistent with a southward shift of the current, allowing warm water to move south into areas previously occupied by cooler water. The southward movement of the ACC has been linked to a southward shift and strengthening of the westerly winds. The change in the winds, in turn, has been linked both to loss of ozone in the polar stratosphere and to enhanced greenhouse warming. The transport of the ACC has apparently not changed much over this time period, despite the change in wind forcing. Recent studies suggest the ACC may be in a regime in which the transport is insensitive to changes in wind: as the wind forcing increases, the

The eastward flow of the ACC is dynamically linked to a weaker circulation in the north–south plane. The distribution of water properties on transects across the Southern Ocean clearly reveals water masses spreading across the ACC. For example, the salinity maximum of the Lower CDW and oxygen minimum of the Upper CDW can be traced as they shoal from depths of 2000–3500 m north of the ACC to approach the sea surface south of the Polar Front (Figure 2). The high-oxygen, low-salinity waters formed in the Southern Ocean (Antarctic Intermediate Water and Antarctic Bottom Water) can be followed as they cross the ACC and enter the basins to the north. These distributions reflect an ocean circulation pattern known as the overturning circulation (Figure 6). Deep water spreads to the south across the ACC and upwells at the sea surface. Some of the upwelled deep water is driven north beneath the westerly winds, gains heat and fresh water from the atmosphere, and therefore becomes less dense, and ultimately sinks to form Antarctic Intermediate Water and Subantarctic Mode Water. Deep water that upwells further south and closer to Antarctica is converted to denser Antarctic Bottom Water and returns to the north. The result of these water mass transformations is a circulation in the north–south plane that consists of two counter-rotating cells. According to the residual mean theory, the strength of the net overturning circulation (mean flow plus the eddy contribution) is determined by the surface buoyancy forcing. The Southern Ocean overturning cells play an important part in the global-scale overturning circulation. The Southern Ocean imports deep water from the basins to the north, and exports bottom water and intermediate water. Recent studies suggest the conversion of deep water to intermediate water in the Southern Ocean is a key link in the global overturning circulation. For decades it has been assumed that the sinking of dense water in the polar regions was balanced by widespread upwelling at lower latitudes. However, measurements of mixing rates and large-scale tracer budgets suggest that mixing in the interior of the ocean is an order of magnitude too weak to support the upwelling required. The transformation of deep water to



Buoyancy loss

Buoyancy gain

SAF

PF

189

STF SAMW

Continental shelf

AAIW

Depth (m)

1000

Antarctica

UCDW

2000

NADW

AABW 3000

4000 80° S

Mid-ocean ridge

70°

LCDW

60°

50°

40°

30°

Latitude Figure 6 A schematic view of the Southern Ocean overturning circulation. Upper Circumpolar Deep Water (UCDW) upwells at high latitude and gains buoyancy (from heating, precipitation, and ice melt) as it is driven north, to ultimately sink as Antarctic Intermediate Water (AAIW) or Subantarctic Mode Water (SAMW). Lower Circumpolar Deep Water (LCDW) and North Atlantic Deep Water (NADW) upwell closer to the Antarctic continent, are made more dense (by cooling and salt rejected during sea ice formation), and sink to form Antarctic Bottom Water (AABW). The major southern ocean fronts are indicated. STF, Subtropical Front; SAF, Subantarctic Front; PF, Polar Front. Reproduced with permission from Speer K, Rintoul SR, and Sloyan B (2000) The diabatic Deacon cell. Journal of Physical Oceanography 30: 3212–3222. & Copyright [2007] AMS.

intermediate water by air–sea buoyancy exchange in the Southern Ocean provides an alternative means of connecting the upper and lower limbs of the global overturning circulation. Mixing likely also makes a contribution in regions of rough bathymetry, including the Southern Ocean, where elevated mixing rates have been measured. Eddies spawned by the ACC make an important contribution to the overturning circulation. The eddies transfer mass across the Drake Passage gap, where the absence of land barriers means that there can be no net east–west pressure gradient and therefore no net north–south flow. Furthermore, the same forces that drive the overturning circulation (wind and buoyancy exchange) also drive the ACC. The west-to-east flow of the ACC and the overturning circulations cannot be understood in isolation. The two are intimately linked, and eddy fluxes, topographic interactions, and wind and buoyancy forcing are all important ingredients of the dynamical coupling between them. The Southern Ocean overturning also has a large influence on global biogeochemical cycles. Upwelling of nutrient-rich deep water south of the ACC returns nutrients to the surface layer. The nutrients are exported from the Southern Ocean to lower latitudes by the overturning circulation, ultimately supporting a large fraction of global primary production. Water

masses at the surface of the Southern Ocean exchange oxygen and carbon dioxide with the atmosphere and carry ‘ventilated’ water to the interior of the ocean when the water masses sink. Where deep water upwells south of the ACC, carbon dioxide is released to the atmosphere; where water sinks from the sea surface, carbon dioxide is carried from the atmosphere into the ocean. As a result of the overturning circulation, more of the carbon released by human activities is accumulating just north of the ACC than in any other latitude band of the ocean.

Summary The ACC is the largest current in the world ocean, carrying about 13778 106 m3 s 1 from west to east around Antarctica. By connecting the ocean basins, the ACC allows water masses and climate anomalies to propagate between the basins. The current flow is concentrated in a number of circumpolar fronts, which extend from the sea surface to the seafloor. The fronts also mark the boundaries between zones with distinct physical, chemical, and ecological characteristics. Eddies produced by dynamical instabilities of the fronts play an important part in the dynamics of the ACC by transporting momentum vertically and heat and mass poleward.


190


Both wind and buoyancy forcing contribute to driving the ACC. Interaction between the deep-reaching flow and the bottom topography establish bottom form stresses to balance the wind forcing. The strong eastward flow of the ACC is intimately connected to an overturning circulation made up of two counterrotating cells. Water mass transformations driven by exchange of heat and moisture with the atmosphere connect the upper and lower limbs of the thermohaline circulation. The transport and storage of heat, fresh water, and carbon dioxide by the ACC have a significant influence on global and regional climate.

See also Bottom Water Formation. Carbon Cycle. Current Systems in the Southern Ocean. Ekman Transport and Pumping. Energetics of Ocean Mixing. Heat Transport and Climate. Mesoscale Eddies. Ocean Circulation. Ocean Subduction. Satellite Altimetry. Shelf Sea and Slope Sea Fronts. Ocean Circulation: Meridional Overturning Circulation. Water Types and Water Masses. Weddell Sea Circulation. Wind- and Buoyancy-Forced Upper Ocean. Wind Driven Circulation.

Further Reading Cunningham S, Alderson SG, King BA, and Brandon MA (2003) Transport and variability of the Antarctic Circumpolar Current in Drake Passage. Journal of Geophysical Research 108: 8084 (doi:10.1029/2001 JC001147). Deacon G (1984) The Antarctic Circumpolar Ocean. London: Cambridge University Press. Gille ST (2002) Warming of the Southern Ocean since the 1950s. Science 295: 1275--1277. Gordon AL and Molinelli E (1986) Southern Ocean Atlas. Washington, DC and New Delhi: National Science Foundation and Amerind Publishing. Hallberg R and Gnanadesikan A (2006) The role of eddies in determining the structure and response of the Southern Ocean overturning: Results from the

Modelling Eddies in the Southern Ocean project. Journal of Physical Oceanography 36: 2232--2252. Munk WH and Palmeń E (1951) Note on the dynamics of the Antarctic Circumpolar Current. Tellus 3: 53--55. Niiler PP, Maximenko NA, and McWilliams JC (2003) Dynamically balanced absolute sea level of the global ocean derived from near-surface velocity observations. Geophysical Research Letters 30: 2164. Nowlin WD Jr. and Klinck JM (1986) The physics of the Antarctic Circumpolar Current. Review of Geophysics and Space Physics 24: 469--491. Olbers D, Borowski D, Volker C, and Wolff J-O (2004) The dynamical balance, transpor‘t and circulation of the Antarctic Circumpolar Current. Antarctic Science 16: 439--470. Orsi AH and Whitworth T, III (2005) In: Sparrow M, Chapman P, and Gould J (eds.) Hydrographic Atlas of the World Ocean Circulation Experiment (WOCE), Vol. 1: Southern Ocean. Southampton: International WOCE Project Office (ISBN 0-904175-49-9). Orsi AH, Whitworth T, III, and Nowlin WD (1995) On the meridional extent and fronts of the Antarctic Circumpolar Current. Deep-Sea Research I 42: 641--673. Rintoul SR, Hughes C, and Olbers D (2001) The Antarctic Circumpolar Current system. In: Siedler G, Church J, and Gould J (eds.) Ocean Circulation and Climate, pp. 271--302. London: Academic Press. Sokolov S and Rintoul SR (2007) Multiple jets of the Antarctic Circumpolar Current. Journal of Physical Oceanography 37: 1394--1412. Speer K, Rintoul SR, and Sloyan B (2000) The diabatic Deacon cell. Journal of Physical Oceanography 30: 3212--3222. Whitworth T, III (1980) Zonation and geostrophic flow of the Antarctic Circumpolar Current at Drake Passage. Deep-Sea Research 27: 497--507. Whitworth T, III and Nowlin WD (1987) Water masses and currents of the Southern Ocean at the Greenwich Meridian. Journal of Geophysical Research 92: 6462--6476.

Relevant Website http://www.soc.soton.ac.uk – Electronic Atlas of WOCE Data.


ANTARCTIC FISHES I. Everson, Anglia Ruskin University, Cambridge, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction Antarctica is a continental landmass much of which is covered by an ice cap, consequently the ichthyofauna is totally marine. Surrounding the continent is the Southern Ocean, approximately 36 million km2, continuous with the Atlantic, Indian, and Pacific Ocean basins to the north and whose northern limit is generally taken as the Antarctic Polar Frontal Zone (APFZ). There is a clear separation between the Antarctic and the Southern Hemisphere continents, the nearest connection being with South America via the Scotia Arc, a series of islands separated from each other by deep water. The Antarctic Circumpolar Current (ACC) and the general oceanographic regime mean that marine isotherms are more or less concentric around the continent. Close to the continent the seasonal variation in temperature is rarely more than 1 1C while even at the northern limit, as for example at South Georgia, the range is little more than 4 1C. These two factors, geographical isolation and constant low temperature, have a major effect on Antarctic fish.

marine environment. Antarctic notothenioids with their high species diversity and endemism form a species flock comparable to that of Lake Baikal. Early taxonomic studies were based on the traditional methods of morphometric and meristic analyses. Recent studies have used molecular biological analyses not only of nuclear material but also of antifreeze compounds to indicate phylogeny.

Adaptations Cold Adaptation

Some of the earliest studies on the physiology of Antarctic fish concerned the measurement of oxygen uptake rates. Initially it had been assumed that, since many biochemical processes are temperaturedependent, the metabolic rates of Antarctic fish might be very low. The initial experiments indicated that rates were substantially higher than those of temperate fish when studied at low temperature and the degree of elevation of the metabolic rate in Antarctic fish was attributed to a phenomenon termed ‘cold adaptation’. Subsequent studies demonstrated that the greater part of this elevation was caused by handling stress and the extended recovery Table 1

Composition of Southern Ocean ichthyofauna

Taxon

Fish Fauna The Southern Ocean ichthyofauna is relatively sparse and unusual in composition, consisting of 213 species belonging to only 18 families (Table 1 and Figure 1). Nearly half the species belong to one group, the perciform notothenioids, which make up 45% of the fish fauna. Restricting consideration to the shelf, and particularly in the highest latitudes, notothenioids make up 77% of the species and 90–95% of the biomass of fish. Notothenioids are morphologically and ecologically diverse, and have variegated into a wide variety of niches, mainly demersal, and also in the water column and even within sea ice. As a group, this makes them more diverse than, for example, the finches of the Gala´pagos archipelago. The concept of species flocks has been developed for freshwater fish to identify groups that have a close affinity; typically such flocks are to be found in ancient lake systems and it is extremely unusual for such a flock to be identified from a large

Agnatha Chondrichthyes Osteichthyes Notacanthiformes Anguilliformes Salmoniformes Stomiiformes Aulopiformes Myctophiforms Gadiformes Ophidiiformes Lophiiformes Lampriformes Beryciformes Zeiformes Scorpaeniformes Perciformes Zoarcidei Notothenioidei Blennioidei Scombroidei Stromateoidei Pleuronectiformes


Benthic 2 8

Benthopelagic

2

Pelagic

1

2 2 4

9 1

11

5 12 9 35 1 1 3 2 6

1 32 22 95 1 2 1 4

191

192

ANTARCTIC FISHES

Anchor ice (to 30 m)

Annual ice Platelet ice Trematomus nicolai

Pagothenia borchgrevinki Pleuragramma antarcticum

Dissostichus mawsoni Trematomus bernacchii Trematomus loennbergii

Figure 1 These six species from McMurdo Sound demonstrate some of the life-history types included in the Nototheniidae. Pelagic, cryopelagic, epibenthic, and benthic species are illustrated. Dots indicate typical habitat, although most species have considerable depth ranges. Modified from Eastman and DeVries (1986).

time, of the order of 24 h or more, following introduction into respirometers. In spite of this, it is now accepted that some slight elevation of metabolic rate remains that cannot be explained wholly by experimental technique. Consideration of the phenomenon has raised some controversy between different workers. The existence of the phenomenon has been demonstrated experimentally, although it does not appear to confer any evolutionary advantage because it implies a higher energy requirement on the part of the fish. All these studies have been undertaken on whole fish; the overall oxygen uptake rate being the balance between all the component metabolic pathways that are present. As such it has been argued that the term ‘cold adaptation’ has little meaning and that it is more sensible to consider each metabolic component separately to provide an overall balance. Antifreeze

Pure water freezes at 0 1C, but the presence of salts causes the freezing point to be depressed such that normal seawater freezes at around 1.85 1C. At McMurdo Sound the annual mean water temperature is 1.87 1C and varies within the range 1.40 to 2.15 1C. Body fluids, such as the blood plasma, of most teleost fish have a freezing point of c. 0.7 1C. Even though this difference is small it is important, because living in waters close to the freezing point of seawater, Antarctic fish require some mechanism to prevent their body fluids from freezing. In the absence of ice, fish could live in a supercooled state. Unfortunately this is not a stable state because very few ice crystals are required to cause a supercooled liquid to freeze. An alternative adaptation is required. The ionic concentration of the blood of most marine teleosts is 320–380 mOsm kg 1, only about one-third of that of Antarctic seawater (1050 mOsm kg 1). The freezing

point depression of some notothenioids at McMurdo Sound is 2.2 1C, although their blood osmolality is 550–625 mOsm kg 1, equivalent to a freezing point depression of 1.02 to 1.16 1C. Thus although there appears to be some compensation as measured by the osmolality, it is insufficient to explain all of the depression in freezing point. Compensation for this difference comes in the form of antifreeze glycopeptides (AFGPs) which exert their effect by a mechanism known as adsorption-inhibition (Figures 2 and 3). Even though ice crystals can form, their further growth is prevented when AFGPs are adsorbed onto them because the AFGP molecule prevents growth of the ice crystal along its main axis. Thus the AFGPs have an antifreeze function, lowering the freezing point beyond that which would be expected from the osmolality. The AFGPs however do not lower the melting point. The AFGP molecules are of such a size that they would be lost through the glomeruli of normal teleost kidneys. In glomerular nephrons of normal teleosts, molecules with a molecular weight of o68 000 Da pass through the filtration barrier. As the urine passes through the different parts of the nephron, it is modified by reabsorption of nonwaste products and secretion of waste products. The AFGP molecules are of such a size that they would pass through the glomeruli but would need to be reabsorbed later on in the nephron. The kidneys of all Antarctic fish which possess AFGPs are aglomerular, obviating this requirement. Thus the evolution of the aglomerular trait in Antarctic fish complements that of the presence of antifreeze. Cardiovascular

A continuous low water temperature means that the oxygen-carrying potential of seawater is high. Thus, as long as the partial pressure of oxygen in the


ANTARCTIC FISHES

seawater remains high, so will the available oxygen. It is against this background that further cardiovascular adaptations have evolved. Early taxonomic studies relied on specimens preserved in alcohol or formalin, both of which affect the color of the fish. Because fish typically possess red H

O H

N

C C

CH3

H

C

N

N

O H C

C

C

CH 3 H CH2OH

O H

C

O

O

H

OH

CH3

H H

CH 2OH O

O

HO H OH H

H H H

H

NH C

O

CH 3

OH

Figure 2 Basic repeating structural unit of the AFGPs of notothenoids. The peptide consists of amino acids in the sequence [alanyl-alanyl-threonine]n. Each threonine is joined to a disaccharide through a glycosidic linkage. In low-molecularweight AFGPs 6–8, proline is periodically substituted for alanine at position 1 of the tripeptide. Reproduced from Eastman JT (1993) Antarctic Fish Biology: Evolution in a Unique Environment. London: Academic Press.

th grow Ice sence b in a GPs h owt e of AF r g Ice esenc r in p GPs F of A

193

blood, until the 1950s no mention was made of the anemic appearance of the gills of some species of Antarctic fish. At that time, it was noticed that members of the Channichthyidae (at that time called Chaenichthyidae) were white, as a result of which they were called ‘white-blooded fish’ or ‘icefish’. The blood of channichthyids is devoid of hemoglobin, although small numbers of nonfunctional erythrocytes have been described in a few species. Initial consideration was given to determine whether, because channichthyids do not possess scales, cutaneous respiration might be a major factor in oxygen uptake. However, the absorptive area and vascularization relative to the gills mitigated against that mechanism. Alternatively it was thought possible that channichthyids possessed either a more efficient oxygen utilization mechanism or else lowered oxygen requirement. This second consideration was being examined at a time when the concept of metabolic cold adaptation was under discussion. The viscosities of the plasma of red-blooded notothenioids and channichthyid fish are very close, although the blood of the former is approximately 25% higher than the latter. Studies on oxygen uptake rates indicated that channichthyids utilized oxygen at a slightly lower rate as compared with equivalent red-blooded notothenioids. In the absence of hemoglobin, the oxygen-carrying capacity of channichthyid blood is only about one-tenth that of

c-axis

a-axis a-axis Step

Long straight growth fronts

Basal plane

Adsorbed AFGP molecules

Prism face

Small, highly curved growth fronts

Figure 3 Model of an ice crystal depicting adsorption-inhibition as a mechanism for the freezing point depression of water by antifreezes. In the absence of AFGPs, ice crystal growth occurs as water molecules are added to the crystal in a regular fashion at steps on the basal planes. When the AFGPs are adsorbed, ice cannot propagate over them and long straight fronts become divided into many small curved fronts. Reproduced from Eastman JT (1993) Antarctic Fish Biology: Evolution in a Unique Environment. London: Academic Press.


194

ANTARCTIC FISHES

red-blooded fish. Two mechanisms are possible to compensate for this effect: either channichthyid blood is circulated at a much faster rate or there is much more of it in the system. The latter has proven to be the case and channichthyid blood takes up 8–9% of the total volume of the fish (2–4 times that of other teleosts); the heart rate and blood pressure are low but the stroke volume and resultant cardiac output are large. To reduce the resistance to flow, the capillaries are larger than in other teleosts and the blood is less viscous. Even though the hemoglobin-less condition is clearly effective, it is a feature that confines the fish to areas of high oxygen tension such as those present in Antarctic waters. Only one channichthyid species, Champsocephalus esox, is found outside of the Antarctic zone. Experimental studies have demonstrated that channichthyids are particularly sensitive to hypoxia, indicating that in their natural habitat the oxygen saturation is always consistently high.

Weddell Sea Circulation. Wind- and BuoyancyForced Upper Ocean.

Further Reading Cheng C-HC, Cziko PA, and Evans CW (2006) Nonhepatic origin of notothenioid antifreeze reveals pancreatic synthesis as common mechanism in polar fish freezing avoidance. Proceedings of the National Academy of Sciences of the United States of America 103: 10491--10496. Clarke A (1991) What is cold adaptation and how should we measure it? American Zoologist 31: 81--92. Di Prisco G, Pisano E, and Clarke A (eds.) (1998) Fishes of Antarctica; A Biological Overview. Milan: Springer. Eastman JT (1993) Antarctic Fish Biology: Evolution in a Unique Environment. London: Academic Press. Gon O and Heemstra PC (eds.) (1990) Fishes of the Southern Ocean. Grahamstown: JLB Smith Institute of Ichthyology. Kock K-H (1992) Antarctic Fish and Fisheries. Cambridge, MA: Cambridge University Press.

See also Antarctic Circumpolar Current. Current Systems in the Southern Ocean. Fish Ecophysiology. Fish Feeding and Foraging. Fish Reproduction.


ANTHROPOGENIC TRACE ELEMENTS IN THE OCEAN E. A. Boyle, Massachusetts Institute of Technology, Cambridge, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 162–169, & 2001, Elsevier Ltd.

Introduction Human activities have increased the fluxes of several chemical elements into the ocean above natural levels. Despite convincing evidence for this enhancement of elemental fluxes (see Further Reading section for references relevant to the discussion in this article), there is only one element – lead (Pb) – where abundant evidence proves that open-ocean seawater concentrations are substantially higher than they were in preindustrial times. For a few other elements – e.g. cadmium (Cd) and mercury (Hg) – there is some evidence suggesting a detectable anthropogenic impact (or models indicating that an anthropogenic enhancement must exist even if it has not been observed). For most other elements, the size of the oceanic reservoir of these elements overwhelms relatively large anthropogenic fluxes, and it may require centuries of further inputs before the human impact can be discerned.

Anthropogenic Lead in the Ocean Sampling and analysis for Pb have been difficult because of low concentrations and abundant contamination sources: lead paint, lead weights, and gasoline exhausts. The latter source is now substantially reduced, so perhaps the current Pb contamination problem is less serious than it was in the 1980s. Patterson and co-workers were the first to call attention to the overwhelming anthropogenic augmentation of Pb fluxes into the environment, particularly that resulting from alkyl leaded gas utilization. Patterson’s evidence began with the demonstration that Pb deposition in remote Greenland snows had increased by two orders of magnitude. In the late 1970s/early 1980s, his laboratory obtained the first valid data (uncontaminated and properly analyzed) for the vertical distribution of Pb in water sample profiles from the North Atlantic, North Pacific, and South Tropical Pacific (Figure 1). These data demonstrated that the highest concentrations of

Pb occurred in the surface ocean and that concentrations decreased with increasing depth in the water column. At that time, the highest Pb concentrations were found in the North Atlantic Ocean (160 10 12 moles kg 1 at the surface decreasing to 26 10 12 moles kg 1 at 3000 m water depth). The high Pb concentrations in this basin are emitted from the major industrial nations surrounding the basin. Lower Pb concentrations were seen in North Pacific surface waters (60 10 12 moles kg 1 at the surface decreasing to 5 10 12 moles kg 1 at 3500 m depth), and the lowest concentrations in the south tropical Pacific (20 10 12 moles kg 1 at the surface decreasing to 4 10 12 moles kg 1 at 4000 m). In Patterson’s view, this evidence proved the anthropogenic origin of Pb in the modern ocean. The lead industry attempted to discredit Patterson’s evidence on environmental Pb pollution by many spurious arguments. Although they were not clever enough, they might have attempted to discredit Patterson’s interpretation of his oceanic Pb data by pointing out that similar comparative concentration variations occur for aluminum (Al) in the ocean, even though oceanic Al is entirely of natural origin. Al is released from terrestrial dusts blown into the ocean and ‘scavenged’ onto sinking biologically produced particles that remove it from the deep ocean. Therefore Al is high in the surface ocean, highest downwind of major dust sources such as north-west Africa, and lowest in the South Pacific because of low dust inputs to the surface and cumulative scavenging in the deep waters. Lead was phased out of gasoline in the USA (followed soon by Canada and Japan, and somewhat later by western Europe and a few other countries) beginning in 1970 when the US Environmental Protection Agency mandated emissions controls on gasoline exhausts. Originally, controls on lead emissions were not the goal of the regulations; the regulations were formulated to minimize emissions of hydrocarbons, nitrogen oxides, and carbon monoxide. As it turned out, the technological fix for those problems was to use catalytic converters on the exhaust stream, and the activity of the catalysts was destroyed by lead exhausts. Hence leaded gasoline could not be used with catalytic converters, and regulations mandated the elimination of leaded gasoline. Later, regulations specifically directed at minimizing lead emissions into the environment were also enacted.


195

196

ANTHROPOGENIC TRACE ELEMENTS IN THE OCEAN _12

Pb (10 50

0

_

moles kg 1)

100

150

200

0 South Pacific

North Pacific North Atlantic

1000

Depth (m)

2000

3000

4000

5000 Figure 1 Oceanic Pb profiles obtained by Patterson and co-workers on samples collected between 1976 and 1981.

Changing leaded gasoline utilization patterns offer an opportunity to establish that Pb decreases in the ocean in response to the decreasing Pb emissions from leaded gasoline. Pb gas utilization increased from its introduction in the 1920s until the 1970s, then decreased rapidly – resulting in nearly complete elimination from the USA by 1990 and from western Europe by the turn of the century (Figure 2). Surface waters of the ocean should respond within a few years to changes in fluxes from the atmosphere, and the upper layers of the ocean should respond on decadal timescales. These expectations are based upon the penetration of the nuclear bomb tritium into the ocean and natural radioisotope 210Pb (halflife 22.3 years). Studies of tritium penetration show that surface waters ventilate the upper thermocline of the ocean on a timescale of several years to several decades. 210Pb is supplied to the ocean from windblown aerosols which have acquired 210Pb from the decay of the radioactive noble gas 222Rn. 222Rn is released into the atmosphere as a decay product of

naturally occurring crystal 238U which decays in several stages to the immediate parent of 222Rn, 226 Ra. The flux of 210Pb out of the atmosphere has been measured at numerous sites, therefore the flux of 210Pb into the ocean can be estimated reasonably well. In the surface ocean, 210Pb is rapidly scavenged from sea water by newly formed biological material whose residues eventually sink out of the surface waters carrying 210Pb. The steady-state concentration of 210Pb in the mixed layer of the ocean is thus determined by a balance between 210Pb supply from the atmosphere and 210Pb removal by sinking particles. The time constant for this process is estimated by dividing the measured surface water 210Pb reservoir by the flux of 210Pb from the atmosphere. The rate of removal varies with the rate of biological activity, but in typical open-ocean ‘deserts’ such as the Sargasso Sea, 210Pb is removed from surface waters every 2 years. Thus as the flux of anthropogenic lead from the atmosphere decreases, the concentration of lead in surface waters should follow this decrease with a 2 year lag.


ANTHROPOGENIC TRACE ELEMENTS IN THE OCEAN

197

US and European gasoline Pb consumption, 1930 _93 300

200 USA

3

Gasoline Pb consumption (10 metric tonnes per year)

250

150

100

50 Italy UK France Germany

0 1930

1940

1950

1960

1970

1980

1990

Year Figure 2 Alkyl Pb gas consumption, 1930–93. Four European countries (accounting for 70% of EC gasoline consumption) stacked to compare with the much higher US consumption.

Decreasing concentrations of Pb in surface waters of the Atlantic Ocean have been documented by observations from 1979 until the present (Figure 3). The decreasing concentration of Pb in the upper layers of the ocean is also observed (Figure 4), with a slower response in the deeper waters that are replaced on decadal timescales. These decreases are seen in other parts of the Atlantic and North Pacific as well. This evidence amply demonstrates that the phasing out of leaded gasoline in the USA has been closely followed by decreases of the Pb concentration of the ocean on the appropriate timescale. In the period before direct observations of Pb in the ocean, the response of oceanic Pb to earlier changes in emissions are documented by the Pb content of reef-building corals. Corals precipitate calcium carbonate skeletons with annual variations in their density that can be counted in the same way

as tree rings, hence providing a chronology of the time of deposition. In corals from near Bermuda, concentrations of Pb increased from very low levels in the 1880s to higher levels in the 1920s, in pace with the emissions of Pb from high temperature industrial activities (smelting, coal combustion, etc.) (Figure 5). With the introduction of leaded gasoline, Pb increased far more until the mid-1970s, when coralline Pb began to decrease with the phasing out of leaded gasoline, confirming the picture provided by direct observations in this period. Other information on the sources and dispersion of anthropogenic lead can be derived from the stable isotope ratios of lead (e.g. 206Pb/207Pb) which vary from one mining source to another because of differing Pb/U and Pb/Th for geological sources of lead. In particular, US leaded gasoline for many years had a 206Pb/207Pb ratio of 41.20, whereas European


198


_

Pb (pmol kg 1)

150

100

50

0 1980

1985

1990 1995 Year (year marked at Jan. 1)

2000

Figure 3 Pb in Bermuda surface waters, 1979–99. Squares (near surface waters) and triangles (40–50 m depth) are samples collected and analyzed by the MIT group (see Wu and Boyle 1997 for data through 1996; 1997–2000 from Boyle et al., unpublished); circular symbols are samples collected and analyzed by Schaule and Patterson 1981 and Veron et al 1993.

leaded gasoline was o1.15. Eventually anthropogenic Pb will be removed from the water column and reside in marine sediments, and there is already evidence from surficial sediments of the North Atlantic Ocean for significant quantities of anthropogenic Pb.

Anthropogenic Mercury in the Ocean Volatile mercury (Hg) and organomercury compounds are emitted from the land into the atmosphere naturally from wildfires, volcanoes, and microbial activity. Mercury is emitted into the atmosphere by humans as a result of high temperature processes (e.g. smelting, coal combustion, incineration) combined with commercial uses of elemental mercury (e.g. thermometers, batteries), as well as disposal of mercury-laden wastes (e.g. from gold mining operations) that are then converted into volatile forms in the environment. Estimates indicate that the anthropogenic mercury emissions from the land into the atmosphere exceeded natural sources by approximately a factor of three during the past century, not nearly so large as the Pb emission enhancement, but nonetheless substantial. A large fraction of this Hg is carried over great

distances by the atmosphere and deposited into the ocean. As for Pb, contamination is a major problem in Hg sampling and analysis (perhaps even more so), because Hg0 is volatile and broken mercury thermometer residues exist in most chemical laboratories. Evidence for anthropogenic perturbation of the oceanic Hg reservoir is less straightforward than it is for Pb. Observations in the modern Atlantic marine atmosphere show that the Northern Hemisphere has two- to three-fold higher Hg concentrations than the Southern Hemisphere, as might be expected for an anthropogenic source. However, land area is also higher in the Northern Hemisphere, and it is likely that even natural emissions are higher in the Northern Hemisphere, so observation of an interhemisphere difference does not by itself prove anthropogenic perturbation. Measurements of Hg in sea water have attempted to determine elemental (Hg0), organometallic (e.g. CH3Hg, (CH3)2Hg), and ‘reactive’ forms (ionic and inorganically complexed Hg, organically complexed Hg, and labile organic and particulate Hg). Reported Hg concentrations range from 10 11 to 10 9 moles kg 1. There are no direct seawater measurements or indirect seawater



199

_

Pb (pmol kg 1) 50

0

100

200

150

0 1997 100

1993

200 1989 300 1984

Depth (m)

400

500

600 1979 700

800

900

1000 Figure 4 Pb in ocean profiles in the Sargasso Sea near Bermuda: 1979 data from Schaule and Patterson 1983; 1984 data from Boyle et al. (1986); data from Veron et al 1993; 1993 data from Wu and Boyle 1997; 1997 data (Wu and Boyle, unpublished).

80 Bermuda D. strigosa

Pb/Ca, nmol mol

_1

60

Southern Preserve 40 North Rock

20

0 1880

1900

1920

1940

1960

1980

Figure 5 Pb in corals near Bermuda in annually laminated layers deposited between 1880 and 1997 (Shen and Boyle, 1987).


200


indicators such as corals which have yet documented an anthropogenic enhancement, although it has been estimated that Hg concentrations in the surface ocean have increased from 0.5 to 1.5 pmol kg 1 during the past century.

Other Anthropogenic Elements in the Ocean Several other elements have had significant anthropogenic enhancements of global fluxes, but in most cases, the size of the oceanic reservoir of the elements is too large for an observable shift in open-ocean concentrations. A two- to three-fold increase in the concentration of cadmium (Cd) has been observed in corals near Bermuda during the past century. However, the concentrations of Cd in surface waters near Bermuda are low even today (certainlyo10 11 moles kg 1 and sometimes as low as 10 12 moles kg 1; because concentrations in deep waters are two or three orders of magnitude higher, the oceanic reservoir is large compared with the anthropogenic enhancement of Cd in surface waters. Although concentrations of Zn in corals were not measured, a similar anthropogenic enhancement of emissions to the atmosphere may well have caused slight enhancements to the Zn concentrations of the surface ocean, but evidence of such an increase has not been reported. Because of the use of tributyl tin as a stabilizer in PVC plastic (hence an incineration source for inorganic Sn) and as an antifouling agent in marine paints, there has been some interest in the fate of anthropogenic Sn in the oceanic environment. Inorganic Sn is very low in the surface waters of the Sargasso Sea (B3 10 12 moles kg 1). Tributyl tin has been observed in enclosed harbors, but not in the open ocean. On the whole, however, the size of the ocean precludes major enhancements of oceanic trace metal concentrations. Even where the enhancement is clear, such as for Pb, the concentrations are quite low and there is no reason to suppose that these enhancements pose a threat to marine biota.

Anthropogenic Trace Metal Enhancements in Coastal Waters and Embayments As opposed to minimal evidence for anthropogenic trace metal contamination of open-ocean waters, it is clear that major impacts on the trace metal chemistry of coastal sea water have occurred and that these have sometimes had serious consequences. The discharge of untreated industrial and human wastes into coastal waters, and the release of toxic metals and

metal compounds from antifouling paints have clearly increased metal concentrations in shallow ocean waters in many coastal regions. One of the most notorious examples of coastal pollution was in Minimata Bay, Japan where from the 1930s until the 1960s a chemical company dumped tonnes of mercury, which contaminated the bay sediments. This Hg pollution entered the marine food chain. Thousands of people developed severe neurological symptoms as a result of eating Hgcontaminated fish. This is perhaps the most extreme example where pollution in the marine environment has demonstrably harmed humans directly. More recently, it has been demonstrated that high Cd, Zn, Cu, and Pb concentrations in Atlantic coastal waters of southern Spain have risen by many orders of magnitude due to acid mine drainage from the Rio Tinto massive sulfides. As yet, little is known about any human health effects from this trace element pollution, although the region has an active coastal fishing industry. There is too much detail in the many local examples to discuss here, but there is little doubt that many coastal embayments proximate to big cities have elevated metal levels, and that an understanding of the consequences of this pollution is still being developed. Unfortunately, proper measurement capabilities did not exist until the past two decades, so there is usually no direct evidence for the degree of enhancement of dissolved metals relative to preindustrial times. One exception to this situation is in San Francisco Bay, where the anthropogenic enhancement of dissolved cadmium has been documented by measuring the Cd content of benthic foraminifera. An anthropogenic impact on metals in coastal marine sediments has been demonstrated repeatedly. Lead has been shown to be enriched in coastal sediments. Even more exotic elements such as silver (photography) and osmium (electron microscopy) are enriched in the sewage of some cities and can be traced from their source.

Conclusions The anthropogenic dominance of lead in the openocean environment has been conclusively demonstrated. Although there is evidence that the fluxes of other elements to the ocean have been enhanced to a significant degree, there is less direct evidence documenting their enhancements in the open-ocean environment. Coastal waters and sediments near major cities show large enhancements in several trace metals.



See also Metal Pollution. Radioactive Wastes. Refractory Metals. Transition Metals and Heavy Metal Speciation.

Further Reading Bruland KW and Franks RP (1983) Mn, Ni, Cu, Zn, and Cd in the western North Atlantic. In: Wong EBCS, Bruland KW, Burton JD, and Goldberg ED (eds.) Trace Metals in Seawater, pp. 395--414. New York: Plenum. Byrd JT and Andreae MO (1982) Tin and methyltin species in seawater; concentrations and fluxes. Science 218: 565--569. Chow TJ, Bruland KW, Bertine K et al. (1973) Lead pollution: records in Southern California coastal sediments. Science 181: 551–552. Fitzgerald WF, Engstrom DR, Mason RP, and Nater EA (1997) The case for atmospheric mercury contamination in remote areas. Environmental Science and Technology 32: 1--7. Flegal AR and Patterson CC (1983) Vertical concentration profiles of lead in the Central Pacific at 15N and 20S. Earth and Planetary Science Letters 64: 19--32. Hamelin B, Grousset F, and Sholkovitz ER (1990) Pb isotopes in surficial pelagic sediments from the North Atlantic. Geochimica et Cosmochimica Acta 54: 37--47. Hamelin B, Ferrand JL, Alleman L, and Nicolas E (1997) Isotopic evidence of pollutant lead transport from North America to the subtropical North Atlantic gyre. Geochimica et Cosmochimica Acta 61: 4423. Harrison RM and Laxen DPH (1981) Lead Pollution Causes and Control. London: Chapman and Hall. Helmers E and van bulleted Loeff MMR (1993) Lead and aluminum in Atlantic surface waters (501N to 501S) reflecting anthropogenic and natural sources in the eolian transport. Journal of Geophysical Research 98: 20261--20273. Mason RP, Fitzgerald WF, and Morel FMM (1994) The biogeochemical cycling of elemental mercury: anthropogenic influences. Geochimica et Cosmochimica Acta 58: 3191--3198. Jenkins WJ (1980) Tritium and He-3 in the Sargasso Sea. Journal of Marine Research 38: 533--569. Measures CI and Edmond JM (1988) Aluminum as a tracer of the deep outflow from the Mediterranean. Journal of Geophysical Research 93: 591--595. Measures CI and Edmond JM (1990) Aluminium in the south Atlantic: steady state distribution of a short residence time element. Journal of Geophysical Research 95: 5331--5340. Measures CI, Edmond JM, and Jickells T (1986) Aluminum in the northwest Atlantic. Geochimica et Cosmochimica Acta 50: 1423--1429.

201

Orians KJ and Bruland KW (1985) Dissolved aluminum in the central North Pacific. Nature 316: 427--429. Orians KJ and Bruland KW (1986) The biogeochemistry of aluminum in the Pacific Ocean. Earth and Planetary Science Letters 78: 397--410. Murozumi M, Chow TJ, and Patterson C (1969) Chemical concentrations of pollutant lead aerosols, terrestrial dusts and sea salts in Greenland and Antarctic snow strata. Geochimica et Cosmochimica Acta 33: 1247--1294. Nriagu JO (1989) A global assessment of natural sources of atmospheric trace metals. Nature 338: 47--49. Ravizza GE and Bothner MH (1996) Osmium isotopes and silver as tracers of anthropogenic metals in sediments from Massachusetts and Cape Cod bays. Geochimica et Cosmochimica Acta 60: 2753--2763. Schaule B and Patterson CC (1983) Perturbations of the natural lead depth profile in the Sargasso Sea by industrial lead. In: Wong CS, Boyle EA, Bruland KW, Burton JD, and Goldberg ED (eds.) Trace Metals in Seawater, pp. 487--504. New York: Plenum. Schaule BK and Patterson CC (1981) Lead concentrations in the northeast Pacific: evidence for global anthropogenic perturbations. Earth and Planetary Science Letters 54: 97--116. Shen GT and Boyle EA (1987) Lead in corals: reconstruction of historical industrial fluxes to the surface ocean. Earth and Planetary Science Letters 82: 289--304. Shen GT and Boyle EA (1988) Thermocline ventilation of anthropogenic lead in the western North Atlantic. Journal of Geophysical Research 93: 15715--15732. Shen GT, Boyle EA, and Lea DW (1987) Cadmium in corals as a tracer of historical upwelling and industrial fallout. Nature 328: 794--796. Sherrell RM, Boyle EA, and Hamelin B (1992) Isotopic equilibration between dissolved and suspended particulate lead in the Atlantic Ocean: evidence from Pb-210 and stable Pb isotopes. Journal of Geophysical Research 97: 11257--11268. Slemr F and Langer E (1992) Increase in global atmospheric concentrations of mercury inferred from measurements over the Atlantic Ocean. Nature 355: 434--437. Turekian KK, Benninger LK, and Dion EP (1983) Be-7 and Pb-210 total deposition fluxes at New Haven, Connecticut and at Bermuda. Journal of Geophysical Research 88: 5411--5415. van Geen A and Luoma SN (1999) A record of estuarine water contamination from the Cd content of foraminiferal tests in San Francisco Bay, California. Marine Chemistry 64: 57. van Geen A, Adkins JF, Boyle EA, Nelson CH, and Palanques A (1997) A 120 year record of metal contamination on an unprecedented scale from mining of the Iberian Pyrite Belt. Geology 25: 291--294. Veron AJ, Church TM, Flegal AR, Patterson CC, and Erel Y (1993) Response of lead cycling in the surface Sargasso Sea to changes in tropospheric input. Journal of Geophysical Research 98: 18269--18276.


202


Veron AJ, Church TM, Rivera-Duarte I, and Flegal AR (1999) Stable lead isotope ratios trace thermohaline circulation in the subarctic North Atlantic. Deep-Sea Research II 46: 919--935. Wolff EW and Peel DA (1985) The record of global pollution in polar snow and ice. Nature 313: 535--540.

Wu JF and Boyle EA (1997) Lead in the western North Atlantic Ocean: completed response to leaded gasoline phaseout. Geochimica et Cosmochimica Acta 61: 3279--3283.


ANTIFOULING MATERIALS D. J. Howell and S. M. Evans, Newcastle University, Newcastle, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction Any unprotected surface which is introduced into the sea, such as a buoy, a ship’s hull, an oil rig support, or a fish cage, will become fouled by the growth of marine organisms. Colonization of a newly immersed surface involves a typical ecological succession of different flora and fauna. Initially, the surface undergoes biochemical changes as macromolecules, such as glycoproteins, proteoglycans, and polysaccharides, become adsorbed on to it. Bacteria then colonize it within c. 1 h of submersion, followed by diatoms, yeasts, and protozoa. Subsequently, invertebrate larvae and algal spores settle on the surface and metamorphose and/or grow, often forming a dense covering over it. In the case of wooden structures, there are even some organisms, such as the shipworm (actually a mollusk, Teredo), which bore beneath the surface. The climax community can be highly diverse. Overall, there are 44000 known species of fouling organisms, including barnacles, hydroids, tube worms, bryozoans, seaweeds, and others. This article considers the problem of fouling on ships’ hulls, and measures taken to prevent it. Coatings containing the biocide tributyltin, more commonly known as TBT, were particularly effective antifoulants but TBT leached from them into the marine environment, harming nontarget species. Their use was initially regulated, then eventually banned altogether. The extent to which the paint industry has been able to meet the challenge of finding equally effective, but less environmentally damaging, alternative antifoulants is discussed.

Historical Development Fouling has caused problems to shipping throughout history. Teredo have been responsible for the destruction of many wooden sailing ships, even causing them to break up at sea. Modern steel hulls are not penetrated by these mollusks; nevertheless, their surfaces are highly vulnerable to fouling. Coverings by the growth of organisms can be remarkable if they are left unchecked. The total weight of fouled organisms can be as high as 150 kg m 2, representing

6000 t of fouling attached to a large commercial vessel which has an underwater surface area of 40 000 m2. However, it is not solely the increased weight which causes the problem to shipping. Fouling leads to an increased friction between the hull and seawater, causing the so-called ‘hull roughness’. As a result of these combined effects, badly fouled ships suffer from loss of speed and maneuverability. They are also expensive to operate, incurring fuel penalties of Z50%. Not surprisingly, there have been extensive efforts to reduce fouling, and its impacts, down the ages. The ancient Carthaginians and Phoenicians used pitch, and possibly copper sheathing, on the bottoms of ships to prevent fouling, and later coated hulls with sulfur and arsenical compounds. The Greeks and Romans both introduced lead sheathing. Copper cladding was then ‘reinvented’ in the seventeenth and eighteenth centuries, but it became redundant after the introduction of steel vessels at the end of the eighteenth century, because of problems caused by galvanic corrosion. A new technology was needed to protect ships’ hulls from fouling, but it took some 60 or so years until the mid-nineteenth century for the development of antifouling paints to provide it. There have been substantial developments in paint technology since the first antifouling paints were introduced, and the search for more effective coatings still continues today. However, the large majority of antifouling paints incorporate biocides into the matrix and, in this respect, the technology has not changed. The biocides leach slowly from the paints, killing organisms which attempt to settle on the hull. Cuprous oxide was used in early formulations, but numerous other toxins – including organomercury, lead and arsenical compounds, and DDT – were added as boosters, in order to enhance the effectiveness and life expectancy of the paints. However, such compounds pose severe environmental and human health risks, and they were withdrawn voluntarily by the paint industry during the 1960s. They were replaced largely by the highly effective biocide tributyltin (TBT). The properties of the paint matrix are another important consideration because they determine qualities such as the rate of biocide release and therefore the long-term effectiveness of a paint as an antifoulant. Originally, active biocidal ingredients were dispersed in a soluble resinous matrix, which released the biocide as it dissolved slowly in seawater. A problem with these free association paints,


203

204

ANTIFOULING MATERIALS

as they became known, was that the release rates of biocides from them were uncontrolled. The initial rate was rapid so that the paint was highly effective when it was first applied (Figure 1). However, subsequent release of the biocide from the matrix declined steadily and antifouling performance diminished with time; it might be virtually exhausted within c. 12 months. Consequently, free association paints typically offered a maximum period of only 1–2 years of service before dry-docking and repainting was necessary. There were limited improvements when matrices which were insoluble, using chlorinated rubber and vinyl resins, were developed in the 1940s. They had greater mechanical strength and allowed thicker coatings of paint to be applied, thereby providing effective antifouling performance over a full 2-year period. However, a major breakthrough in antifouling paint technology occurred with the introduction of so-called self-polishing copolymer paints in the late 1960s. The term self-polishing was used because the biocide was released slowly as the paint surface was gradually worn away. The essential difference from free association paints was that the biocide was chemically bonded in a copolymer resin system, via an organotin–ester linkage. This ester group hydrolyzed at the surface of the paint where it was in contact with seawater, and this resulted in a slow and controlled release of the biocide. The remaining surface of the paint was mechanically weakened by breakage of these bonds, and was eroded by moving seawater, resulting in the exposure of a fresh surface layer. The hydrolysis/erosion process was then repeated until there was no paint left.

The technology has been particularly successful with TBT copolymer (poly)tributyltinmethacrylatemethylmethacrylate, although cuprous oxide or other boosters were included in the formulations. They were needed even though TBT is active against most fouling organisms, since some slime-forming diatoms are resistant to it. TBT-based self-polishing copolymer antifouling systems (TBT SPC systems) brought enormous benefits to the shipping industry. They could provide effective antifouling cover for Z5 years, more than doubling the performance of free association paints. Reduced fuel costs and less frequent need to drydock and repaint vessels were estimated to be worth some US$5.7 billion per annum to the industry during the mid-1990s. There were also huge environmental benefits because lower fuel consumption by the world’s shipping fleet reduced the release of ‘greenhouse’ gases and emissions which were responsible for acid rain. Annual fuel savings, which could be attributed to the use of TBT-based antifoulants, were believed to be 7.2 million tons, reducing carbon dioxide emissions by 22 million tons, and sulfur dioxide by 0.6 million tons, per year. An additional benefit was that effective antifouling prevented the transport of invasive (nonnative) organisms across the world on ships’ hulls. Such species can cause enormous economic and ecological damage once they become established. It was originally thought that these organisms hitchhiked their way across the oceans in ballast water, but it is becoming increasingly clear that fouled hulls are an additional means of transport. The effect on local economies can be devastating. It has been estimated that the

Free association paints

Copolymer paints

Leaching

Polishing

Leaching rate

Leaching rate

Time

Time

Figure 1 A diagrammatic comparison of the mode of action of free association paints and self-polishing copolymer paints, and the rates at which they release tributyltin into the water. Reproduced from Stebbing ARD (1985) Organotins and water quality: Some lessons to be learned. Marine Pollution Bulletin 16: 383–390.



cost of control measures that were required to limit the destruction caused by the introduced European zebra mussel Dreissena polymorpha to the USA, which had infested over 40% of internal waterways, was US$1 billion between 1989 and 2000. TBT SPC systems came into widespread use. They were, for example, first used by major shipping lines in Europe and the Far East in the mid-1970s, and were registered by the Environment Protection Agency in the USA in 1978. Subsequently, it was estimated that they were used on 70% of the world’s commercial shipping fleet, and on high proportions of fishing vessels and pleasure craft.

The Downside of TBT-Based Antifoulants: Their Regulation and Eventual Ban Unfortunately, there are environmental costs of using TBT-based paints. The ideal antifouling biocide would be one which degraded into harmless residues immediately after release into the water column so that the toxic effects occurred at the ship’s hull but nowhere else in the marine environment. In fact, TBT does degrade reasonably rapidly in seawater, where it has a residence time of only a few days. Nevertheless, this is sufficient time for it to become adsorbed on to particles, and to aggregate in sediments in which there are high levels of organic matter. Here degradation processes are considerably slower, and the half-life of TBT may then be a matter of months or years. Regrettably, TBT is so toxic that even low concentrations can harm marine life. For example, the lethal dose concentration (15-day LC50) for larvae of the mussel Mytilus edulis is 10 ng l 1, and that (96-h LC50) for larvae of the sole Solea solea is 21 ng l 1. Sublethal effects occur at even lower concentrations. A dose of 40.4 ng l 1 can affect the growth and reproduction of phytoplankton and zooplankton, and one of 42 ng l 1 can affect the process of shell formation in the oyster Crassostrea gigas. This latter concentration is also sufficient to cause the development of the condition known as imposex in the dogwhelk Nucella lapillus. In this case, TBT acts as a hormone disruptor affecting gender differentiation. Female dogwhelks develop male genital organs, including a pseudopenis and vas deferens, which become superimposed on their reproductive systems (Figure 2). Not surprisingly, the widespread and uncontrolled use of TBT-based paints worldwide during the 1970s and 1980s (especially free association paints) resulted in unacceptably high levels of TBT in some coastal areas of intense boat use. The main problems were in

205

enclosed bodies of water with poor flushing characteristics, such as harbors, dry-docks, marinas, estuaries, and bays. The first documented cases of serious biological impact on nontarget organisms came from Arcachon Bay in west France. The bay is a center of both oyster culture and of high yachting activity, and TBT originating from paints used on yacht hulls was held responsible for abnormal growth and reproductive failure of cultured oysters. The oyster farming industry was in a state of near collapse. Production of oysters fell in the early 1980s to c. 33–50% of the normal harvest (Table 1). At about the same time, TBT pollution was also linked to serious decline in populations of dogwhelks (N. lapillus). They were suffering from imposex and it was so severe that populations of these organisms in areas of high boating activity had become partially or totally sterile. Subsequently, imposex was described in more than 100 species of gastropods worldwide, and the condition in many of them has been used as a biological indicator of TBT contamination. It became evident from surveys using this indicator that TBT contamination had become a global problem. It was clearly necessary to control the use of TBT-based paints. Pleasure craft, especially yachts, were believed to be the main source of pollution in coastal waters and several countries banned the use of these antifoulants on vessels o25 m in length. The French Government was the first to react, due to the oyster crisis, introducing controls in 1982. The UK followed in 1987, the USA in 1988, and subsequently Canada, New Zealand, Australia, South Africa, Hong Kong, and most European countries reacted in similar ways. The regulations were effective in reducing TBT contamination. Declining concentrations of TBT were reported during the 1990s in water samples, sediments (although less rapidly due to their persistence in them), and tissues of mollusks in monitoring programs worldwide. More dramatically, there have been enormous improvements in the health of marine biota which had been affected by TBT pollution in the previous decade. Oyster production in Arcachon Bay recovered spectacularly to former levels almost immediately after the introduction of regulations in France in 1982 (Table 1), and there was widespread recovery of populations of dogwhelks suffering from imposex. Nevertheless, commercial harbors, especially those with dry-docking and repair facilities, were still hot spots of TBT contamination (see Table 2), due to the accumulation of butyltins themselves and paint chips containing them in sediments (Figure 3). This has remained an unfortunate legacy of the TBT era and there are still management implications with respect to the disposal of contaminated sediments. TBT may


206


p

gp vd

a

v Stage

cg

1

2

3

4

n b A

B

A

6

5

B

Figure 2 Six stages in the development of imposex in the dogwhelk N. lapillus. Abbreviations: a, anus; b, ‘blister’; cg, capsule gland; gp, genital papilla; n, ‘nodule’; p, penis; v, vulva; vd, vas deferens. Reproduced from Gibbs PE, Bryan GW, Pascoe PL, and Burt GR (1987) The use of the dogwhelk (Nucella lapillus) as an indicator of TBT contamination. Journal of the Marine Biological Association of the United Kingdom 67: 507–524.

be remobilized from them either as a result of dredging or natural events, such as storms or tidal flow. A ban on the use of TBT as the active biocide in antifoulants became inevitable. The International Maritime Organization’s Convention on the Control of Harmful Antifouling Systems (AFS 2001) resolved that the application of TBT-based paints be prohibited from 1 January 2003. Furthermore, it stated that by 1 January 2008 (effective date), ships either (1) shall not bear such compounds on their hulls or

external parts or surfaces; or (2) shall bear a coating that forms a barrier to such compounds leaching from the underlying noncompliant antifouling system.

Meeting the Challenge: Alternative Antifoulants The ban on the use of TBT created a need for new coating systems that have the same antifouling



efficiency as TBT-based coatings, but are more environmentally acceptable. The success of TBT-based coatings inhibited research on alternatives and this did not commence in earnest until the 1990s when the ban became probable. Since then, there has been substantial research effort on the biology and chemistry of marine fouling and natural antifouling systems. It has greatly increased our understanding of how to reduce and perhaps prevent fouling in the future. Research on alternative coatings has followed two major lines: (1) the development of antifouling Table 1 Production of oysters in Arcachon Bay, France, between 1978 and 1985 Period

Production (t)

1978–79 1979–80 1980–81 1981–82 1982–83 1983–84 1984–85

10 000 6000 3000 5000 8000 12 000 12 000

Reproduced from Alzieu C (1996) Biological effects of tributyltin on marine organisms. In: de Mora SJ (ed.) Tributyltin: Case Study of an Environmental Contaminant, pp. 167–211. Cambridge: Cambridge University Press.

207

coatings that use the same binder delivery mechanism as the banned TBT coatings but contain different biocidal products; and (2) the development of nonbiocidal alternatives. Biocidal Antifouling Coatings

There are, broadly, two types of biocidal coatings on the market today. First, those that have hydrating binders, also known as ablative or solid matrix coatings, and, second, those that have hydrolyzing binders and mimic TBT SPC coatings. Copper compounds, such as cuprous oxide (Cu2O), copper thiocyanate (CuSCN), or metallic copper, are now used as the principal biocide. There is nevertheless concern about levels of this metal in the marine environment, and regulatory bodies are already turning their attention to copper contamination in ports, harbors, and coastal zones. Some actions have already been taken. For instance, the use of copper has been banned as an antifouling biocide on pleasure craft in Baltic Sea and west coast areas of Sweden, and release rate limits have been imposed in Canada, Denmark, and are about to be enforced in the USA. However, the situation is complex due to the issue relating to the bioavailablity of copper. Copper, when released into the marine environment, is only bioavailable in the ionic form. Due to the high

Table 2 Two ‘hot spots’ of TBT contamination in Puget Sound (USA): gradients of diminishing concentrations of organotins (total TBT and its metabolites DBT and MBT) in tissue samples, and of measures of imposex (VDSI and RPSI), in the dogwhelks Nucella emarginata and N. lamellosa from a shipyard at Anacortes and the complex of commercial harbors and shipyards at Seattle Source

Species

Distance of site from source (km)

Total organotins % with imposex (ng g 1)

Anacortes shipyard

N. lamellosa

0

668

a

1.3 2.4 6.5 8.0

212 61 43 24

0.9 1.9 2.3 8.0 9.0 0.9 1.9 2.3 8.0 9.0

Seattle

Seattle

N. emarginata

N. lamellosa

VDSI

RPSI

a 100 100 100

4.0 2.8 3.3

1.2 o0.1 o0.1

146 65 42 72

100 100 100 94

4.0 3.9 1.9 1.9

14.1 9.1 o0.1 o0.1

131 141 68 66 42

100 100 100 100 100

4.0 4.0 3.1 3.5 3.2

19.3 4.3 6.7 0.5 o0.1

a

Population consisted of males only. DBT, dibutyltin; MBT, monobutyltin; VDSI, vas deferens sequence index; RPSI, relative penis size index. Reproduced from Evans SM, Barnes N, Birchenough AC, Brancato MS, and Hardman E (2001) Tributyltin contamination in two estuaries and adjacent coasts: Puget sound, Washington and Narragansett Bay, Rhode Island (USA). Invertebrate Reproduction and Development 39: 221–229.


208


Figure 3 Residual paint flakes in sediments may be the cause of long-term pollution problems. The illustration shows (from left), tin and copper elemental maps, and scanning electron microscope microphotograph, of a paint flake recovered from a dredge in a working shipyard. Twelve layers of paint are visible in the flake, reflecting the coatings cycle of a commercial vessel with the newest layers at the top. Antifouling layers are typified by high copper content. Two layers of coating containing high tin and copper levels are present. Published with permission from Dickon Howell.

proportion of humic substances in seawater, this ionic copper is quickly bound to forms that are not bioavailable. The result of this is that environmental copper may exceed Environmental Quality Standards (EQS) levels if it is measured using the total dissolved form, even though the bioavailable copper is actually below the approved limits. In any case, copper as the sole biocide may not give complete fouling protection. There are several algal species (e.g., Enteromorpha spp., Ectocarpus spp., Achnanthes spp.) that show marked physiological tolerance to copper, and effective coatings must include organic booster biocides to combat them. These boosters include both copper and zinc pyrithione (antifungal agents used in anti-dandruff shampoos and outdoor paints), diuron (a substituted urea herbicide), ChlorothalonilTM (tetrachloroisophthalonitrile), dichlofluanid (a phenylsulfamide acaricide), Irgarol 1051TM (a triazine), Sea Nine 211TM (dichloro-isothiazolone), zineb and maneb (both dithiocarbomate fungicides containing zinc and manganese, respectively). Many of them, including zineb, maneb, ChlorothalonilTM, and dichlofluanid have a history of use as agricultural pesticides. When the new biocidal coating formulations were brought to market, there was initial concern that little research had been done on the environmental persistence of these booster biocides in the marine environment in terms of their degradation, partitioning, and bioavailability. This concern has been justified as both diuron and Irgarol 1051TM have been shown to have relatively high environmental persistence. Following monitoring studies, Irgarol 1051TM, a photosystem II inhibitor, has been shown to have very slow degradation (half-life of 100 days in marine sediments) and a high bioaccumulation

factor. The latter was shown, for example, in monitoring studies of the seagrass Zostera marina, a species of high international conservation importance. Concentrations of Irgarol 1051TM were found to be 25 000 times higher in Zostera tissues than in ambient (o0.003 mg l 1) seawater. It is for these reasons that Irgarol 1051TM has also been withdrawn from use by the UK Health and Safety Executive and also in Denmark. Similarly, diuron has since been withdrawn from use by the UK Health and Safety Executive and is being considered for withdrawal by the European Commission. The ban may spread to some of the other formulations, even those that seem to be more environmentally friendly. For example, Sea Nine 211TM and both copper and zinc pyrithione have relatively low environmental persistence, with half-lives of under 48 h, but some people believe that their introduction into the marine biosphere is still undesirable. Even a short pulse of a rapidly degradable pollutant at low concentrations can affect plankton populations significantly, thus causing effects that can cascade through a community. It can also be argued, with the benefit of hindsight, that some of the biocides that have been used as antifoulants should never have been allowed to come on to the market in the first place. Some of them had been used successfully in terrestrial applications but their behavior, once released in the marine environment, was unforeseen. This is well illustrated by experiences with TBT, whose sublethal impacts at extremely low, almost undetectable concentrations in seawater were not appreciated until they were in widespread use in marine coatings. A major problem when TBT coatings were introduced was that there was no standardized method for testing antifoulants



that gave results that were comparable to a ‘realworld situation’. Regrettably this is still the case. Particular biocides can still ‘pass’ ecotoxicological assays carried out in isolation of other assessments. However, such tests provide no information on, for example, how a biocidal coating system performs when it is applied to a ship’s hull, what synergistic effects there may be between active biocides, and precisely what the biocide’s environmental fate will be. Fortunately, progress is being made and considerable attention is now given to the environmental persistence and toxicity of new biocides and evidence is accumulating that the non-TBT SPC coatings are performing well. There are reports that there are intervals of 4–5 years between recoatings of new formulations, rivaling the achievements of the old TBT SPC coatings. Nonbiocidal Alternatives

In an ideal world, it would be unacceptable to release any toxic compound into the environment, and the ultimate aim of research workers and the regulators must be to eliminate the use of biocides in antifouling coatings altogether. There are many promising areas of research and development of nonbiocidal alternatives. These nontoxic coating systems fall into three categories: (1) foul release coatings; (2) socalled ‘smart’ coatings; and (3) hard marine coatings in conjunction with regular cleaning. Foul release coatings prevent the settlement of marine biota by having very low surface energy and lower surface roughness than biocidal coatings, providing a surface onto which organisms have great difficulty attaching themselves. Settlement can occur if vessels are stationary for extended periods but the adhesion between the fouling and the coating surface is very weak, so hydrodynamic forces will remove the organism when the vessel is traveling at relatively high speed of 15 knots or more. The majority of foul release systems are based on silicone as it has a low surface energy, a low elastic modulus, a microphase separation structure of hydrophilic and hydrophobic layers, and a nonfixed surface layer containing oil materials, which plasticize the coating and reduce adhesion strength. Silicones also provide a smoother surface than SPCs, making them more hydrodynamically efficient. Nevertheless, they are not yet seen as a viable option for slow-moving ocean-going container ships and are primarily used on high-speed coastal vessels. ‘Smart’ coatings are at the cutting edge of antifouling research. They include super-hydrophobic/ hydrophilic surfaces, use of magnetic fields, piezoelectric stimulation, bubble curtains to prevent fouling while the vessel is stationary, and biomimetic

209

surfaces. A promising approach is in the field of biomimetics. Coatings have been developed mimicking natural surfaces such as the shells of mollusks and crustaceans or the skin of fish and marine mammals. These surfaces have distinct nanoprofiles which inhibit fouling due to their bioenergetic properties. The underlying principle is that of attachment theory which states that attachment of a sphere (spore) to a flat substratum leads to a net increase in total surface area and an increase in work or energy exerted on the system. Attachment therefore will be preferential in valleys that are of similar dimension to the body size of the spore, resulting in a net decrease in surface area and therefore a reduction in hydrodynamic stress. If a surface can be created that is energetically unsuitable to a settling spore (i.e., has few attachment points), settlement will be inhibited (e.g., Figure 4). Although these technologies have not been tested on a working vessel as yet, initial results are encouraging and commercial-scale testing will become a reality in the near future, particularly in the case of the SharkletTM coating. The use of hard marine coatings that require regular in-service cleaning (every 2–3 weeks) is a trend that is rising, especially in the luxury cruise liner market. However, this process brings with it an increased threat of the spreading of invasive species, because organisms picked up in one part of the world may be deposited at another when the vessel hull is cleaned. This threat has not been well publicized because legislative drivers for regulating against the translocation of invasive species have mainly focused on ballast water. It is, nevertheless, a real one, and is causing particular concern in countries, such as Australia and New Zealand, where threats from

Figure 4 A scanning electron microscope microphotograph of the surface profile of the newly developed SharkletTM coating. Courtesy of Dr. Tony Brennan, University of Florida.


210


invasive species present particularly acute problems. Further development of this particular method of fouling control will probably depend on significant improvement of underwater hull cleaning facilities that can ensure containment of all organic matter cleaned from the hull.

See also Eutrophication. Metal Pollution. Pollution: Approaches to Pollution Control. Pollution: Effects on Marine Communities. Pollution, Solids. Shipping and Ports. Ships. Transition Metals and Heavy Metal Speciation.

Further Reading Alzieu C (1996) Biological effects of tributyltin on marine organisms. In: de Mora SJ (ed.) Tributyltins: Case Study of an Environmental Contaminant, pp. 167--211. Cambridge, UK: Cambridge University Press. Alzieu C (1998) Tributyltin: Case study of a chronic contaminant in the coastal environment. Ocean and Coastal Management 40: 23--36. de Mora SJ (ed.) (1996) Tributyltin: Case Study of an Environmental Contaminant. Cambridge, UK: Cambridge University Press.

Evans SM (1999) TBT contamination: The catastrophe that never happened. Marine Pollution Bulletin 38: 629--636. Evans SM (1999) TBT or not TBT? That is the question. Biofouling 14: 117--129. Evans SM, Barnes N, Birchenough AC, Brancato MS, and Hardman E (2001) Tributyltin contamination in two estuaries and adjacent coasts: Puget sound, Washington and Narragansett Bay, Rhode Island (USA). Invertebrate Reproduction and Development 39: 221--229. Gibbs PE, Bryan GW, Pascoe PL, and Burt GR (1987) The use of the dogwhelk (Nucella lapillus) as an indicator of TBT contamination. Journal of the Marine Biological Association of the United Kingdom 67: 507--524. Hunter JE and Cain P (1996) Shipping and the environment: Is compromise inevitable? Anti-Fouling Coatings in the 1990s – Environmental, Economic and Legislative Aspects, Paper 16, IMAS 96, 22–24 October 1996. London: The Institute of Marine Engineers. Stebbing ARD (1985) Organotins and water quality: Some lessons to be learned. Marine Pollution Bulletin 16: 383--390. Yebra DM, Kiil S, and Dam-Johansen K (2004) Antifouling technology – past, present and future steps towards efficient and environmentally friendly antifouling coatings. Progress in Organic Coatings 50: 75--104.


ARCTIC OCEAN CIRCULATION B. Rudels, Finnish Institute of Marine Research, Helsinki, Finland & 2009 Elsevier Ltd. All rights reserved.

Introduction The Arctic Ocean is the northernmost part of the Arctic Mediterranean Sea, which also comprises the Greenland Sea, the Iceland Sea, and the Norwegian Sea (the Nordic Seas) and is separated from the North Atlantic by the 500–850-m-deep Greenland– Scotland Ridge. The Arctic Ocean is a small, 9.4 106 km2, enclosed ocean. Its boundaries to the south are the Eurasian continent, Bering Strait, North America, Greenland, Fram Strait, and Svalbard. The shelf break from Svalbard southward to Norway closes the boundary. The Arctic Ocean lies almost entirely north of and occupies most of the region north of the polar circle. More than half of its area, 53%, consists of large, shallow shelves, the broad Eurasian marginal seas: the Barents Sea (200–300 m), the Kara Sea (50–100 m), the Laptev Sea (o50 m), the East Siberian Sea (o50 m) and the Chukchi Sea (50–100 m), and the narrower shelves north of North America and Greenland. The deep Arctic Ocean comprises two major basins, the Eurasian Basin and the Canadian (also called Amerasian) Basin, separated by the approximately 1600-m-deep Lomonosov Ridge. The Eurasian Basin is further divided into the Nansen and Amundsen basins by a mid-ocean ridge (the Gakkel Ridge), while the Canadian Basin is separated by the Alpha Ridge and the Mendeleyev Ridge into the Makarov and the Canada Basins. The Amundsen Basin is the deepest (B4500 m), while the maximum depths of the Makarov and the Nansen Basins are B4000 m. The Canada Basin is slightly shallower (B3800 m) but by far the largest (Figure 1). The Arctic Ocean water masses are primarily of Atlantic origin. Atlantic waters (AWs) enter the Arctic Ocean from the Nordic Seas through the 2600-m-deep Fram Strait and over the B200-m-deep sills in the Barents Sea. The Arctic Ocean also receives low-salinity Pacific water through the shallow (45 m) and narrow (50 km) Bering Strait. The outflows occur through Fram Strait and through the shallow (150–230 m) and narrow channels in the Canadian Arctic Archipelago.

The physical oceanography of the Arctic Ocean is shaped by the severe high-latitude climate and the large freshwater input from runoff, B0.1 Sv (1 Sv ¼ 1 106m3 s 1), and net precipitation, B0.07 Sv. The Arctic Ocean is a strongly salinity stratified ocean that allows the surface water to cool to freezing temperature and ice to form in winter and to remain throughout the year in the central, deep part of the ocean. The water masses in the Arctic Ocean can, because of the stratification, be identified by different vertical layers. Here five separate layers will be distinguished: 1. The B50-m-thick upper, low-salinity polar mixed layer (PML) is homogenized in winter by freezing and haline convection, while in summer the upper 10–20 m become diluted by sea ice meltwater. The salinity, S, ranges from 30 to 32.5 in the Canadian Basin and between 32 and 34 in the Eurasian Basin. 2. The 100–250-m-thick halocline, with salinity increasing with depth, while the temperature remains close to freezing, 32.5oSo34.5 (Figure 2). 3. The 400–700-m-thick Atlantic layer historically defined as subsurface water with potential temperature, y, above 0 1C, 34.5oSo35. 4. The intermediate water below the Atlantic layer that communicates freely across the Lomonosov Ridge, 0.5 1Coyo0 1C, 34.87oSo34.92. 5. The deep and bottom waters in the different basins, 0.55 1Coyo 0.5 1C, 34.92oSo34.96 (Canadian Basin), 0.97 1Coyo 0.5 1C, 34.92o So34.945 (Eurasian Basin). There are large lateral variations of the characteristics in these layers that depend upon the circulation and upon the mixing processes in the Arctic Ocean (Figures 2 and 3). A more detailed classification, especially for the deeper water masses, is given in Table 1. It should be kept in mind that this classification is not unique and that several others exist in the literature.

Circulation The circulation of the uppermost layers of the Arctic Ocean has mainly been inferred from the ice drift, determined from satellites and from drifting buoys, while at deeper levels primarily the distributions of temperature and salinity and more recently of other tracers have been used to deduce the movements of the different water masses, often with


211

212

ARCTIC OCEAN CIRCULATION

135° E

90° E

180°

45° E

35° W

0°

90° W

45° W

Figure 1 Map of the Arctic Mediterranean Sea showing geographical and bathymetric features. The bathymetry is from IBCAO (the International Bathymetric Chart of the Arctic Ocean; Jakobsson et al., 2000; 2008) and the projection is Lambert equal area. The 500 and 2000 m isobaths are shown. All maps used here are made by Martin Jakobsson (personal communication). BIT, Bear Island Trough; CB, Canadian Basin; EB, Eurasian Basin; GFZ, Greenland Fracture Zone; MJP, Morris Jessup Plateau; JMFZ, Jan Mayen Fracture Zone; SAT, St. Anna Trough; YM, Yermak Plateau; VC, Victoria Channel; VS, Vilkiltskij Strait; FJL, Franz Josef Land; BS, Barrow Strait; HG & CS, Hell Gate and Cardigan Sound.

the assumption that the circulation is largely controlled by the bathymetry. Direct, moored current measurements have been scarce and mostly confined to the continental slope and to the Lomonosov Ridge. These measurements have confirmed the importance of the bathymetry for the circulation. In the deep basins current measurements have been made from drifting ice camps and more recently also from autonomous ice-tethered platforms, relaying the observations via satellite to shore. Subsurface drifters are just beginning to be used and observational efforts during the International Polar Year (IPY) 2007– 09 are likely to significantly increase the knowledge of the circulation in the Arctic Ocean. The motions of the ice cover and the surface water are predominantly forced by the wind, and the atmospheric high-pressure cell over the Arctic creates the anticyclonic Beaufort gyre in the Canada Basin. Ice leaks from the offshore side of the gyre and joins the Transpolar Drift (TPD) that brings ice from the Canada Basin toward Fram Strait. A second branch originating from the Siberian shelves, mainly from

the Laptev Sea, carries ice across the Eurasian Basin. About 90% of the ice export (0.09 Sv) passes through Fram Strait (Figure 4). The PML and the halocline are maintained by river runoff, ice melt, and the inflow of low-salinity water through Bering Strait. The lowest surface salinities and the thickest halocline are therefore observed in the Canada Basin. The inflow through Bering Strait, although affected by local winds, is in the last instance driven by a higher sea level in the North Pacific as compared to the Arctic Ocean. This creates a pressure gradient that forces the Pacific water northward into the Arctic Ocean. The Bering Strait inflow continues across the Chukchi Sea in four branches. One branch enters the East Siberian Sea, while one of the central inflow branches enters the Arctic Ocean along the Herald Canyon west of the Chukchi Plateau, and the other passes via the Central Gap east of the Chukchi Plateau into the Canada Basin. The easternmost branch reaches the Canada Basin along the Barrow Canyon close to Alaska. River runoff, mainly from the Mackenzie



213

0 PML

CB 400

600 −2

1 −1 0 Potential temperature (°C)

32.0

2

33.0

34.0

35.0

Salinity

3

0

5

27

1

-27.

2

26.5

−1 28

Potential temperature (°C)

Halocline

AB MB

26

Pressure (dbar)

NB 200

−2 32.0

32.5

33.0

33.5 Salinity

34.0

34.5

35.0

Figure 2 Potential temperature and salinity profiles and yS curves from the upper layers of the Nansen Basin (NB, dark yellow), Amundsen Basin (AB, green), Makarov Basin (MB, magenta), and Canada Basin (CB, blue). The PML and the halocline are indicated in the salinity profiles. Above the PML, the low-salinity layer due to seasonal ice melt is seen. The temperature maximum in the Canada Basin is due to the presence of Bering Strait Summer Water (BSSW) and the temperature minimum below indicates the upper halocline with S B 33.1, deriving from the colder, more saline Bering Strait Winter Water (BSWW) and from brine release and haline convection in the Chukchi Sea. No halocline is present in the Nansen Basin, only a deep winter mixed layer between the thermocline and the seasonal ice melt layer with temperature close to freezing. The curved shape of the Nansen Basin thermocline as seen in the yS diagram suggests wintertime haline convection with dense, saline parcels penetrating into the thermocline. Adapted from Rudels B, Jones EP, Schauer U, and Eriksson P (2004) Atlantic sources of the Arctic Ocean surface and halocline waters. Polar Research 23: 181–208.

River, adds freshwater to the Canada basin. The runoff peaks in early summer (June). The river runoff as well as most of the Pacific water becomes trapped in the anticyclonic Beaufort gyre, forming an oceanic high-pressure cell in the southern Canada Basin. The Pacific water leaves the Beaufort gyre and the Arctic Ocean mainly through the Canadian Arctic Archipelago, but a smaller fraction also exits, at least intermittently, through Fram Strait. Warm AW crosses the Greenland–Scotland Ridge and continues toward the Arctic Ocean in the Norwegian Atlantic Current. The Norwegian Atlantic Current splits as it reaches the Bear Island Trough. One part enters the Barents Sea together with the Norwegian Coastal Current, which carries low-salinity water from the Baltic Sea and runoff from the Norwegian coast. The remaining part continues as the West Spitsbergen Current to Fram Strait. There the current again splits. Some AW

recirculates westward in the strait to join the southward-flowing East Greenland Current, and the rest enters the Arctic Ocean in two streams. One stream flows over the Svalbard shelf and slope, the other passes west and north around the Yermak Plateau and then continues eastward, eventually joining the inner stream at the continental slope east of Svalbard. The deeper outer stream also transports intermediate and deep waters from the Nordic Seas into the Arctic Ocean. As the AW enters the Arctic Ocean, it encounters, and melts, sea ice, and the upper part becomes colder and less saline. In winter this upper layer is homogenized by convection and mechanical stirring and cooled to freezing temperature. The salinity of the upper layer is 34.2–34.4, and it is advected with the warm AW core eastward along the Eurasian continental slope. In summer, seasonal ice melt creates a low-salinity upper layer, which is removed in


214

Pressure (dbar)


0

0

500

500

1000

1000

1500

1500

2000

2000

2500

2500

3000

3000

3500

3500

4000 −2

2

0

4000 34.7

4

34.8

34.9

Potential temperature (°C) 0.5


.70

27

AW

.97

0.0

27 2

dAW(\) 1

AAW

dAAW(/) CBDW

0

uPDW(\) 0.5 30.444 1.5 35.142 2.5 39.738

uPDW(\) −1

CBDW −0.5

AIW(/)

−1.0

EBDW

EBDW AIW(/)

NDW

NDW −2 34.7

35.1

Salinity

4 3

35

−1.5 34.8

34.9 Salinity

35

35.1

34.86 34.88

34.9

34.92 34.94 34.96

Salinity

Figure 3 Characteristics of the water columns in different parts of the Arctic Mediterranean. Upper row: potential temperature and salinity profiles; lower row: yS curves. Green: The Greenland Sea, the ultimate source of the Arctic Intermediate Water (AIW) and the Nordic Seas Deep Water (NDW) entering the Arctic Ocean. Red: The West Spitsbergen Current in Fram Strait carrying warm Atlantic Water (AW), AIW, and NDW into the Arctic Ocean. Dark yellow: The Fram Strait branch at the continental slope of the Nansen Basin. Magenta: The interior Nansen Basin. Cyan: The Amundsen Basin. Black: The Makarov Basin. Blue: The Canada Basin. The shift in depth of the temperature maximum between the Makarov Basin and the Canada Basin is due to the stronger presence of Pacific water (PW) in the Canada Basin, displacing the deeper part of the water column. Note that the Canadian Basin deep waters becomes warmer than the Eurasian Basin deep waters below 1000 m and more saline below 1500 m, above the sill depth of the Lomonosov Ridge. In the yS diagrams, (/) indicates a stratification unstable in temperature or salinity, (\) indicates a stratification stable in both components. AAW, Arctic Atlantic Water; dAW, dense Atlantic Water; dAAW, dense Arctic Atlantic Water; uPDW, upper Polar Deep Water; CBDW, Canadian Basin Deep Water; EBDW, Eurasian Basin Deep Water. The s1.5 isopycnal shows the density at the sill depth of the Lomonosov Ridge and the s2.5 isopycnal the density at sill depth in Fram Strait.

winter by freezing and the upper layer is homogenized down to the thermocline. No cold halocline is present between the mixed layer and the AW in the Nansen Basin (see Figure 2). The Arctic Intermediate Water (AIW) and the Nordic Seas Deep Water (NDW) that enter the Arctic Ocean in the

West Spitsbergen Current can be identified west and north of the Yermak Plateau as less saline and colder anomalies. Farther to the east these signals in temperature and salinity gradually disappear, but signs of the Nordic Seas waters can still be detected by other tracers, for example, CFCs.



Table 1

215

Simplified water mass classification for the Arctic Ocean

Water masses Upper waters (syo27.70) Polar mixed layer (Upper) halocline (Lower) halocline Intermediate waters I (27.70osyo27.97) Atlantic Water Arctic Atlantic Water

Abbreviation

Definition

Main source or origin

PML

32oSo34 32.5oSo33.5 33.5oSo34.5

Arctic Ocean Chukchi Sea Nansen Basin, Barents Sea

AW AAW

2oy 0oyo2

West Spitsbergen Current Arctic Ocean (transformed)

Intermediate waters II (27.97osy, s0.5o30.444) Dense Atlantic Water DAW Dense Arctic Atlantic Water DAAW Arctic Intermediate Water AIW Upper Polar Deep Water UPDW

0oy, 0oy, yo0, yo0,

Deep waters (30.444os0.5) Nordic Seas Deep Water Canadian Basin Deep Water Eurasian Basin Deep Water

So34.915 0.6oy, 34.915oS yo 0.6, 34.915oS

NDW CBDW EBDW

In the Barents Sea, the AW remains in contact with the atmosphere during most of its transit. It is cooled significantly and becomes freshened by net precipitation and by the melting of sea ice. Some of the water entering the Barents Sea returns as colder, denser water to the Norwegian Sea in the Bear Island Channel, but the major part reaches the eastern Barents Sea. In the Barents Sea, the AW becomes separated into three different water masses. (1) The bulk of the inflow is cooled and freshened by air–sea interaction and becomes denser. (2) Some of the AW reaches the shallow areas west of Novaya Zemlya and becomes transformed into saline, dense bottom water by the ice formation and brine rejection. (3) The upper part of the AW interacts with sea ice and a less saline, upper layer is formed by ice melt, which in winter becomes homogenized down to the thermocline. These waters all enter the Arctic Ocean, mainly by passing between Novaya Zemlya and Franz Josef Land into the Kara Sea and then continuing in the St. Anna Trough to the Arctic Ocean. However, some water reaches the Arctic Ocean directly from the Barents Sea along the Victoria Channel. The Barents Sea branch follows the eastern side of the St. Anna Trough and then continues along the continental slope as an almost 1000-m-thick, cold, and weakly stratified water column, displacing the warmer, more saline Fram Strait inflow branch from the slope and deflecting the denser basin waters toward larger depths. Strong isopycnal mixing between the two branches takes place and inversions and irregular intrusive layers are formed (Figure 5, right column).

unstable stable in unstable stable in

in S (/) y and S (\) in S or y (/) S and y (\)

West Spitsbergen Current Arctic Ocean (transformed) Greenland Sea Arctic Ocean Greenland Sea Canadian Basin Eurasian Basin

North of the Laptev Sea the two inflow branches have largely merged and bottom temperatures above 0 1C indicate that Fram Strait branch water again is present at the slope. The boundary current splits at the Lomonosov Ridge with one part continuing into the Canadian Basin, the rest returning along the ridge and the Amundsen Basin toward Fram Strait. The AW temperature in the boundary current is significantly reduced already north of the Laptev Sea. This could partly be due to mixing between the two branches and partly be the result of heat loss to the upper layers and the ice. However, it may also be an indication that some of the Fram Strait branch recirculates already in the Nansen Basin. The returning Fram Strait branch then becomes more prominent in the northern Nansen Basin, while the Barents Sea branch dominates in the Amundsen Basin and along the Lomonosov Ridge (Figure 5, left column). The bulk of the Barents Sea branch in the St. Anna Trough is denser, colder, and less saline than the Atlantic layer and forms a distinct salinity minimum beneath the AW in the Amundsen Basin and above the Gakkel Ridge. This minimum identifies the upper Polar Deep Water (uPDW) in the Eurasian Basin and is located higher in the water column and is more distinct than the AIW minimum in the Nansen Basin deriving from the Nordic Seas. This is about the only area of the Arctic Ocean where a part of the water column is unstably stratified in salinity (Figures 3 and 5). One part of the Barents Sea inflow, mostly comprising water from the Norwegian Coastal Current, enters the Kara Sea through the Kara Strait south of Novaya Zemlya. It mixes with the runoff from Ob


216


90° E

135° E 180° SCC AC

45° E SB MC TPD ACC

ESC BI C

BG

WSC

135° W

NAC C NC

JMC EGC EIC

WG C

60

BC

°N

IC

NAD

0°

50

°N

90° W

45° W

Figure 4 The circulation of the upper layers of the Arctic Mediterranean Sea. Warm Atlantic currents are indicated by red arrows, cold, less-saline polar and arctic currents by blue arrows. Low-salinity transformed currents are shown by green arrows. The maximum ice extent is shown by a blue and the minimum ice extent by red striped line. AC, Anadyr Current; ACC, Alaskan Coastal Current; BC, Baffin Current; BIC, Bear Island Current; BG, Beaufort gyre; EGS, East Greenland Current; EIC, East Iceland Current; ESC, East Spitsbergen Current; IC, Irminger Current; JMC, Jan Mayen Current; MC, Murman Current; NAD, North Atlantic Drift; NAC, Norwegian Atlantic Current; NCC, Norwegian Coastal Current; SB, Siberian branch (of the Transpolar Drift); SCC, Siberian Coastal Current; TPD, Transpolar Drift; WGC, West Greenland Current; WSC, West Spitsbergen Current.

and Yenisey, forming the low-salinity water present on the Kara Sea shelf. Most of this shelf water flows through the Vilkiltskij Strait into the Laptev Sea, where it receives additional freshwater, mostly from the Lena River. The main export of shelf water from the Eurasian shelves to the deep basins occurs across the Laptev Sea shelf break into the Amundsen Basin, and the low-salinity shelf water overruns the boundary current and the two inflow branches. The upper parts of the Fram Strait and the Barents Sea branches become isolated from the ice, the sea surface, and the atmosphere and evolve into halocline waters. The Fram Strait branch supplies the halocline in the Amundsen Basin and the Makarov Basin and the lower halocline, beneath the Pacific water, in the northern part of the Canada Basin. The Barents Sea branch halocline, which remains at the continental slope, is affected by the stronger mixing occurring over the steep topography and becomes warmer by mixing with AW from below. It eventually supplies

the lower halocline in the southern Canada Basin. The shelf water outflow supplies the PML in the Amundsen and Makarov Basins but it has high enough salinity also to contribute to the halocline waters in the Canada Basin, beneath most of the Pacific waters (Figure 6). The properties of the Atlantic and intermediate waters in the Canada Basin are distinctly different from those in the Eurasian Basin, indicating that the waters of the boundary current become transformed along their paths in the Canadian Basin. The temperature maximum in the Atlantic layer becomes colder and less saline, while in the intermediate water range the temperature and the salinity increase and the intermediate uPDW salinity minimum disappears. The yS curves below the temperature maximum approach a straight line toward higher salinities and lower temperatures with increasing depth (Figure 3). The boundary current circulates cyclonically around the Canada Basin and splits at the different


Pressure (dbar)


0

0

500

500

1000

1000

1500

1500

2000

2000

2500

2500

Pressure (dbar)

3000 −1

0 2 1 Potential temperature (°C)

3000 −1

3

0

0

500

500

1000

1000

1500

1500

2000

2000

2500

2500

3000 34.7

34.9

34.8

35

0 2 1 Potential temperature (°C)

3000 34.7

35

Salinity

3.0

3.0

2.5

2.5

2.0

2.0

1.5

1.5

0

.7

0.5

.7 0

0 −0.5 −1.0 34.7

34.8

34.9

27

7

.9

1.0

97 27 .

1.0

27


3

34.9

34.8

Salinity

217

0.5 30.444 1.5 35.142 2.5 39.738 35

27

0.5 0 −0.5 −1.0 34.7

Salinity

34.8

34.9

35

Salinity

Figure 5 Potential temperature and salinity profiles and yS curves showing the interaction and interleaving between the Fram Strait branch and the Barents Sea branch north of Severnaya Zemlya (right column) and the water properties of the Nansen and Amundsen basins offshore of the Fram Strait branch (left column). Red stations: Fram Strait branch, blue station: the Barents Sea branch, black and cyan stations: active mixing between the branches. Interleaving is present not only in the AW but also in the deeper layers. Offshore of the Fram Strait branch, the warm, saline intrusions in the Nansen Basin (black and magenta stations) suggest a close recirculation of the Fram Strait branch in the Nansen Basin, while the colder, less saline intrusions in the Amundsen Basin (green and blue stations) indicate that the intermediate part of the water column here is dominated by Barents Sea branch waters.


218


135° E

90° E

180°

45

135° W

°N

60

0°

°N

50

90° W

45° W

Figure 6 Circulation of the Atlantic-derived halocline waters and the distribution of Pacific water (orange) and Eurasian river runoff (green). The proposed source areas for the Fram Strait branch lower halocline water (black) and the Barents Sea branch lower halocline water (blue), and the circulation of these waters in the Arctic Ocean are indicated. RR, river runoff; PW, Pacific water; AW, Atlantic water. The cross indicates possible contribution of Barents Sea branch lower halocline water to the Baffin Bay bottom water. The isoline shown is 500 m. Based on Rudels B, Jones EP, Schauer U, and Eriksson P (2004) Atlantic sources of the Arctic Ocean surface and halocline waters. Polar Research 23: 181–208.

ridges (Figure 7). One loop enters the Makarov Basin along the Mendeleyev Ridge, and one loop penetrates into the northern Canada Basin at the Chukchi Plateau. The boundary current also enters the Canada Basin at the Alpha Ridge and the Makarov Basin at the North American side of the Lomonosov Ridge. These circulation loops and their interactions with the boundary current induce a time lag that makes the Atlantic and intermediate waters of the different basins distinct, even if the waters originate from the same inflow across the Lomonosov Ridge.

Processes The climatic forcing on the Arctic Ocean is strong; the large variation of incoming solar radiation, the severe cooling during the polar night, and the intense weather systems, either locally formed polar lows or cyclones advected from lower latitudes, all contribute in forming the Arctic Ocean characteristics. The strong stability of the deep Arctic Ocean basins confines the effects of these forcing fields to the ice-ocean surface and to the PML, but the shallow Arctic shelves, which

experience the largest impact of the forcing, the strong runoff in summer and excessive ice formation in winter, do not only add low-salinity upper water and ice to the central basins but also create waters dense enough to break through the stratification, transforming the deeper advected layers. Ice formation releases brine, and in lee polynyas, areas of open water where the ice is removed by offshore winds, sufficient ice may form to allow the released brine to overcome the initial low salinity on the shelves and create saline dense bottom waters at freezing temperature. These bottom waters eventually cross the shelf break and descend into the deep basins until their densities match those of the surrounding. They then merge with the ambient water. Less-dense plumes add colder water to the upper layers, supplying water to the halocline. This occurs in the Chukchi Sea, where the water of the Canada Basin upper halocline with salinity 33.1 is formed. In the Nansen Basin no halocline is present and the density range between the upper mixed layer and the AW is small. The shelf outflows then either enter the mixed layer or sink into and cool the thermocline and the Atlantic layer. The formation of



135° E

219

90° E

180°

45°

135° W

°N

60

0°

°

50 N

90° W

45° W

Figure 7 Schematics showing the circulation in the subsurface Atlantic and intermediate layers in the Arctic Mediterranean Sea. The interactions between the Barents Sea and Fram Strait inflow branches north of the Kara Sea as well as the recirculation and different inflow streams in Fram Strait and the overflows across the Greenland–Scotland Ridge are indicated. The isoline shown is 500 m. Based on Rudels B, Jones EP, Anderson LG, and Kattner G (1994) On the intermediate depth waters of the Arctic Ocean. In: Johannesen OM, Muench RD, and Overland JE (eds.) AGU Geophysical Monographs 85: The Polar Oceans and Their Role in Shaping the Global Environment, pp. 33–46. Washington, DC: American Geophysical Union.

halocline waters in the Eurasian Basin occurs, when the boundary current is overrun by the less saline shelf water, and the initial mixed layers become transformed into halocline waters. Comparison between the intermediate and deep waters in the Nordic Seas and in the Arctic Ocean shows that the shelf-slope convection not only brings salt but also heat into the deeper layers (Figures 3 and 5). This implies that the denser, more saline plumes are entraining warmer water, as they pass through the intermediate Atlantic layer. Observations of outflowing, cold and dense water from Storfjorden in southern Svalbard have shown that as the outflow reaches deeper than 1500 m, it has become warmer than the ambient water, supporting the idea of entraining boundary plumes. No plumes have yet been followed from their sources down the slope in the Arctic Ocean, but saline, warmer, and denser bottom layers have been observed deeper than 2000 m at the continental slope north of Severnaya Zemlya. The deep waters of the Canadian and Eurasian Basins differ significantly, the Canadian Basin Deep

Water (CBDW) being warmer and more saline than the Eurasian Basin Deep Water (EBDW). The difference in temperature was first believed due to the presence of the Lomonosov Ridge, blocking the cold deep water from the Nordic Seas, but the temperature difference starts well above the sill depth of the Lomonosov Ridge, showing that heat as well as salt is added to the uPDW and the CBDW in the Canadian Basin. The differences between the Amundsen and Nansen Basins are small but clear. The Nansen Basin is slightly less saline than the Amundsen Basin between 1600 and 2600 m due to a stronger presence of NDW and AIW. In the Amundsen Basin a mid-depth (1700 m) salinity maximum or at least a sharp bend in the yS curves is observed. It is strongest closer to Greenland but is present over most of the basin. This maximum derives from the CBDW that crosses the Lomonosov Ridge and then penetrates into the central Amundsen Basin from the Greenland continental slope. In the Nansen, Amundsen, and Canada Basins, the salinity of the deep water increases toward the


220


bottom, while the temperature goes through a minimum and then increases slightly before an isothermal and isohaline bottom layer is reached (Figure 3). Shelf-slope convection cannot explain the temperature minimum in the deep waters, but it could be due to advection between the basins. The deep water in the Makarov Basin has at the sill depth (2400 m) of the Alpha Ridge similar yS characteristics as the temperature minimum in the Canada Basin and could supply the minimum. In the Eurasian Basin this explanation does not work. A minimum is present in both the Amundsen and Nansen Basins and these minima lie too deep to derive from the inflow of colder NDW. One possibility is that the St. Anna Trough intermittently conduits colder, denser bottom water, formed in the Barents Sea, into the Nansen Basin. This cold water would enter below the warm layers of the Fram Strait branch and by entraining less heat it could contribute colder water to the deeper part of the water column than the slope convection occurring farther east around Severnaya Zemlya. Geothermal heat flux has been proposed as an energy source for heating and stirring the bottom layers, keeping them homogenous and gradually increasing their temperature, thereby creating the overlying temperature minimum. The high bottom salinity could then be a remnant from convection events that took place several hundred years ago and brought dense, saline, and cold water to the bottom, where it now is isolated and gradually becomes warmer, thicker, and less dense. Whether the bottom layers are kept homogenous by geothermal heating and convection from below, or if the stirring is due to mechanical mixing as the boundary plumes enter the bottom layer is an open question. However, geothermal heating and ventilation by shelf-slope convection are not two mutually exclusive processes and could operate simultaneously. The Makarov Basin is different. The salinity becomes constant with depth, while the temperature still decreases, creating a temperature-stratified layer above the isothermal and isohaline bottom water (Figure 3). This rules out the gradual heating of an old, fossil bottom layer but requires that colder, rather than more saline, water renews the bottom layer. This then excludes shelf-slope convection as a dominant process. One possibility is that colder water from the Amundsen Basin is brought across the Lomonosov Ridge, either intermittently, forced by, for example, topographically trapped waves, or flowing through yet undetected passages in the ridge. The central part of the Lomonosov Ridge, where such an exchange was believed to occur, has recently been surveyed. A gap was found but not as deep as expected, barely 1900 m, and more critically, the

water mass properties observed during the survey indicated that the flow of the densest water was from the Makarov Basin to the Amundsen Basin – in the wrong direction. This inflow of Makarov Basin water followed the Lomonosov Ridge toward Greenland and is likely to be one source, perhaps the most important one, for the mid-depth salinity maximum in the Amundsen Basin. A flow onto the shelves is needed to compensate for the supply to the PML in summer and the outflow of dense water to the deeper layers and the export of ice in winter. This flow could come from the Arctic Ocean, along the bottom in summer and at the surface in winter, or along horizontally separated inflow and outflow paths across the shelf break. It could also be supplied from behind, like the Bering Strait inflow and the inflow to the Barents Sea. These inflow shelves differ from the inner shelves of the Kara, Laptev, and East Siberian Seas, where a two-way exchange across the shelf break might be more likely, even necessary. However, the difference is not as large as it might appear. Most of the Barents Sea inflow actually passes the Kara Sea before it enters the deep basins and much of the runoff from Ob and Yenisey, together with a substantial fraction of the Barents Sea inflow, continues into the Laptev Sea and then enters either the Amundsen Basin or the East Siberian Sea. Such eastward flow is the usual fate of river plumes, which tend to follow the coastline eastward. This also holds for the East Siberian Coastal Current, which carries the runoff from perhaps the most ‘inner’ shelf of the Arctic Ocean, that of the East Siberian Sea, into the Chukchi Sea. The East Siberian Coastal Current occasionally, but rarely, passes south through Bering Strait into the Bering Sea. The circulation pattern in the Arctic Ocean outlined above gives the impression that water masses are advected, with little mixing, in loops through the different basin, replacing one vintage water mass with another. This is not entirely the case. The structure of the salinity and temperature profiles shows sign of intrusive mixing and the presence of lenses or eddies of anomalous water masses in the water column. Eddies were first reported in the Canadian Basin halocline and were there considered related to the outflow of dense water from the Chukchi Sea. These eddies are mostly anticyclonic, 10–20 km in diameter and highly energetic with maximum velocities above 0.3 m s 1. Eddies have subsequently been observed in all water masses of the Arctic Ocean, although the velocities associated with the deeper-lying eddies have so far not been determined. Their water characteristics imply that the eddies have traveled considerable distances as



coherent water bodies without their anomalous properties being removed by smaller-scale mixing, for example double-diffusive convection, merging the eddies with the Arctic Ocean water columns in different basins (Figure 8). The intrusive layers are particularly intriguing (Figures 5 and 8). The largest property amplitudes are found at the frontal zones but the layers appear to reach over entire basins, making their extent perhaps the largest one observed in the World Ocean. The inversions allow for the release of the potential energy stored in the unstably stratified component by double-diffusive convection, which can drive the interleaving layers across the basins, and this has been suggested as a mechanism for transfer of heat from the boundary current into the deep basins. For the layers to expand, along-layer gradients in heat and salt are required. In the interior of the basins such gradients are often absent, and an alternative explanation for the wide extent of the intrusive layers is that they are formed in the frontal zones, expand initially, driven by double-diffusive fluxes, until the potential energy, available in the stratification, is removed. After this the layers remain as

221

fossil structures that are advected around the main circulation gyres. The intrusions between the Fram Strait and the Barents Sea branch water observed in the Amundsen Basin and over the Gakkel Ridge have been interpreted as fossil structures, initially formed by the interactions between the Fram Strait and the Barents Sea branches north of the Kara Sea and now being advected toward Fram Strait (Figure 5). Interleaving layers are observed in all background stratifications: saltfinger unstable, diffusively unstable, as well as when both components are stably stratified. In the stable–stable situation, disturbances extending across the front are necessary to create the initial inversions. Differential diffusion, taking into account the more rapid diffusion of heat in weakly turbulent surroundings, has been proposed as a mechanism for creating interleaving in a stable–stable stratification. The time required to create layers would then be in the order of years. This appears long, since interleaving layers are found very close to the area where the parent water masses first meet also when both components are stably stratified.

Variability 2.5


2.0 1.5 1.0 0.5

7

27.9

0.5 30.444 1.5 35.142 2.5 39.738

0 −0.5 −1.0

34.86

34.88

34.9 Salinity

34.92

34.94

Figure 8 yS diagram showing eddies present in the intermediate and deep layers on a section taken in 1996 across the Eurasian Basin. The red station shows the warm and saline Canadian Basin Deep Water, the blue station indicates an isolated lens of cold, low-salinity Barents Sea branch water. The cyan and black stations show an eddy of Barents Sea branch water (cyan station) with warmer and more saline water in both the slope and the basin directions (black stations) and surrounded by interleaving structures like a ‘meddy’. This is in contrast to the slope to basin decrease in salinity and temperature seen in the interleaving in the Atlantic layer (although here also an eddy (not shown) was detected). Finally an eddy of warm, saline Fram Strait branch water, yellow station, was present in the colder, Barents Sea branch dominated Atlantic layer in the Amundsen Basin.

Until 1990 an underlying assumption has been that the Arctic Ocean, at least in its deeper parts, is reasonably quiet and unchanging and that observations made during a longer period, 10–20 years, could be merged and used to describe the basic hydrographic conditions. The observations in the 1990s proved the Arctic Ocean to be as variable and changing as any other ocean. An inflow of anomalously warm AW was reported in 1990 and has been observed propagating in the boundary current and into the different basins, changing the characteristics of the Atlantic layer in the Arctic Ocean. This warm inflow event persisted for almost a decade. Colder water then entered the Arctic Ocean for a short period after which the inflowing AW again became warmer. These inflow events can be traced upstream and originate from the input of warmer AW across the Greenland–Scotland Ridge. The pulses have also been followed in the Arctic Ocean, tracing several of the suggested circulation loops in the basins, giving timescales for the movements along the different loops. Model work has reproduced many of these events and their pathways around the Arctic Ocean. The first pulse has now left the Eurasian Basin and partly exited the Arctic Ocean through Fram Strait. In the Amundsen Basin it has been replaced by colder water, while the part that entered the Makarov Basin at the Mendeleev Ridge has circulated around the


222


Makarov Basin and is now found at the Makarov Basin side of the Lomonosov Ridge, practically removing the temperature front previously present along the ridge. The pulse has also entered the northern Canada Basin at the Chukchi Plateau. Its movements around and south of the Chukchi Plateau have taken comparably long, and it has been proposed that the AW here enters not mainly in the boundary current but directly into the basin as intrusive, doublediffusively driven layers. Older water, previously found in the southern Canada Basin, has shifted northward along the slope and is seen penetrating along the Alpha Ridge into the northern Canada Basin. The 1990 inflow event coincided with a strong, positive state of the North Atlantic Oscillation (NAO) and of the Arctic Oscillation (AO), which also affected the distribution of the runoff from the Siberian rivers. Instead of entering the Amundsen Basin from the Laptev Sea it continued eastward to the Makarov Basin and the northern Canada Basin. The upper Pacific water lens, which in the 1970s extended over the entire Canadian Basin to the Lomonosov Ridge, then contracted to the Canada Basin. Similar scenarios have been reproduced in model studies. The shifting of the Pacific/Atlantic surface front as well as the river water (shelf water) front counterclockwise toward the Canada Basin elevated the effects of the warmer AW. In the Amundsen Basin, and partly in the Makarov Basin, it approached closer to the sea surface, into the levels previously occupied by the halocline, thus magnifying the increase of both temperature and salinity at these levels. The shifting of the river water front also caused an increase in salinity of the surface layers of the Amundsen and Makarov Basins. The area with a deep winter mixed layer, previously confined to the Nansen Basin, expanded into the Amundsen basin, almost reaching the Lomonosov Ridge. The winter convection reached down to 120–130 m and actually caused a temperature decrease immediately above the Atlantic layer. The river water front has during the 2000s moved back into the Amundsen Basin, almost as far as the Gakkel Ridge, indicating that the shelf water outflow from the Laptev Sea again primarily enters the Amundsen Basin, recreating the PML–halocline structure in that basin. The inflow in the Barents Sea branch also appears to vary. The intermediate salinity minimum in the Eurasian Basin has become more pronounced and less saline uPDW has crossed the Lomonosov Ridge in the boundary current along the continental slope and entered the Canadian Basin. This has made the uPDW characteristics in the Makarov Basin more similar to those in the Eurasian Basin, with a curved rather than

a straight yS curve below the temperature maximum. A question is, if these anomalies in the Canadian Basin will stay long enough to be removed by shelfslope convection and interior mixing processes, recreating the older, smooth, yS characteristics, or if this inflow will create entirely new yS structures? If so, does this mean that the gyre circulation has changed and the communication between the Eurasian and Canadian Basins has increased, leading to a more rapid ventilation of the Canadian Basin? Historical data from the Russian archives indicate that sudden changes have occurred before, and the situation now observed may not be unique. The use of CTDs instead of Nansen bottles also reveals structures in the water column previously not resolved. The circulation pattern in the Arctic Ocean responds to long periodic (decadal) variations in the atmospheric circulation, the NAO and the AO. The positive AO state increases the inflow of AW and weakens the Beaufort gyre, while the negative state leads to a well-developed Beaufort gyre and a smaller inflow of AW. The negative state then retains the fresh water and the ice, while the positive state acts to reduce the storage of ice and low-salinity upper waters, which, together with a larger inflow of AW, increases the mean salinity of the Arctic Ocean waters. Perhaps the most prominent change observed in the Arctic Ocean has been the retreat of the ice cover. A reduction of the minimum ice extent of 1.5–2 106 km2 (420%) between 1979 and 2005 has been observed. Comparison between submarine observations of ice thicknesses 20 years apart indicates a thinning of the ice, from 3.1 to 1.8 m. However, the magnitude of this change has been contested. Thickness observations at the same position years apart might not be relevant, since the distribution of the ice will depend upon the forcing of the ice field, which will vary between the different years. Nevertheless, even disregarding changes in ice thickness, the ice cover has become significantly reduced. A thinner and less extensive ice cover indicates a loss of ice storage, which could either be due to reduced formation of ice, or to an increased export of ice out of the Arctic Ocean. The amount of freshwater released by the reduction of the ice cover could have contributed to perhaps the largest fraction of the freshening of the Nordic Seas and to the subpolar gyre reported in recent years, but this does not give any information about where the phase change occurred, in the Arctic Ocean or south of Fram Strait. The time series of the ice export are not long enough to provide an answer and the knowledge of the liquid freshwater export is uncertain, if not completely absent. The presence of the halocline between the PML and the Atlantic layer makes the stirring in the PML



entrain cold water from below, and the ice cover is isolated from the heat stored in the Atlantic layer. The situation in the mid-1990s with no river (shelf) water present in the Amundsen Basin could lead to the entrainment of warm AW into the uppermost layer and thus bring heat to the sea/ice surface. This sensible heat rather than latent heat of ice formation could then be supplied to the atmosphere and less ice would form in winter. The heat might also directly melt the ice, but if the stirring of the mixed layer, driving the entrainment of heat from below, is due to haline convection, a change from freezing to melting would stop the convection, and the entrainment as well as the heat transport from below would cease.

Transports The last 10 years have seen large programs focusing on measuring the transports through the key passages of the Arctic Ocean. Much of this activity has been coordinated by the ASOF (Arctic and Subarctic Ocean Fluxes) program but other projects have also been involved. The observations of the Bering Strait inflow have largely confirmed the transport estimates proposed by Russian researchers 50 years ago. A mean inflow of 0.8 Sv of low-salinity (Sr32) water takes place. The seasonal variations are large; the inflow is 1.2 Sv in summer and 0.4 Sv in winter. The recent observations have shown that the freshwater transport through Bering Strait might be substantially (20%) higher than previously estimated, making it twothirds of the river runoff. This reevaluation is due to the inclusion of the transport in the low-salinity Alaskan Coastal Current. The transports through the Canadian Arctic Archipelago, notoriously difficult to measure due to the remoteness, the severe climate, and the proximity to the magnetic north pole that makes direction determinations extremely difficult, have in recent years been measured in the Hell Gate and Cardigan Strait (the Jones Sound) and in Barrow Strait (Lancaster Sound). This gives observed transports through two of the three main passages through the Archipelago. Transport measurements have also been made in Nares Strait, but year-long transport estimates from Nares Strait are not yet available. The fluxes through the narrow Hell Gate are directed out of the Arctic Ocean, barotropic, and almost constant, while in the neighboring Cardigan Strait the flow is weaker and reversals are observed. The combined average transport is estimated to 0.3 Sv. The transports in the wider Barrow Strait are largely barotropic on the southern side of the channel

223

and directed eastward, toward Lancaster Sound and Baffin Bay. A weak, baroclinic westward flow is observed on the northern side, indicating transport of runoff and penetration of water from Baffin Bay and Lancaster Sound. The flow is highly variable with an estimated mean transport of 0.75 Sv from the Arctic Ocean to Baffin Bay. Models indicate that the Barrow Strait (Lancaster Sound) might contribute one-third to one-half of the total outflow through the Canadian Arctic Archipelago. Should this be correct the total transport would be 1.5–2 Sv. It is low-salinity, primarily Pacific, water that passes through the archipelago, but the bottom water in Baffin Bay, 0.5 1C and 34.5, likely derives from the Arctic Ocean through Nares Strait and would then be supplied by lower halocline waters. The transport of the AW to the Barents Sea between Bjørnøya and Norway has been measured continuously for 10 years. The mean net transport is into the Barents Sea, but there is a return flow of transformed, colder, and denser water. The mean net transport of AW to the Barents Sea has been estimated to 1.5 Sv. However, short periodic variations are large and in spring, due to changing wind conditions, a whole month of small net inflow, occasionally even outflow, has been observed. There are also indications of variability on longer timescales, 3–4 years, but no trend has been detected. The AW is warm and saline at the entrance, but it loses much of its heat during transit and does not contribute heat to the central Arctic Ocean. To the inflow of AW should be added a transport of 0.7 Sv of less saline (34.4) water from the Norwegian Coastal Current, increasing the net inflow to B2.2 Sv. Fram Strait, which also has been monitored regularly since 1997, has a two-directional flow, and not only polar surface water (PSW), comprising the PML and halocline waters, and AW but also intermediate and deep water masses pass through the strait. The flow is largely barotropic and highly variable in space and time. Eddies are present, both barotropic and baroclinic, which complicates the transport estimates. Most of the steady flow occurs in the northward-flowing West Spitsbergen Current to the east and the southward-flowing East Greenland Current to the west. The inflow comprises warm AW and colder, less-saline AIW and NDW, while the outflow carries sea ice, low-salinity PSW, cool Arctic Atlantic Water (AAW), uPDW with temperature between 0 and 0.5 1C and CBDW, seen as a salinity maximum at 1700 m, and the colder, but also saline EBDW close to the bottom. The observed total northward and southward transports are large, 10–15 Sv. The mean net transport is much smaller and lies between 1.5 and 2.5 Sv


224


southward, out of the Arctic Ocean. A large recirculation is present in the strait. This appears partly associated with the barotropic eddies that drift westward along the sill of the strait and carry water from the West Spitsbergen Current to the East Greenland Current. The transports through Fram Strait of largest climate importance are the export of ice and the heat carried by the AW. There is a large interannual variability but no trend has been found. The estimated mean ice export is B0.9 Sv, while the outflow of PSW is B1 Sv. The inflow of AW and heat also varies strongly from year to year but larger transports and higher temperatures have been observed in the last years, giving a flux B50 TW. The average heat transport since 1997 is 35–40 TW. This is estimated relative to an assumed mean temperature of 0.1 1C in the Arctic Ocean. However, since the mass budget is not closed, the heat transport will depend upon the choice of reference temperature. The long-term mean inflow of AW, 42 1C, is c. 3 Sv. Whether an increased heat transport through Fram Strait, connected with warmer AW and perhaps a stronger flow, also leads to more heat being available for the Arctic, for ice melt and for the atmosphere, is not yet clear. The fact that pulses of warm AW can be traced around the Arctic Ocean could imply that most of the heat is not lost but only stored and will eventually leave the Arctic Ocean through Fram Strait. The heat advected into the Arctic Ocean through Bering Strait is located closer to the sea surface and could have a larger impact on the heat flux to the ice cover from below and on the thickness of the ice cover. The large retreat in ice extent reported in the last years has mostly occurred in the southern Canada Basin, close to the Chukchi Sea.

and in the Labrador Sea. However, the fresh water largely remains in the East Greenland Current and mostly bypasses the Greenland Sea and Iceland Sea gyres. In recent years, ice has not been formed in the central Greenland Sea and the Greenland Sea has been dominated by thermal convection. The AAW and the uPDW contribute to the Denmark Strait overflow, while the CBDW and the EBDW mainly enter the Greenland Sea, where they supply the mid-depth (1800 m) temperature maximum and the deep salinity maximum. In recent years, the local convection in the Greenland Sea has not penetrated through the temperature maximum and only less-dense AIW has been formed. The production of AIW has lead to a more direct contribution of the Greenland Sea to the AMOC, especially to the East Greenland Current and the Denmark Strait overflow. The denser Arctic Ocean deep water masses, previously continuing in the East Greenland Current to Denmark Strait are now entering the Greenland Sea and gradually replacing the old, colder, and less saline Greenland Sea deep and bottom waters, making the Greenland Sea water column more ‘Arctic’ in character.

See also Bottom Water Formation. Double-Diffusive Convection. Heat Transport and Climate. Intrusions. Meddies and Sub-Surface Eddies. Non-Rotating Gravity Currents. North Atlantic Oscillation (NAO). Ocean Circulation. Ocean Circulation: Meridional Overturning Circulation. Open Ocean Convection. Overflows and Cascades. Polynyas. Rotating Gravity Currents. Sea Ice: Overview. Upper Ocean Heat and Freshwater Budgets. Upper Ocean Mixing Processes. Water Types and Water Masses.

Significance for Climate The influence of the Arctic Ocean on the circulation at lower latitudes is mainly through the export of freshwater as ice and as low-salinity PSW, and through the export of dense intermediate and deep waters that contribute to the Greenland–Scotland overflow and to the North Atlantic Deep Water, enforcing the thermohaline circulation and the Atlantic meridional overturning circulation (AMOC). Of the around 6-Sv overflow water supplied to the North Atlantic Deep Water by the Arctic Mediterranean about 3 Sv have passed through the Arctic Ocean. The outflow of ice and less-dense surface water could increase the stability of the water column in the convection areas to the south, in the Nordic Seas,

Further Reading Bjo¨rk G, Jakobsson M, Rudels B, et al. (2007) Bathymetry and deep-water exchange across the central Lomonosov Ridge at 88–891 N. Deep-Sea Research I 54: 1197--1208 (doi:10.1016/j.dsr.2007.05.010). Bjo¨rk G, So¨derqvist J, Winsor P, Nikolopoulos A, and Steele M (2002) Return of the cold halocline to the Amundsen Basin of the Arctic Ocean: Implications for the sea ice mass balance. Geophysical Research Letters 29(11): 1513 (doi:10.1029/2001GL014157). Coachman LK and Aagaard K (1974) Physical oceanography of the Arctic and sub-Arctic seas. In: Herman Y (ed.) Marine Geology and Oceanography of the Arctic Ocean, pp. 1--72. New York: Springer.



Dickson B, Meincke J, and Rhines P (eds.) (2008) Arctic– Subarctic Ocean Fluxes: Defining the Role of the Northern Seas in Climate. New York: Springer. Dickson RR, Rudels B, Dye S, Karcher M, Meincke J, and Yashayaev I (2007) Current estimates of freshwater Arctic and subarctic seas. Progress in Oceanography 73: 210--230 (doi:10.1016/j.pocean.2006. 12.003). Fahrbach E, Meincke J, Østerhus S, et al. (2001) Direct measurements of volume transports through Fram Strait. Polar Research 20: 217--224. Jakobsson M, Cherkis NZ, Woodward J, Macnab R, and Coakley B (2000) New grid of Arctic bathymetry aids scientists and mapmakers. EOS, Transactions of American Geophysical Union 81(9): 89--96. Jakobsson M, Macnab R, Mayer L, et al. (2008) An improved bathymetric portrayal of the Arctic Ocean: Implications for ocean modelling and geological, geophysical and oceanographic analyses. Geophysical Research Letters 35: L07602 (doi:10.1029/2008 GL0335220). Johannesen OM, Muench RD, and Overland JE (eds.) (1994) AGU Geophysical Monographs 85: The Polar Oceans and Their Role in Shaping the Global Environment. Washington, DC: American Geophysical Union. Jones EP, Rudels B, and Anderson LG (1995) Deep water of the Arctic Ocean: Origin and circulation. Deep-Sea Research 42: 737--760. Leppa¨ranta M (ed.) (1998) Physics of Ice-Covered Seas. Helsinki: Helsinki University Press. Lewis EL, Jones EP, Lemke P, Prowse T, and Wadhams P (eds.) (2000) The Freshwater Budget of the Arctic Ocean. Dordrecht: Kluwer Academic Publishers. Merryfield WJ (2002) Intrusions in double-diffusively stable Arctic waters: Evidence for differential mixing? Journal of Physical Oceanography 32: 1452--1459. Nansen F (1902) Oceanography of the North Polar Basin. Scientific Results III(9): The Norwegian North Polar Expedition 1893–96. Oslo: Jacob Dybwad. Peterson BJ, McClelland J, Curry R, Holmes RM, Walsh JE, and Aagaard K (2006) Trajectory shifts in the Arctic and subarctic freshwater cycle. Science 313: 1061--1066. Prinsenberg SJ and Hamilton J (2005) Monitoring the volume, freshwater and heat fluxes passing through Lancaster Sound in the Canadian Arctic Archipelago. Atmosphere-Ocean 43: 1--22. Quadfasel D, Sy A, Wells D, and Tunik A (1991) Warming of the Arctic. Nature 350: 385. Rudels B, Jones EP, Anderson LG, and Kattner G (1994) On the intermediate depth waters of the Arctic Ocean. In: Johannesen OM, Muench RD, and Overland JE (eds.) AGU Geophysical Monographs 85: The Polar Oceans and Their Role in Shaping the Global Environment, pp. 33--46. Washington, DC: American Geophysical Union. Rudels B, Jones EP, Schauer U, and Eriksson P (2004) Atlantic sources of the Arctic Ocean surface and halocline waters. Polar Research 23: 181--208.

225

Rudels B, Muench RD, Gunn J, Schauer U, and Friedrich HJ (2000) Evolution of the Arctic Ocean boundary current north of the Siberian shelves. Journal of Marine Systems 25: 77--99. Schauer U, Fahrbach E, Østerhus S, and Rohardt G (2004) Arctic warming through Fram Strait: Oceanic heat transports from 3 years of measurements. Journal of Geophysical Research 109: C06026 (doi:10.1029/2003 JC001823). Serreze MC, Barrett A, Slater AJ, et al. (2006) The largescale freshwater cycle in the Arctic. Journal of Geophysical Research 111: C11010 (doi:10.1029/2005 JC003424). Smith WO, Jr. (ed.) (1990) Polar Oceanography, Part A: Physical Sciences. San Diego: Academic Press. Smith WO, Jr. and Grebmeier JM (eds.) (1995) Coastal and Estuarine Studies: Arctic Oceanography, Marginal Ice Zones and Continental Shelves. Washington, DC: American Geophysical Union. Steele M and Boyd T (1998) Retreat of the cold halocline layer in the Arctic Ocean. Journal of Geophysical Research 100: 881--994. Stein R and MacDonald RW (eds.) (2004) The Organic Carbon Cycle in the Arctic Ocean, 362pp. Berlin: Springer. Swift JH, Aagaard K, Timokhov L, and Nikiforov EG (2005) Long-term variability of Arctic Ocean waters: Evidence from a reanalysis of the EWG data set. Journal of Geophysical Research 110: C03012 (doi:10.1029/ 2004JC002312). Timmermanns M-L, Garrett C, and Carmack E (2003) The thermohaline structure and evolution of the deep water in the Canada Basin, Arctic Ocean. Deep-Sea Research I 50: 1305--1321 (doi:10.1016/S0967-0637(03)00125-0). Untersteiner N (ed.) (1986) The Geophysics of Sea Ice. New York: Plenum. Wadhams P, Gascard J-C, and Miller L (1990) Topical studies in oceanography: The European Subpolar Ocean Programme: ESOP. Deep-Sea Research II 46: 1011--1530. Walsh D and Carmack EC (2002) A note on evanescent behavior of Arctic thermohaline intrusions. Journal of Marine Research 60: 281--310. Wassmann P (ed.) (2006) Special Issue: Structure and Function of Contemporary Food Webs on Arctic Shelves: A Pan-Arctic Comparison. Progress in Oceanography 71(2–4): 123–477. Wheeler PA (1997) Topical studies in oceanography: 1994 Arctic Ocean section. Deep-Sea Research II 44: 1483--1758. Woodgate RA and Aagaard K (2005) Revising the Bering Strait freshwater flux into the Arctic Ocean. Geophysical Research Letters 32: L02602 (doi:1029/2004 GL021747). Woodgate RA, Aagaard K, Muench RD, et al. (2001) The Arctic boundary current along the Eurasian slope and the adjacent Lomonosov Ridge. Deep-Sea Research I 48: 1757--1792.


ARTIFICIAL REEFS W. Seaman and W. J. Lindberg, University of Florida, Gainesville, FL, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Artificial reefs are intentionally placed benthic structures built of natural or man-made materials, which are designed to protect, enhance, or restore components of marine ecosystems. Their ecological structure and function, vertical relief, and irregular surfaces vary according to location, construction, and degree to which they mimic natural habitats, such as coral reefs. For ages, humans have taken advantage of the behavior of some organisms to seek shelter at submerged objects by introducing structures into shallow waters, where biological communities could form and fishes could be harvested. Contemporary purposes for artificial reefs include increasing the efficiency of artisanal, commercial, and recreational fisheries, producing new biomass in fisheries and aquaculture, boosting underwater recreation and ecotourism opportunities, preserving and renewing coastal habitats and biodiversity, and advancing research. This article reviews the evolution, spread, and increased scale of such practices globally, discusses their scientific basis, and describes trends concerning artificial reef planning, evaluation, and their appropriate application in natural resource management.

Utilization of Reefs in the World’s Oceans The near-shore ocean ecosystems of all inhabited continents contain artificial reefs. Their scales and purposes vary widely, ranging, for example, from placement in local waters of small structures by individuals in artisanal fishing communities, to deploying complex systems of heavy modules in distant areas by organizations concerned with commercial seafood production, to pilot studies using experimentally based structures for restoration of seagrass, kelp, and coral ecosystems. The origins of artificial reef technology are traced to places as diverse as Japan and Greece. Modern deployment of reefs has the longest history and has been most widespread in Eastern Asia, Australia, Southern Europe, and North America. In recent years, the technology has been introduced or more extensively established in scores of nations worldwide, including in Central and South America and the Indo-Pacific basin. No organization maintains a database for artificial reef development globally, so that assessments of reef-related activity must be derived from the records of economic development, fishery, and environment organizations, as well as from scientific journal articles which serve as a proxy for gauging national efforts. For centuries structures of natural materials have been used to support artisanal and subsistence fishing in coastal communities, particularly in tropical areas (Figure 1). In India, traditionally tree branches have been weighted and sunken. Brush parks made of

Figure 1 In some areas of long-standing artisanal fisheries, structures such as the pesquero (left), a bundle of mangrove tree branches, are used for benthic structure, as in Cuba. More recently, small fabricated modules, such as this example from India (right), are used to complement or replace natural materials.

226


ARTIFICIAL REEFS

branches stuck into the substrates have been used in estuaries of several African countries, Sri Lanka, and Mexico. In the Caribbean basin, casitas of wood logs provide benthic shelters from which spiny lobsters (Panulirus argus) are harvested. Indigenous knowledge of local fisherfolk is important in sustainably managing these reef systems, even when in their initial stages they may function principally to aggregate larger mobile fishes, such as around bundles of mangrove branches, pesqueros, in Cuba. In recent decades, some artisanal fishing communities have deployed newer designs of artificial reefs, and at larger scales. In part, this has been in response to damage of habitat and fisheries by coastal land-use practices and more intense fishing practices, such as trawling. In India, for example, a national effort has deployed reef modules of steel plates, to increase the time fishing cooperatives actually devote to harvest as a means of promoting social and economic well-being. The coastal waters of Thailand received concrete modules as part of a plan to balance small- and large-scale fisheries, whereby areas of 14– 22 km2 received between 2400 and 3300 units. The most extensive deployment of artificial reefs, for any purpose, is in Japan. There, fishing enhancement is the goal of a national plan begun in 1952 and since greatly expanded. By 1989, 9% of the coastal shelf (o200 m depth) had been affected by reef development. Early emphasis was on engineering and construction of reefs to withstand rigorous high seas environmental conditions. Industrial manufacturers have used materials such as steel, fiberglass, and concrete in structures that are among the largest units in the world, attaining heights of 9 m, widths of 27 m, and volumes of 3600 m3 (Figure 2). Research at sea has been augmented by laboratory studies in which scale models of reefs have been analyzed for effects including deflection of

Figure 2 The largest reef modules have been fabricated and deployed in Japan.

227

bottom water currents to create nutrient upwellings for enhancing primary production. Observations of fish behavior led to a classification of the affinity of fishes to associate with physical structure. The Japanese aims have been to develop nursery, reproduction, and fishing grounds to supply seafood including abalone, clams, sea urchins, and finfishes. This partially was in response to the closure of distant waters to fishing by other nations. In contrast to Japan (and many other nations more recently involved), the United States, while also an early entrant in this field, built thousands of artificial reefs to enhance recreational fishing through numerous, independent small-scale efforts organized by local interest groups or governmental organizations at the state level. Until recently, materials of opportunity predominated, including heavier and more durable materials such as concrete rubble from bridge and building demolition sites. Designed structures are becoming more common, especially as reefs gain acceptance in habitat restoration. One state, Florida, has about half of the nation’s total number of reefs, and in one four-county area annual expenditures by nonresidents and visitors on fishing and diving at dozens of artificial reefs were US $1.7 billion (2001 values). Recreational fishing has also been the focus of widespread reef building in Australia, and lesser efforts in areas such as Northern Ireland and Brazil. Other regions more recently deploying artificial reefs to enhance or simply maintain fisheries to supply human foodstuffs include southern Europe and Southeast Asia. From inception, their approach has been to design, experimentally test, and document reefs. As early as the 1970s, 8 m3, 13 t concrete blocks (and other designs) were deployed in the Adriatic Sea of Italy, forming reefs of volumes up to 13 000 m3 and coverage up to 2.4 ha. They serve multiple, fishery-related objectives that are sought in many nations: shelter for juvenile and adult organisms; reproduction sites; capture nutrients in shellfish biomass; protect fish spawning and nursery areas from illegal trawling; and protect artisanal set fishing gear from illegal trawling damage. Higher catches and profits for small-scale fisheries were the result. This is especially true in Spain, where heavy concrete and steel rod reef designs have been used effectively to protect seagrass beds from illegal trawling (Figure 3). Artificial reefs are being used as part of a larger strategy to implement marine ranching programs. In Korea, for example, the decline of distant water fisheries led to a transition from capture-based to culturebased fisheries. From 1971 to 1999, artificial reef placements have affected 143 000 ha of submerged


228

ARTIFICIAL REEFS

Figure 3 Concrete reef modules with projecting steel beams such as railroad tracks are deployed in protected marine areas of Spain to discourage illegal trawling, so as to promote restoration of seagrass habitats and fish populations. Photograph Courtesy of Juan Goutayer.

coastal areas, out of the total 307 000 ha planned. At certain reef sites, hatchery-reared juvenile invertebrates and fishes have been released, and enhanced fishery yield reported, as part of long-term experiments. Artificial reefs are also used exclusively for aquaculture, such as in settings where excess nutrients stimulate primary production that is transferred into biomass of harvestable species, such as mussels (Mytilus galloprovincialis) and oysters (Ostrea edulis and Crassostrea gigas) in Italy. Restoration of marine ecosystems using artificial reefs has focused on plant and coral communities. On the Pacific coast of the United States, for example, a project to mitigate for electric powerplant impacts is developing a 61 ha artificial reef for kelp (Macrocystis pyrifera) colonization, survival, and growth. Along the coast of the Northwest Mediterranean Sea, artificial reefs are deployed in protected areas to preserve and promote colonization of seagrasses (Posidonia). In numerous tropical areas, artificial reefs have been used in repairs to damaged coral reefs (Figure 4), or to hasten replacement of dead or removed coral. In the Philippines, for example, hollow concrete cubes are deployed as sites for coral colonization, sometimes being located in marine reserves. Increasing popularity of artificial reefs to promote tourism is seen in the development of diving opportunities, both for people in submersible vehicles and for scuba divers. In places such as Mexico, the Bahamas, Monaco, and Hawaii, submarine operators provide trips to artificial reefs. On a larger scale, obsolete ships have been sunk to create recreational dive destinations in areas as diverse as British Columbia, Canada, Oman, New Zealand, Mauritius, and Israel. Purposes include generation of economic revenues in

Figure 4 This larval enclosure tent is for settlement of the coral, Montastraea faveolata, at the site of a ship grounding on Molasses Reef, Florida Keys National Marine Sanctuary. Coral are mass-spawned, larvae cultured through the planktonic phase, and then introduced into the tent. In the laboratory, larvae may settle onto rubble pieces which subsequently are cemented to the reef. Photograph by D. Paul Brown.

local communities and diversion of diver pressure from natural reefs. As strictly works of engineering, structures such as rock jetties for shore protection also serve as de facto reefs with biological communities. Submerged breakwaters have been used to create waves for recreational surfboarding.

Progress toward Scientific Understanding Advances in research on artificial reefs include methodologies, subjects and rigor of approaches. Methodological advances include improved underwater biological census techniques, and the application of remote sensing. Investigations have become more ecological process-oriented in explaining biological phenomena such as sheltering, reproduction, recruitment, and feeding, better able to explain physical and hydrodynamic behavior of reefs, and have led to specifying reef designs consistent with the ecology of organisms. Newer approaches include an expansion of hypothesis-driven and experimental research, scaling up of pilot designs into larger reefs, ecological modeling and forecasting, and interdisciplinary studies with various combinations of physical, social, life, and mathematical sciences. Comparative studies of natural and artificial habitats have advanced understanding of both systems. The earlier history of research on artificial reefs may be observed in the subjects, methods, and findings reported at eight international conferences held from


ARTIFICIAL REEFS

1974 to 2005. Subsequent to a peak number of peerreviewed articles contained in a proceedings from the 1991 conference (84), fewer resulted from the latest conferences (1999 (56) and 2005 (20)). Scientists studying artificial reefs increasingly are reporting in journals devoted to fish biology, fisheries, and marine ecology. Colonization of artificial reefs has the oldest and most extensive record of scientific inquiry. The appearance of large fishes at some reef sites almost immediately after placement of structure is the most visible form of colonization (and also a factor in one of the most controversial artificial reef issues, discussed below). However, microbes, plants, and invertebrates also colonize reefs, and patterns of succession occur that converge toward naturally occurring assemblages of species (Figure 5). Augmenting earlier studies describing diversity and abundance of artificial reef flora and fauna, later research addressed the influence of environmental variables upon reef ecology, interactions of individual species with the artificial reef structure and biota, and long-term life history studies of selected species. Comparison of patch reefs of stacked concrete blocks with translocated coral reefs in the Bahamas (Figure 6) determined that the reefs had similar fish species composition, while the natural reefs supported more individuals and species, owing to their greater structural complexity (i.e., variety of hole sizes) and associated forage base. Meanwhile, long-term studies of the spiny lobster experimentally established that artificial reef structures, with dark recessed crevices approximately one body-length deep and multiple entrances, indeed augmented recruitment in a situation of habitat limitation in the Florida Keys, USA. While these situations allowed observation by scuba diving, in deeper waters hydroacoustic and video-recording instruments have been used to document fish species, such as at petroleum production platforms in the North Sea and Gulf of Mexico. Prominent among the scientific issues regarding the efficacy of artificial reefs for fisheries production and management is the ‘attraction-production question’: Do artificial reefs merely attract fishes thereby improving fishing efficiency, or do they contribute to biological production so as to enhance fisheries stocks? While this popularized dichotomy is an oversimplification, and an expectation of one allencompassing answer is unrealistic, this question has had great heuristic value for the advance of reef science and responsible reef development. The origins of this issue derive from early observations and assumptions about artificial reefs. High catch rates and densities of fish at artificial reefs were popularly

229

Figure 5 Aspects of attachment surfaces for marine plants and shelter and feeding sites for invertebrates and fishes provided by artificial reefs. Drawing courtesy of S. Riggio.

Figure 6 Experimental modules such as the one pictured on the right from the Bahamas exemplify a global trend toward greater manipulation of artificial reefs as tools for testing of hypotheses concerning ecological processes in the ocean, and comparison with systems such as coral reefs, pictured left. Drawing courtesy of American Fisheries Society.

taken as proof that artificial reefs benefited fisheries stocks, based on assumptions that hard substrata were limiting and reef fish standing stocks were dependent on food webs that have hard substrata as their foundations. However, a few fisheries scientists challenged this reasoning in the 1980s. They noted that before heavy exploitation the existing natural habitat supported abundant reef fishes, presumably at or near carrying capacity. Fishing then reduced populations to some lower level, yet the amount of natural habitat remained the same, still capable of supporting higher population numbers. The fisheries scientists reasoned that with fish stocks substantially below carrying capacity, the amount of hard-bottom habitat could not be the factor limiting population size, so the addition of artificial reefs would not benefit fish stocks. Thus, observed high densities of


230

ARTIFICIAL REEFS

fish and high catch rates at artificial reefs were considered an artifact of fish behavioral preferences, which simply concentrated them at reef sites and intensified fishing mortality, what conservationists today call an ecological trap. The fisheries scientists’ formal argument was more thorough and complex than the popularized version, as they recognized a continuum from solely attraction, with little if any direct biological benefit, to potentially high levels of added biological production from each unit of additional reef structure. Species fit along that continuum according to their life history characteristics and use of reef habitat ecologically. Thus small, highly site-attached fishes (e.g., blennies and gobies) that derive all resource needs (i.e., food, shelter, and mates) directly from the occupied reef structure, and that complete all but the planktonic phase of their life cycle in one place, would have the greatest net production potential from artificial reef development. Conversely, large, transient reef-visiting fishes (e.g., jacks and mackerels) were expected to have the least production potential from artificial reef development. In between these extremes were the majority of exploited reef fishes for which one would expect some combination of attraction and production affected by reef ecological setting and overall stock abundance. (Implicit in the argument by the fisheries scientists was the assumption that fishing pressure would be intensely focused at the artificial reefs, which has more to do with the management of human activities than the intrinsic ecological characteristics of the reefs themselves.) Differing professional positions concerning attraction-production still remain, and the debate continues to stimulate research questions and artificial reef management issues. In part, this is because the complexity of reef ecological functions and the effects of spatially explicit fishing mortality are beyond simple answers to the seemingly dichotomous question that has been popularized among lay audiences. And in part, this is because researchers necessarily address this issue from their own disciplinary perspectives and study systems, generally not integrating the multiple levels of biological organization or spatial-temporal domains over which this question can be legitimately addressed. Pertinent studies can focus on life history stages of populations, communities, or ecosystems and encompass individual reefs, broad landscapes, or geographic ranges of fisheries stocks, while covering seasonal, interannual, or long-term time frames. Practical solutions to the fisheries management implications will require more specific attention to the desired species and assemblages, and the processes that sustain them.

Now a pivotal question is how the manipulation of habitat affects demographic rates and exploitation rates across spatial and temporal scales. An example of a way to integrate large sets of ecological data to analyze different scenarios of artificial reef development is through ecosystem simulations such as for Hong Kong where fishing levels in and out of protected areas have been predicted as part of a fishery recovery program.

Advances in Planning, Design, and Construction The increase of artificial reefs in the world’s coastal seas, as measured by the growing number, cost, size, intricacy, and footprint of structures deployed, has prompted an emphasis on their planning. This is to promote more efficient and cost-effective structures, enable work at larger scales and with more precision, satisfy regulatory requirements, reduce conflicts with other natural and human aspects of ecosystems, and minimize the prospects of unintended consequences from improperly constructed reefs. The scientific basis for reefs has been strengthened by research efforts in dozens of countries, coupled with the practical experiences of numerous individuals and groups that have worked independently to build, manage, and, increasingly, evaluate reefs. Independently in different geographic areas, common approaches and some published guidelines for developing artificial reefs have emerged. The earliest handbooks focused on physical and oceanographic conditions as first encountered by the Coastal Fishing Ground Enhancement and Development Project of Japan. (Translations of certain documents into English in the 1980s by the United States National Marine Fisheries Service disseminated information.) The United States produced a national plan in 1984 (in revision in 2006) which spurred promulgation of plans among the coastal states. In some nations, planning is done through a coordinated reef program run by a federal or provincial marine or fishery resources organization, as in Italy or China. Finally, where reef deployment is directed to selected localities or areas, but not nationwide, master plans or other guidelines for siting, materials, and other aspects have been produced, such as for the Aegean Sea of Turkey and northern Taiwan. With a legitimate intent assured, the initial component of reef planning is to frame measurable objectives, define expectations for success, and forecast interactions of that reef with the ecosystem. Subsequently, a valid design and site plan for the reef and procedures for its construction must be developed


ARTIFICIAL REEFS

and meet regulatory requirements. The context for management of the reef must be defined. For example, an artificial reef program in Hong Kong was based on an extensive consultation with community interests, leading to stakeholder support for regulatory practices in the local fisheries and protection of marine reserve areas containing some of the reefs. Finally, protocols for evaluation of reef performance and management of assessment data must be established, with communication of results to all interested parties. Construction is a three-phase process: fabrication, transportation, and placement. Representative costs of larger reefs include US $16 000 000 for a 56 ha reef in California, USA, and between US $38 and $57 million annually during 1995–99 for a national program in Korea. The ecological and economic impacts of poorly designed artificial reefs are exemplified by an effort in the United States – projected to cost US $5 000 000 – to remove 2 000 000 automobile tires that had been deployed in the 1970s, but which were drifting onto adjacent live coral reefs and beaches. The design phase of reef development has changed dramatically due to growing emphasis on making the structural attributes of intentional, fabricated reef materials conform and contribute to the biological life history requirements of organisms that are particularly desired as part of the reef assemblage. Reef design should be dictated by: (1) a set of measurable and justified objectives for the reef, (2) the ecology of species of concern associated with the reef, and (3) predicted and understandable environmental and socioeconomic consequences of introduction of the reef into the aquatic ecosystem. This requires a multidisciplinary approach using expertise of biologists, engineers, economists, planners, sociologists, and others. From a physical standpoint, key aspects of design and construction include stability of a reef site, durability of the reef configuration, and potential adverse impacts in the environment. Also, the use of physical processes to enhance reef performance is a desirable aspect. To build a quantitative understanding of the environment into which a reef is to be placed, large-scale oceanographic processes and local conditions must be determined by site surveys early in reef planning, including water circulation driven by tides, wind and baroclinic/density fields, locally generated wind waves, swells propagated from distance, sediment/substrate composition, distribution and transport, and depth. These factors are then coupled with the attributes of the reef material, such as weight, density, dimensions, and strength, in order to forecast reef physical performance.

231

From an ecological standpoint, abiotic and biotic influences considered in design include geographic location, type and quality of substrate surrounding the reef site, isolation, depth, currents, seasonality, temperature regime, salinity, turbidity, nutrients, and productivity. Substrate attributes affecting reef ecology include its composition and surface texture, shape, height, profile, surface area, volume and hole size, which taken together contribute to the structural complexity of the reef. Spacing of individual and groups of reef structure is important. A plethora of designs for artificial reefs exist (Figure 7). For a mitigation reef aiming to create new kelp beds off California, USA, biologists and engineers concluded that the most effective design should place boulders and concrete rubble in low-relief piles (o1 m height), at depths of 12–14.5 m on a sand layer of 30–50 cm overlaying hard substrate; success criteria include (1) support of four adult plants per 100 m2 and (2) invertebrate and fish populations similar to natural reefs. As part of a system of artificial reefs in Portugal, experiments quantified production of sessile invertebrates (upon which fishes could feed) according to location on different facets of cubic settlement structures placed on reefs. For fishes, meanwhile, one of five designs used in Korea is the box reef, a 3 3 3 m concrete cube (Figure 8), targeted to two species: small, dark spaces in the lower two-thirds of the reef are provided for rockfish (Sebastes schlegeli), while the upper third is more open to satisfy behavioral preferences of porgy (Pagrus major). Repeatedly, authorities cite complexity of structure as a primary factor in design. As large individual and sets of reefs are planned, more organizations are using pilot projects to determine physical, biological, economic, and even

Figure 7 Artificial reefs of concrete modules are used worldwide, with designs intended to meet ecological requirements of designated species and habitats.


232

ARTIFICIAL REEFS

Figure 8 The box reef design used in Korea resulted from a collaboration of engineers and biologists. Modules have been placed in research plots within Tongyong Marine Reserve for use in experiments with stocking juvenile fishes. Photograph by Kim Chang-gil.

political feasibility. The 42 000 t Loch Linne Artificial Reef in Scotland was started in 2001 as a platform for scientific investigation of the performance of different structures, ultimately to establish fisheries for target species, specifically lobster (Homarus gammarus). This effort also is representative of a trend toward increased predeployment research – in this case a 4-year study of seabed, water-column, and biological parameters – intended to enable better forecasting of reef impacts and measurement of results. A special case of working with reef materials concerns the deployment of obsolete ships and petroleum/gas production platforms, due to the need to handle, prepare, and place them in environmentally compatible ways. In Canada, for example, decommissioned naval vessels require extensive removal of electrical wiring and other components to eliminate release of pollutants into the sea, in conformance with strict federal rules. In the northern Gulf of Mexico (where over 4000 platforms provide a considerable area of hard surface for sessile organism attachment), eastern Pacific, North Sea, and Adriatic Sea offshore platforms either act as de facto reefs or in a limited number of cases are being toppled in place or transported to new locations to serve as dedicated reefs, which costs less than removal to land.

have led to more realistic expectations bolstered by two decades of scientific advances. National plans (e.g., Japan, Korea, and United States), regional programs (e.g., Hong Kong, Singapore, and Turkey), and large-scale pilot projects (e.g., Loch Linne, Scotland; San Onofre, USA) concerning artificial reefs are better defining their role in ecosystem and fishery management. International scientific bodies such as the North Pacific Marine Sciences Organization (PICES) have addressed the relevance of artificial reefs to core fisheries issues, including stock enhancement, fishing regulations, and conservation. Integrated coastal management in the Philippines, India, Spain, and elsewhere now includes artificial reefs in multifaceted responses to issues of habitat destruction, fishery decline, and socioeconomic development. Finally, private consultants and businesses have found markets for their services and products, with one patented design being deployed in over 40 countries, and ecological engineers have recognized reefs as constructed ecosystems for use in restoration ecology. The evaluation of reef performance is fostering an increased acceptance and utilization of artificial reef technology by resource management organizations. Quantitative evaluation is increasingly driven by agency concerns for demonstrating positive returns on their investments, and the increasing scale of artificial reef projects. Evaluations vary in intensity and complexity, ranging from descriptive studies of short duration (e.g., pre- and postdeployment) to extensive studies of ecological processes and dynamics, to the synthesis of complex databases through quantitative modeling. One response of the scientific community was through formation of the European Artificial Reef Research Network, which promulgated priorities and protocols for research across its membership. A significant trend is for artificial reef projects to be planned and evaluated in an adaptive management framework, in which expectations are more explicitly stated, the projects implemented and rigorously evaluated, and then adjustments made to the management practices based on findings from the evaluations. When applied consistently, this cycle will continue to evolve the application of artificial reef technologies toward an ever-increasing standard of practical effectiveness.

Integration of Reefs in Ecological and See also Human Systems Increasingly, artificial reef technology is being applied globally in fisheries and ecosystem management. Early disappointments and healthy skepticism

Coastal Zone Management. Cold-Water Coral Reefs. Coral Reef and Other Tropical Fisheries. Coral Reef Fishes. Coral Reefs. Fiordic Ecosystems. Fisheries: Multispecies Dynamics. Fisheries


ARTIFICIAL REEFS

Overview. Fishery Management, Human Dimension. Fishery Manipulation through Stock Enhancement or Restoration. Fishing Methods and Fishing Fleets. Large Marine Ecosystems. Mariculture Diseases and Health. Mariculture, Economic and Social Impacts. Mariculture of Aquarium Fishes. Mariculture of Mediterranean Species. Mariculture Overview. Ocean Gyre Ecosystems. Upwelling Ecosystems.

Further Reading American Fisheries Society (1997) Special Issue on Artificial Reef Management. Fisheries 22: 4--36. International Council for the Exploration of the Sea (2002) ICES Journal of Marine Science 59(supplement S15363).

233

Jensen AC, Collins KJ, and Lockwood APM (eds.) (2000) Artificial Reefs in European Seas. Dordrecht: Kluwer. Lindberg WJ, Frazer TK, Portier KM, et al. (2006) Density-dependent habitat selection and performance by a large mobile reef fish. Ecological Applications 16(2): 731--746. Love MS, Schroeder DM, and Nishimoto MM (2003) The ecological role of oil and gas production platforms and natural outcrops on fishes in southern and central California: A synthesis of information. OCS Study MMS 2003-032. Seattle. WA: US Department of the Interior, US Geological Survey, Biological Resources Division. Seaman W (ed.) (2000) Artificial Reef Evaluation. Boca Raton, FL: CRC Press. Svane I and Petersen JK (2001) On the problems of epibioses, fouling and artificial reefs: A review. Marine Ecology 22: 169--188.


ATLANTIC OCEAN EQUATORIAL CURRENTS S. G. Philander, Princeton University, Princeton, NJ, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 188–191, & 2001, Elsevier Ltd.

because they provide invaluable checks on the theories and models that explain and simulate oceanic currents. Those currents play a central role in the Earth’s climate, by influencing sea surface temperature patterns for example.

Introduction

Time-averaged Currents

The circulations of the tropical Atlantic and Pacific Oceans have much in common because similar trade winds, with similar seasonal fluctuations, prevail over both oceans. The salient features of these circulations are alternating bands of eastward- and westward-flowing currents in the surface layers (see Figure 1). Fluctuations of the currents in the two oceans have similarities not only on seasonal but even on interannual timescales; the Atlantic has a phenomenon that is the counterpart of El Ninõ in the Pacific. The two oceans also have significant differences. The Atlantic, but not the Pacific, has a net transport of heat from the southern into the northern hemisphere, mainly because of an intense, crossequatorial coastal current in the Atlantic, the North Brazil Current. The similarities and differences between the tropical Atlantic and Pacific (and also the Indian Ocean) are of enormous interest to modelers

Although the trade winds that prevail over the tropical Atlantic Ocean have a westward component, the currents driven by those winds include the eastward North Equatorial Countercurrent, between the latitudes 31 and 101N approximately. Sverdrup, in one of the early triumphs of dynamical oceanography, first pointed out that this current is attributable to the curl of the wind. Flanking this eastward current are westward currents to its north, the North Equatorial Current, and to its south, the South Equatorial Current. The latter current is particularly intense at the equator, where it can attain speeds in excess of 1 m s 1. Figure 1, a schematic map of the various currents, actually depicts conditions between July and September when the south-east trades are particularly intense and penetrate into the northern hemisphere. Centered on the equator, and below the westward surface flow, is an intense eastward jet known as the

70˚W 20˚N

60˚

Caribbean Current Guyana Current

40˚

50˚

Latitude

20˚

10˚

0˚

10˚E

North Equatorial Current

10˚

0˚

Longitude 30˚

North Equatorial Countercurrent

No Br r th Cu azil rre nt

Guinea Current

South Equatorial Current South Equatorial Current

10˚ Brazil Current

South Equatorial Current

20˚S

Figure 1 Schematic map showing the major surface currents of the tropical Atlantic Ocean between July and September when the North Equatorial Countercurrent (NECC) flows eastward into the Guinea Current in the Gulf of Guinea. From January to May the North Equatorial Countercurrent disappears and the surface flow is westward everywhere in the western tropical Atlantic.

234


235

ATLANTIC OCEAN EQUATORIAL CURRENTS

Seasonal Variations of the Currents The seasonal variations of the winds are associated with the north–south movements of the Intertropical Convergence Zone (ITCZ), the band of cloudiness

and heavy rains where the south-east and north-east trades meet. The south-east trades are most intense and penetrate into the northern hemisphere during the northern summer when the ITCZ is between 101 and 151N. During those months the surface currents are particularly strong. The North Brazil Current, after crossing the equator, veers sharply eastward to feed the North Equatorial Countercurrent. The Equatorial Undercurrent is also strongest during this season when the east–west slope of the equatorial thermocline is at a maximum. During the summer of the southern hemisphere, the zone where the north-east and south-east trades meet (the ITCZ) shifts equatorward so that the winds are relaxed along the equator. The North Brazil Current no longer veers offshore after crossing the equator, but continues to flow along the coast into the Gulf of Mexico. It is fed by surface flow that is westward at practically all latitudes in the tropics because, during this season, the eastward North Equatorial Countercurrent disappears from the surface layers, as is evident in Figure 2. At this time, the northward heat transport across 101N is huge – on the order of a peta-watt; during the northern summer it is practically zero.

15˚N NECC _ 10

_ 10 0

10˚

20

Latitude

34

_ 15

5˚

_ 10

0˚

_ 60

_ 30

_5

0 _ 3 0 _ 2

Equatorial Undercurrent which amounts to a narrow ribbon that precisely marks the location of the equator. The undercurrent attains speeds on the order of 1 m s 1 has a half-width of approximately 100 km; its core, in the thermocline, is at a depth of approximately 100 m in the west, and shoals towards the east. The current exists because the westward trade winds, in addition to driving divergent westward surface flow (upwelling is most intense at the equator), also maintain an eastward pressure force by piling up the warm surface waters in the western side of the ocean basin. That pressure force is associated with equatorward flow in the thermocline because of the Coriolis force. At the equator, where the Coriolis force vanishes, the pressure force is the source of momentum for the eastward Equatorial Undercurrent which, in a downstream direction, continually loses water because of intense equatorial upwelling which sustains the divergent, poleward Ekman flow in the surface layers. Along the African coast, cold equatorward coastal currents, the Canary Current off north-west Africa, and the Benguela Current off south-west Africa, are driven by the components of the winds parallel to the coast. These currents, which are associated with intense coastal upwelling and low sea surface temperatures, feed the westward North and South Equatorial Currents respectively. Along the coast of South America, the most prominent current is the North Brazil Current, which carries very warm water from about 51N across the equator. Some of that water feeds the Equatorial Undercurrent, but much of it continues into the northern hemisphere. Further south along the coast of Brazil, the flow is southward. The net north–south circulation associated with the various currents is a northward flow of warm surface waters, and a southward return flow of cold water at depth, resulting in a transport of heat from the southern into the northern Atlantic. The southward flow below the thermocline is part of the global thermohaline circulation, which involves the sinking of cold, saline waters in the northern Atlantic. The absence of such formation of deep water in the northern Pacific – that ocean is less saline than the northern Atlantic – is part of the reason why there is a northward transport of heat across the equator in the Atlantic but not the Pacific.

_ 40

_ 10_ 20

_ 39

_ 30

5˚ _ 20 _ 10 10˚S J

F

M

A

M

J

J A Month

S

O

N

D

J

Figure 2 The seasonal disappearance of the North Equatorial Countercurrent from the western tropical Atlantic. The eastward velocity in cm s 1 (negative values correspond to westward flow) is shown as a function of latitude and month, starting in January. The data, which have been averaged over a band of longitudes in the western equatorial Atlantic from 231W to 331W, are from shipdrift records.


236


The upwelling along the west African coast, and the coastal currents too, are subject to large seasonal fluctuations in response to the variations in the local winds. Thus upwelling is most intense off south-western Africa, and surface temperatures there are at a minimum, during the late northern summer when the local alongshore winds are most intense. Off northwestern Africa the season for such conditions is the late northern winter. The northern coast of the Gulf of Guinea (along 51N approximately) also has seasonal upwelling, with lowest temperatures during the northern summer, even though the local winds along that coast have almost no seasonal cycle. In that region, changes in the depth of the thermocline (which separates warm surface waters from the cold water at depth) depend on winds everywhere in the equatorial Atlantic, even the winds off Brazil which are most intense during the northern summer when they cause a shoaling of the thermocline throughout the Gulf of Guinea. If the winds over the ocean were suddenly to stop blowing, how long would it be before the currents in Figure 1 disappear? The answer to this question (which is the same as asking how long it would take for the currents to be generated from a state of rest) is of central importance in climate studies because, associated with the currents, are sea surface temperature patterns that profoundly affect climate. (From a strictly atmospheric perspective, the cause of El Ninõ is a change in the surface temperature pattern of the tropical Pacific.) The Indian Ocean is ideal for studying these matters because there the abrupt onset of the south-west monsoons in May quickly generates the intense Somali Current along the eastern coast of Africa. Theoretical studies indicate that the generation of such currents, and more generally the adjustment of the ocean to a change in the winds, depend critically on waves (known as Rossby waves) that propagate across the ocean basin along the thermocline. The speed of those waves increases with decreasing latitude, reaching a maximum at the equator, which serves as a guide for the fastest waves – there they travel westward at about 50 cm s 1. The equator is also a guide for a very rapid eastward traveling wave, a Kelvin wave with a speed on the order of 150 cm s 1. The Somali Current near the equator can therefore be generated far more rapidly than can the Gulf Stream in mid-latitudes. The time it takes for the ocean to adjust (for the currents to be generated) depends not only on the speed of certain oceanic waves, but also on the width of the ocean basin. Hence it takes longer to generate the Kuroshio Current in the very wide Pacific, than the Gulf Stream in the smaller Atlantic. If the winds change gradually rather than abruptly, then the timescale of the gradual changes relative to

the time it takes the ocean to adjust determines the nature of the oceanic response. Thus winds that fluctuate on a timescale much longer than the adjustment time of the ocean will force an equilibrium response in which the ocean, at any given time, is in equilibrium with the winds at that time. (The currents and winds fluctuate essentially in phase.) From results such as these it can be inferred that the seasonally varying trade winds over the tropical Atlantic and Pacific Oceans should force an equilibrium response near the equator in the case of the small ocean basin, the Atlantic, but not in the case of the much larger Pacific. The measurements confirm this theoretical result: the seasonal variations of the currents and of the thermocline slope are in phase with the variations of the winds in the equatorial Atlantic, but not in the equatorial Pacific.

Interannual Variations Given the similarities between the climates of the tropical Atlantic and Pacific – arid, cool conditions on the eastern sides, along the shores of Peru and south-western Africa, and warm moist conditions on the western sides – it should come as no surprise that the climate fluctuation known as El Ninõ has an Atlantic counterpart. As in the Pacific, such events involve a relaxation of the trades so that the warm waters that are usually confined to the western side of the basin flow eastward, causing a rise in sea surface temperatures off the south-west African coast where rainfall can increase significantly. To attribute this phenomenon to a relaxation of the trades is of course an oceanographic perspective. From a meteorological point of view, the warming of the eastern tropical Atlantic is the reason for the weakening of the winds and for several other changes in atmospheric conditions. This circular argument – changes in sea surface temperature are both the cause and consequence of changes in the winds – implies that interactions between the ocean and atmosphere are at the heart of the matter. Those interactions give rise to a variety of natural modes of oscillation which, in the Pacific, appear to be neutrally stable so that random atmospheric disturbances are able to sustain a continual oscillation, the Southern Oscillation, with a distinctive timescale on the order of 4 years. In the Atlantic the possible natural modes appear to be strongly damped and hence are far more sporadic than in the Pacific; there is no distinctive timescale for interannual fluctuations in the Atlantic. The main reason for this difference is the modest dimensions of the Atlantic relative to those of the Pacific. Some of the natural modes attributable to



ocean–atmosphere interactions depend on the delayed response of the ocean to changes in the winds. If that delay is small, that is the case in an ocean basin of modest size – then the natural modes tend to be damped. Another factor that can inhibit interannual fluctuations is a particularly strong seasonal cycle. That cycle has a larger amplitude in the equatorial Atlantic than Pacific, because the influence of continents on the seasonal changes in the winds can exceed those of ocean–atmosphere interactions in a basin of small dimensions. For a damped mode of oscillation to appear, a suitable perturbation is necessary. The occurrence of El Ninõ in the Pacific provides such a perturbation in the Atlantic by causing an intensification of the trade winds, and unusually low surface temperatures, in the Atlantic. (This is the impact of the presence of deep atmospheric convection over the eastern tropical Pacific during El Ninõ.) Apparently El Ninõ in the Pacific can amount to a preconditioning of the Atlantic because, on several occasions, El Ninõ in the Pacific was followed a year later by a similar phenomenon in the Atlantic. The amplitude of El Ninõ is generally much larger in the Pacific than Atlantic – the reason why the Pacific but not the Atlantic phenomenon is capable of a global impact. El Ninõ, in the Atlantic and Pacific, has a structure that is essentially symmetrical about the equator. The Atlantic has an additional climate fluctuation that is anti-symmetrical relative to the equator, with sea surface temperatures that are high on one side of that line, low on the other side. The cross-equatorial winds then blow towards the warm side with exceptional intensity. If the higher temperatures are to the north, then the zonal band of heavy rains, the ITCZ, persists in a northerly position, bringing drought to north-eastern Brazil, and good rains to the Sahel, the region to the south of the Sahara desert in west Africa. The reverse happens when the ocean temperatures are high south of the equator, cool to the north.

Stability of the Currents During the northern summer, the currents in the western equatorial Atlantic are so intense that they become unstable. One important factor is the enormous latitudinal shear between the eastward North Equatorial Countercurrent and the adjacent westward South Equatorial Current. The instabilities

237

result in meanders that drift westward at a speed near 50 cm s 1, that have a wavelength on the order of a 1000 km, and a period of approximately 1 month. The unstable conditions are confined to the western equatorial region where there is room for two or three waves at most – they sometimes appear in satellite photographs of sea surface temperature. The waves persist for a few months at most so that approximately three oscillations appear during the summer. The counterparts of these unstable waves in the eastern equatorial Pacific have a shorter period (close to 3 weeks) than in the Atlantic, cover a much larger region, and persist far longer. In the Pacific it is possible to observe very long wave trains – they can extend from the Galapagos Islands in the east to the dateline – that persist for many months.

See also Brazil and Falklands (Malvinas) Currents. Current Systems in the Atlantic Ocean. Coastal Trapped Waves. El Ninõ Southern Oscillation (ENSO). El Ninõ Southern Oscillation (ENSO) Models. Florida Current, Gulf Stream and Labrador Current. Rossby Waves. Satellite Remote Sensing of Sea Surface Temperatures.

Further Reading The Journal of Geophysical Research Volume 103 (1998) is devoted to a series of excellent and detailed review articles on tropical ocean–atmosphere interactions, including an article on oceanic currents. Carton J and Huang B (1994) Warm events in the tropical Atlantic. Journal of Physical Oceanography 24: 888--903. Chang P, Ji L, and Li H (1997) A decadal climate variation in the tropical Atlantic ocean from thermodynamic air– sea interaction. Nature 385: 516--518. Merle J, Fieux M, and Hisard P (1980) Annual signal and interannual anomalies of sea surface temperature in the eastern equatorial Atlantic. Deep Sea Research 26: 77--101. Philander SGH (1990) El Ninõ, La Ninã and the Southern Oscillation. New York: Academic Press. Richardson PL and Walsh DW (1986) Mapping climatological seasonal variations of surface currents in the tropical Atlantic using ship drifts. Journal of Geophysical Research 91: 10537--10550.


ATMOSPHERIC INPUT OF POLLUTANTS R. A. Duce, Texas A&M University, College Station, TX, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 192–201, & 2001, Elsevier Ltd.

Introduction For about a century oceanographers have tried to understand the budgets and processes associated with both natural and human-derived substances entering the ocean. Much of the early work focused on the most obvious inputs – those carried by rivers and streams. Later studies investigated sewage outfalls, dumping, and other direct input pathways for pollutants. Over the past decade or two, however, it has become apparent that the atmosphere is also not only a significant, but in some cases dominant, pathway by which both natural materials and contaminants are transported from the continents to both the coastal and open oceans. These substances include mineral dust and plant residues, metals, nitrogen compounds from combustion processes and fertilizers, and pesticides and a wide range of other synthetic organic compounds from industrial and domestic sources. Some of these substances carried into the ocean by the atmosphere, such as lead and some chlorinated hydrocarbons, are potentially harmful to marine biological systems. Other substances, such as nitrogen compounds, phosphorus, and iron, are nutrients and may enhance marine productivity. For some substances, such as aluminum and some rare earth elements, the atmospheric input has an important impact on their natural chemical cycle in the sea. In subsequent sections there will be discussions of the input of specific chemicals via the atmosphere to estuarine and coastal waters. This will be followed by considerations of the atmospheric input to open ocean regions and its potential importance. The atmospheric estimates will be compared with the input via other pathways when possible. Note that there are still very large uncertainties in all of the fluxes presented, both those from the atmosphere and those from other sources. Unless otherwise indicated, it should be assumed that the atmospheric input rates have uncertainties ranging from a factor of 2 to 4, sometimes even larger.

238

Estimating Atmospheric Contaminant Deposition Contaminants present as gases in the atmosphere can exchange directly across the air/sea boundary or they may be scavenged by rain and snow. Pollutants present on particles (aerosols) may deposit on the ocean either by direct (dry) deposition or they may also be scavenged by precipitation. The removal of gases and/or particles by rain and snow is termed wet deposition. Direct Deposition of Gases

Actual measurement of the fluxes of gases to a water surface is possible for only a very few chemicals at the present time, although extensive research is underway in this area, and analytical capabilities for fast response measurements of some trace gases are becoming available. Modeling the flux of gaseous compounds to the sea surface or to rain droplets requires a knowledge of the Henry’s law constants and air/sea exchange coefficients as well as atmospheric and oceanic concentrations of the chemicals of interest. For many chemicals this information is not available. Discussions of the details of these processes of air/sea gas exchange can be found in other articles in this volume. Particle Dry Deposition

Reliable methods do not currently exist to measure directly the dry deposition of the full size range of aerosol particles to a water surface. Thus, dry deposition of aerosols is often estimated using the dry deposition velocity, vd. For dry deposition, the flux is then given by: Fd ¼ vd Ca

½1

where Fd is the dry deposition flux (e.g., in g m2 s1), vd is the dry deposition velocity (e.g., in m s1), and Ca is the concentration of the substance on the aerosol particles in the atmosphere (e.g., in g m3). In this formulation vd incorporates all the processes of dry deposition, including diffusion, impaction, and gravitational settling of the particles to a water surface. It is very difficult to parameterize accurately the dry deposition velocity since each of these processes is acting on a particle population, and they are each dependent upon a number of factors, including wind speed, particle size, relative humidity, etc. The following are dry deposition velocities that have been


ATMOSPHERIC INPUT OF POLLUTANTS

used in some studies of atmospheric deposition of particles to the ocean:

• • •

Submicrometer aerosol particles, 0.001 m s17 a factor of three Supermicrometer crustal particles not associated with sea salt, 0.01 m s17 a factor of three Giant sea-salt particles and materials carried by them, 0.03 m s17 a factor of two

Proper use of eqn [1] requires that information be available on the size distribution of the aerosol particles and the material present in them. Particle and Gas Wet Deposition

The direct measurement of contaminants in precipitation samples is certainly the best approach for determining wet deposition, but problems with rain sampling, contamination, and the natural variability of the concentration of trace substances in precipitation often make representative flux estimates difficult using this approach. Studies have shown that the concentration of a substance in rain is related to the concentration of that substance in the atmosphere. This relationship can be expressed in terms of a scavenging ratio, S: S ¼ Cr r Ca=g 1

½2

where Cr is the concentration of the substance in rain (e.g., in g kg1), r is the density of air (B1.2 kg m3), Ca/g is the aerosol or gas phase concentration in the atmosphere (e.g., in g m3), and S is dimensionless. Values of S for substances present in aerosol particles range from a few hundred to a few thousand, which roughly means that 1 g (or 1 ml) of rain scavenges D1 m3 of air. For aerosols, S is dependent upon such factors as particle size and chemical composition. For gases, S can vary over many orders of magnitude depending on the specific gas, its Henry’s law constant, and its gas/water exchange coefficient. For both aerosols and gases, S is also dependent upon the vertical concentration distribution and vertical extent of the precipitating cloud, so the use of scavenging ratios requires great care, and the results have significant uncertainties. However, if the concentration of an atmospheric substance and its scavenging ratio are known, the scavenging ratio approach can be used to estimate wet deposition fluxes as follows: Fr ¼ P Cr ¼ P S Ca=g r1

½3

where Fr is the wet deposition flux (e.g., in g m2 year1) and P is the precipitation rate (e.g., in m year1), with appropriate conversion factors to translate rainfall

239

depth to mass of water per unit area. Note that P S r1 is equivalent to a wet deposition velocity.

Atmospheric Deposition to Estuaries and the Coastal Ocean Metals

The atmospheric deposition of certain metals to coastal and estuarine regions has been studied more than that for any other chemicals. These metals are generally present on particles in the atmosphere. Chesapeake Bay is among the most thoroughly studied regions in North America in this regard. Table 1 provides a comparison of the atmospheric and riverine deposition of a number of metals to Chesapeake Bay. The atmospheric numbers represent a combination of wet plus dry deposition directly onto the Bay surface. Note that the atmospheric input ranges from as low as 1% of the total input for manganese to as high as 82% for aluminum. With the exception of Al and Fe, which are largely derived from natural weathering processes (e.g., mineral matter or soil), most of the input of the other metals is from human-derived sources. For metals with anthropogenic sources the atmosphere is most important for lead (32%). There have also been a number of investigations of the input of metals to the North Sea, Baltic Sea, and Mediterranean Sea. Some modeling studies of the North Sea considered not only the direct input pathway represented by the figures in Table 1, but also considered Baltic Sea inflow, Atlantic Ocean inflow and outflow, and exchange of metals with the Table 1 Estimates of the riverine and atmospheric input of some metals to Chesapeake Bay Metal

Riverine input (106 g year1)

Atmospheric input (106 g year1)

% Atmospheric input

Aluminum Iron Manganese Zinc Copper Nickel Lead Chromium Arsenic Cadmium

160 600 1300 50 59 100 15 15 5 2.6

700 400 13 18 3.5 4 7 1.5 0.8 0.4

81 40 1 26 6 4 32 10 14 13

Data reproduced with permission from Scudlark JR, Conko KM and Church TM (1994) Atmospheric wet desposition of trace elements to Chesapezke Bay: (CBAD) study year 1 results. Atmospheric Environment 28: 1487–1498.


240


sediments, as well as the atmospheric contribution to all of these inputs. Figure 1 shows schematically some modeling results for lead, copper, and cadmium. Note that for copper, atmospheric input is relatively unimportant in this larger context, while atmospheric input is somewhat more important for cadmium, and it is quite important for lead, being approximately equal to the inflow from the Atlantic Ocean, although still less than that entering the North Sea from dumping. As regards lead, note that approximately 20% of the inflow from the Atlantic to the North Sea is also derived from the atmosphere. This type of approach gives perhaps the most accurate and in-depth analysis of the importance of

atmospheric input relative to all other sources of a chemical in a water mass. Nitrogen Species

The input of nitrogen species from the atmosphere is of particular interest because nitrogen is a necessary nutrient for biological production and growth in the ocean. There has been an increasing number of studies of the atmospheric input of nitrogen to estuaries and the coastal ocean. Perhaps the area most intensively studied is once again Chesapeake Bay. Table 2 shows that approximately 40% of all the nitrogen contributed by human activity to Chesapeake Bay enters via

Direct atmospheric deposition 1172 (1172)

Direct atmospheric deposition 79 (79) Copper Rivers and direct discharge 2230 (15)

Atlantic outflow 6166 (102)

Dumping 1561 (0)

North Sea

Lead Rivers and direct discharge 1597 (288) Dumping 3015 (0)

Atlantic outflow 4122 (1167) North Sea

Baltic inflow 84 (17)


Atlantic inflow 1262 (252)

Atlantic inflow 3785 (38) Sedimentation 2543 (35)

Sedimentation 3248 (562)

Erosion 600 (0) Sediments


(Tons/ year)

Direct atmospheric deposition 20 (20)

(Tons/ year)

Direct atmospheric deposition 13.3 (13.3) Cadmium Rivers and direct discharge 115 (6)

Atlantic outflow 461 (49)

Dumping 72 (0)

North Sea

Sedimentation 39 (5)

Rivers and direct discharge 3.9 (3.5)

Atlantic outflow 15.4 (15.3)

Dumping 0.0 (0)

North Sea


Baltic inflow 0.2 (0.2)

Atlantic inflow 251 (25)

Atlantic inflow 5.0 (5.0)


Lindane

Sedimentation 7.0 (6.7)

(Tons/ year)

Erosion 0.0 (0) Sediments

(Tons/ year)

Figure 1 Input of copper, lead, cadmium, and lindane to the North Sea. Values in parentheses denote atmospheric contribution. For example, for copper the atmospheric contribution to rivers and direct discharges is 15 tons per year. (Figure reproduced with permission from Duce, 1998. Data adapted with permission from van den Hout, 1994.)



Table 2

241

Estimates of the input of nitrogen to Chesapeake Bay

Source

Total input (109g year1)

Areal input rate (g m2year1)

% of the total

Animal waste Fertilizers Point sources Atmospheric precipitation nitrate ammonium Total

5 48 33

0.4 4.2 2.9

3 34 24

35 19 140

3.1 1.7 12.3

25 14 100

Data reproduced with permission from Fisher D, Ceroso T, Mathew T and Oppenheimer M (1988) Polluted Coastal Waters: The Role of Acid Rain. New York: Environmental Defense Fund.

precipitation falling directly on the Bay or its watershed. These studies were different from most earlier studies because the atmospheric contributions were considered not only to be direct deposition on the water surface, but also to include that coming in via the atmosphere but falling on the watershed and then entering the Bay. Note from Table 2 that atmospheric input of nitrogen exceeded that from animal waste, fertilizers, and point sources. In the case of nitrate, about 23% falls directly on the Bay, with the remaining 77% falling on the watershed. These results suggest that studies that consider only the direct deposition on a water surface (e.g., the results shown in Table 1) may significantly underestimate the true contribution of atmospheric input. The total nitrogen fertilizer applied to croplands in the Chesapeake Bay region is B5.4 g m2 year1, while the atmospheric nitrate and ammonium nitrogen entering the Bay is B4.8 g m2 year1. Chesapeake Bay is almost as heavily fertilized from atmospheric nitrogen, largely anthropogenic, as the croplands are by fertilizer in that watershed!

Table 3

Results from studies investigating nitrogen input to some other estuarine and coastal regions are summarized in Table 3. In this table atmospheric sources for nitrogen are compared with all other sources, where possible. The atmospheric input ranges from 10% to almost 70% of the total. Note that some estimates compare only direct atmospheric deposition with all other sources and some include as part of the atmospheric input the portion of the deposition to the watershed that reaches the estuary or coast. Synthetic Organic Compounds

Concern is growing about the input of a wide range of synthetic organic compounds to the coastal ocean. To date there have been relatively few estimates of the atmospheric fluxes of synthetic organic compounds to the ocean, and these estimates have significant uncertainties. These compounds are often both persistent and toxic pollutants, and many have relatively high molecular weights. The calculation of the atmospheric input of these compounds to the

Estimates of the input of nitrogen to some coastal areasa

Region

Total atmospheric inputb (109g year1)

Total input all sources (109g year1)

% Atmospheric input

North Sea Western Mediterranean Sea Baltic Sea Chesapeake Bay New York Bight Long Island Sound Neuse River Estuary, NC

400c 400c 500 54 – 11 1.7

1500 577d B1200 140 – 49 7.5

27c 69c 42 39 13c 22 23

a

Data from several sources in the literature. Total from direct atmospheric deposition and runoff of atmospheric material from the watershed. c Direct atmospheric deposition to the water only. d Total from atmospheric and riverine input only. b


242


Table 4

Estimates of the input of synthetic organic compounds to the North Sea

Organic compound

Atmospheric input (106 g year1)

Input from other sources (106 g year1)

% Atmospheric input

PCB Lindane Polycyclic aromatic hydrocarbons Benzene Trichloroethene Trichloroethane Tetrachloroethene Carbon tetrachloride

40 36 80 400 300 90 100 6

3 3 90 500 80 60 10 40

93 92 47 44 80 94 91 13

Data reproduced with permission from Warmerhoven JP, Duiser JA, de Leu LT and Veldt C (1989) The Contribution of the Input from the Atmosphere to the Contamination of the North Sea and the Dutch Wadden Sea. Delft, The Netherlands: TNO Institute of Environmental Sciences.

coastal ocean is complicated by the fact that many of them are found primarily in the gas phase in the atmosphere, and most of the deposition is related to the wet and dry removal of that phase. The atmospheric residence times of most of these compounds are long compared with those of metals and nitrogen species. Thus the potential source regions for these compounds entering coastal waters can be distant and widely dispersed. Figure 1 shows the input of the pesticide lindane to the North Sea. Note that the atmospheric input of lindane dominates that from all other sources. Table 4 compares the atmospheric input to the North Sea with that of other transport paths for a number of other synthetic organic compounds. In almost every case atmospheric input dominates the other sources combined.

Atmospheric Deposition to the Open Ocean Studies of the atmospheric input of chemicals to the open ocean have also been increasing lately. For many substances a relatively small fraction of the material delivered to estuaries and the coastal zone by rivers and streams makes its way through the near shore environment to open ocean regions. Most of this material is lost via flocculation and sedimentation to the sediments as it passes from the freshwater environment to open sea water. Since aerosol particles in the size range of a few micrometers or less have atmospheric residence times of one to several days, depending upon their size distribution and local precipitation patterns, and most substances of interest in the gas phase have similar or even longer atmospheric residence times, there is ample opportunity

for these atmospheric materials to be carried hundreds to thousands of kilometers before being deposited on the ocean surface. Metals

Table 5 presents estimates of the natural and anthropogenic emission of several metals to the global atmosphere. Note that ranges of estimates and the best estimate are given. It appears from Table 5 that anthropogenic sources dominate for lead, cadmium, and zinc, with essentially equal contributions for copper, nickel, and arsenic. Clearly a significant fraction of the input of these metals from the atmosphere to the ocean could be derived largely from anthropogenic sources. Table 6 provides an estimate of the global input of several metals from the atmosphere to the ocean and compares these fluxes with those from rivers. Estimates are given for both the dissolved and particulate forms of the metals. These estimates suggest that rivers are generally the primary source of particulate Table 5 Metal

Lead Cadmium Zinc Copper Arsenic Nickel

Emissions of some metals to the global atmosphere Anthropogenic emissions (109 g year1)

Natural emissions (109g year1)

Range

Best estimate

Range

Best estimate

289–376 3.1–12 70–194 20–51 12–26 24–87

332 7.6 132 35 18 56

1–23 0.15–2.6 4–86 2.3–54 0.9–23 3–57

12 1.3 45 28 12 30

Data reproduced with permission from Duce et al., 1991.



243

Estimates of the input of some metals to the global ocean

Table 6 Metal

Atmospheric input

Iron Copper Nickel Zinc Arsenic Cadmium Lead

Riverine input

Dissolved (109 g year1)

Particulate (109 g year1)

Dissolved (109 g year1)

Particulate (109 g year1)

1600–4800 14–45 8–11 33–170 2.3–5 1.9–3.3 50–100

14 000–42 000 2–7 14–17 11–55 1.3–3 0.4–0.7 6–12

1100 10 11 6 10 0.3 2

110 000 1 500 1 400 3 900 80 15 1 600

Data reproduced with permission from Duce et al., 1991.

Pb concentration in seawater, pmol/kg and _ 10 3 Pb concentration in the atmosphere, 10 g / m

metals in the ocean, although again a significant fraction of this material may not get past the coastal zone. For the dissolved phase atmospheric and riverine inputs are roughly equal for metals such as iron, copper, and nickel; while for zinc, cadmium, and particularly lead atmospheric inputs appear to dominate. These estimates were made based on data collected in the mid-1980s. Extensive efforts to control the release of atmospheric lead, which has been primarily from the combustion of leaded gasoline, are now resulting in considerably lower concentrations of lead in many areas of the open ocean. For example, Figure 2 shows that the concentration of dissolved lead in surface sea water near Bermuda has been decreasing regularly over the past 15–20

years, as has the atmospheric lead concentration in that region. This indicates clearly that at least for very particle-reactive metals such as lead, which has a short lifetime in the ocean (several years), even the open ocean can recover rather rapidly when the anthropogenic input of such metals is reduced or ended. Unfortunately, many of the other metals of most concern have much longer residence times in the ocean (thousands to tens of thousands of years). Figure 3 presents the calculated fluxes of several metals from the atmosphere to the ocean surface and from the ocean to the seafloor in the 1980s in the tropical central North Pacific. Note that for most metals the two fluxes are quite similar, suggesting the potential importance of atmospheric input to the

180 Marine dissolved Pb Atmospheric Pb

160 140 120 100 80 60 40 20 0 1970

1975

1980

1985 Year

1990

1995

2000

Figure 2 Changes in concentration of atmospheric lead at Bermuda and dissolved surface oceanic lead near Bermuda from the mid1970s to the mid-1990s. (Data reproduced with permission from Wu J and Boyle EA (1997) Lead in the western North Atlantic Ocean: Completed response to leaded gasoline phaseout. Geochimica et Cosmochimica Acta 61: 3279–3283; and from Huang S, Arimoto R and Rahn KA (1996) Changes in atmospheric lead and other pollution elements at Bermuda. Journal of Geophysical Research 101: 21 033–21 040.)


244


Air/sea exchange

1200

560

0.67

7.8

8.9

67

4.5

7

Al

Fe

Th

V

Cu

Zn

Se

Pb

4700

2600

0.55

5

10

12

0.007

0.3

Flux to sediments

Figure 3 A comparison of the calculated fluxes of aluminum (Al), iron (Fe), thorium (Th), vanadium (V), copper (Cu), zinc (Zn), selenium (Se), and lead (Pb) (in 109 g cm2 year1) from the atmosphere to the ocean and from the ocean to the sediments in the central tropical North Pacific. For each metal note the relative similarity in the two fluxes, except for lead and selenium. (Reproduced with permission from Duce, 1998.)

marine sedimentation of these metals in this region. Lead and selenium are exceptions, however, as the atmospheric flux is much greater than the flux to the seafloor. The fluxes to the seafloor represent average fluxes over the past several thousand years, whereas the atmospheric fluxes are roughly for the present time. The atmospheric lead flux is apparently much larger than the flux of lead to the sediments, primarily because of the high flux of anthropogenic lead from the atmosphere to the ocean since the introduction of tetraethyllead in gasoline in the 1920s. (The atmospheric flux is much lower now than in the 1980s, as discussed above.) However, in the case of selenium the apparently higher atmospheric flux is an artifact, because most of the flux of selenium from the atmosphere to the ocean is simply marine-derived selenium that has been emitted from the ocean to the atmosphere as gases, such as dimethyl selenide (DMSe). DMSe is oxidized in the atmosphere and returned to the ocean, i.e., the selenium input is simply a recycled marine flux. Thus, care must be taken when making comparisons of this type.

surface waters. Table 7 presents a recent estimate of the current input of fixed nitrogen to the global ocean from rivers, the atmosphere, and nitrogen fixation. From the numbers given it is apparent that all three sources are likely important, and within the uncertainties of the estimates they are roughly equal. In the case of rivers, about half of the nitrogen input is anthropogenic for atmospheric input perhaps the most important information in Table 7 is that the organic nitrogen flux appears to be equal to or perhaps significantly greater than the inorganic (i.e., ammonium and nitrate) nitrogen flux. The source of the organic nitrogen is not known, but there are indications that a large fraction of it is anthropogenic in origin. This is a form of atmospheric nitrogen input to the ocean that had not been considered until very recently, as there had been few measurements of organic nitrogen input to the ocean before the mid1990s. The chemical forms of this organic nitrogen are still largely unknown. Of particular concern are potential changes to the input of atmospheric nitrogen to the open ocean in

Nitrogen Species

Table 7 Estimates of the current input of reactive nitrogen to the global ocean

There is growing concern about the input of anthropogenic nitrogen species to the global ocean. This issue is of particular importance in regions where nitrogen is the limiting nutrient, e.g., the oligotrophic waters of the central oceanic gyres. Estimates to date suggest that in such regions atmospheric nitrogen will in general account for only a few percent of the total ‘new’ nitrogen delivered to the photic zone, with most of the ‘new’ nutrient nitrogen derived from the upwelling of nutrient-rich deeper waters and from nitrogen fixation in the sea. It is recognized, however, that the atmospheric input is highly episodic, and at times it may play a much more important role as a source for nitrogen in

Source

From the atmosphere Dissolved inorganic nitrogen Dissolved organic nitrogen From rivers (dissolved inorganic þ organic nitrogen) Natural Anthropogenic From nitrogen fixation within the ocean

Nitrogen input (1012 g year1)

28–70 28–84

14–35 7–35 14–42

Data reproduced with permission from Cornell S, Rendell A and Jickells T (1995) Atmospheric inputs of dissolved organic nitrogen to the oceans. Nature 376: 243–246.



Table 8

245

Estimates of anthropogenic reactive nitrogen production, 1990 and 2020

Region

Energy (NOx)

Fertilizer

1990 2020 (1012g N year1)

D

Factor

% of total increase

1990 2020 (1012g N year1)

D

Factor

% of total increase

USA/Canada Europe Australia Japan

7.6 4.9 0.3 0.8

10.1 5.2 0.4 0.8

2.5 0.3 0.1 0

1.3 1.1 1.3 1.0

10 1 0.4 0

13.3 15.4 — —

14.2 15.4 — —

0.9 0 — —

1.1 1.0 — —

1.6 0 — —

Asia Central/South America Africa Former Soviet Union Total

3.5 1.5 0.7 2.2 21

13.2 5.9 4.2 5.7 45

9.7 4.4 3.5 3.5 24

3.8 3.9 6.0 2.5 2.1

39 18 15 15 100

36 1.8 2.1 10 79

85 4.5 5.2 10 134

49 2.7 3.1 0 55

2.4 2.5 2.5 1.0 1.7

88 5 6 0 100

Data adapted with permission from Galloway et al., 1995.

the future as a result of increasing human activities. The amount of nitrogen fixation (formation of reactive nitrogen) produced from energy sources (primarily as NOx, nitrogen oxides), fertilizers, and legumes in 1990 and in 2020 as a result of human activities as well as the current and predicted future geographic distribution of the atmospheric deposition of reactive nitrogen to the continents and ocean have been evaluated recently. Table 8 presents estimates of the formation of fixed nitrogen from energy use and production and from fertilizers, the two processes which would lead to the most important fluxes of reactive nitrogen to the atmosphere. Note that the most highly developed regions in the world, represented by the first four regions in the table, are predicted to show relatively little increase in the formation of fixed nitrogen, with none of these areas having a predicted increase by 2020 of more than a factor of 1.3 nor a contribution to the overall global increase in reactive nitrogen exceeding 10%. However, the regions in the lower part of Table 8 will probably contribute very significantly to increased anthropogenic reactive nitrogen formation in 2020. For example, it is predicted that the production of reactive nitrogen in Asia from energy sources will increase r fourfold, and that Asia will account for almost 40% of the global increase, while Africa will have a sixfold increase and will account for 15% of the global increase in energy-derived fixed nitrogen. It is predicted that production of reactive nitrogen from the use of fertilizers in Asia will increase by a factor of 2.4, and Asia will account for B88% of the global increase from this source. Since both energy sources (NOx, and ultimately nitrate) and fertilizer (ammonia and nitrate) result in the extensive release of reactive nitrogen to the atmosphere, the predictions above indicate that there

should be very significant increases in the atmospheric deposition to the ocean of nutrient nitrogen species downwind of such regions as Asia, Central and South America, Africa, and the former Soviet Union. This prediction has been supported by numerical modeling studies. These studies have resulted in the generation of maps of the 1980 and expected 2020 annual deposition of reactive nitrogen to the global ocean. Figure 4 shows the expected significant increase in reactive nitrogen deposition from fossil fuel combustion to the ocean to the east of all of Asia, from Southeast Asia to the Asian portion of the former Soviet Union; to the east of South Africa, northeast Africa and the Mideast and Central America and southern South America; and to the west of northwest Africa. This increased reactive nitrogen transport and deposition to the ocean will provide new sources of nutrient nitrogen to some regions of the ocean where biological production is currently nitrogen-limited. There is thus the possibility of significant impacts on regional biological primary production, at least episodically, in these regions of the open ocean. Synthetic Organic Compounds

The atmospheric residence times of many synthetic organic compounds are relatively long compared with those of the metals and nitrogen species, as mentioned previously. Many of these substances are found primarily in the gas phase in the atmosphere, and they are thus very effectively mobilized into the atmosphere during their production and use. Their long atmospheric residence times of weeks to months leads to atmospheric transport that can often be hemispheric or near hemispheric in scale. Thus


246


1 1.5 2

3 4

Figure 4 The ratio of the estimated deposition of reactive nitrogen to ocean and land surfaces in 2020 relative to 1980. (Figure reproduced with permission from Watson AJ (1997) Surface Ocean–Lower Atmosphere Study (SOLAS). Global Change Newsletter, IGBP no. 31, 9–12; data in figure adapted with permission from Galloway JN, Levy H and Kasibhatla PS (1994) Year 2020: Consequences of population growth and development on deposition of oxidized nitrogen. Ambio 23: 120–123.).

Table 9

Estimates of the atmospheric input of organochlorine compounds to the global ocean

Ocean

SHCH (106g year1)

SDDT (106g year1)

SPCB (106g year1)

HCB (106g year1)

Dieldrin (106g year1)

Chlordane (106g year1)

North Atlantic South Atlantic North Pacific South Pacific Indian Global input via the atmosphere Global input via rivers % Atmospheric input

850 97 2600 470 700 B4700 B60 B99%

16 14 66 26 43 B170 B4 B98%

100 14 36 29 52 B230 B60 B80%

17 10 20 19 11 B80 B4 B95%

17 2.0 8.9 9.5 6.0 B40 B4 B91%

8.7 1.0 8.3 1.9 2.4 B22 B4 B85%

(Data reproduced with permission from Duce et al., 1991.)

atmospheric transport and deposition in general dominates all other sources for these chemicals in sea water in open ocean regions. Table 9 compares the atmospheric and riverine inputs to the world ocean for a number of synthetic organochlorine compounds. Note that the atmosphere in most cases accounts for 90% or more of the input of these compounds to the ocean. Table 9 also presents estimates of the input of these same organochlorine compounds to the major ocean basins. Since most of these synthetic organic compounds are produced and used in the northern hemisphere, it is not surprising that the flux into the northern hemisphere ocean is greater than that to southern hemisphere marine regions. There are some differences for specific compounds in different ocean basins. For example, HCH (hexachlorocyclohexane) and DDT have a higher input rate to the North Pacific than the North Atlantic, largely because of the greater use of these compounds in Asia than in North America or

Europe. On the other hand, the input of PCBs (polychlorinated biphenyls) and dieldrin is higher to the North Atlantic than the North Pacific, primarily because of their greater use in the continental regions adjacent to the North Atlantic.

Conclusions The atmosphere transports materials to the ocean that are both harmful to marine life and that are essential for marine biological productivity. It is now apparent that atmospheric transport and deposition of some metals, nitrogen species, and synthetic organic compounds can be a significant and in some cases dominant pathway for these substances entering both estuarine and coastal waters as well as some open ocean regions. Atmospheric input clearly must be considered in any evaluation of material fluxes to marine ecosystems. However, the uncertainties in the



atmospheric fluxes of these materials to the ocean are large. The primary reasons for these large uncertainties are:

• • • • •

The lack of atmospheric concentration data over vast regions of the coastal and open ocean, particularly over extended periods of time and under varying meteorological conditions; The episodic nature of the atmospheric deposition to the ocean; The lack of accurate models of air/sea exchange, particularly for gases; The inability to measure accurately the dry deposition of particles; and The inability to measure accurately the air/sea exchange of gases.

See also Chlorinated Hydrocarbons. Metal Pollution. Refractory Metals. Transition Metals and Heavy Metal Speciation

Further Reading

247

Marine Meteorology Technical Conference on Marine Pollution, World Meteorological Organization TD-No. 890, Geneva, Switzerland. Duce RA, Liss PS, Merrill JT, et al. (1991) The atmospheric input of trace species to the world ocean. Global Biogeochemical Cycles 5: 193--259. Galloway JN, Schlesinger WH, Levy H, Michaels A, and Schnoor JL (1995) Nitrogen fixation: anthropogenic enhancement – environmental reponse. Global Biogeochemical Cycles 9: 235--252. Jickells TD (1995) Atmospheric inputs of metals and nutrients to the oceans: their magnitude and effects. Marine Chemistry 48: 199--214. Liss PS and Duce RA (1997) The Sea Surface and Global Change. Cambridge: Cambridge University Press. Paerl HW and Whitall DR (1999) Anthropogenicallyderived atmospheric nitrogen deposition, marine eutrophication and harmful algal bloom expansion: Is there a link? Ambio 28: 307--311. Prospero JM, Barrett K, Church T, et al. (1996) Nitrogen dynamics of the North Atlantic Ocean – Atmospheric deposition of nutrients to the North Atlantic Ocean. Biogeochemistry 35: 27--73. van den Hout KD (ed.) (1994) The Impact of Atmospheric Deposition of Non-Acidifying Pollutants on the Quality of European Forest Soils and the North Sea. Report of the ESQUAD Project, IMW-TNO Report No. R 93/ 329.

Duce RA (1998) Atmospheric Input of Pollution to the Oceans, pp. 9–26. Proceedings of the Commission for


ATMOSPHERIC TRANSPORT AND DEPOSITION OF PARTICULATE MATERIAL TO THE OCEANS J. M. Prospero, University of Miami, Miami, FL, USA R. Arimoto, New Mexico State University, Carlsbad, NM, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The atmosphere is the primary pathway for the transport of many geochemically important substances to the oceans. Although the magnitude of these wind-borne transports is not accurately known, there is growing evidence that atmospheric deposition significantly impacts chemical and biological processes in the oceans. It is only over the past several decades that marine scientists have come to appreciate the importance of atmospheric transport. Historically it had been assumed that the fluxes of continental materials to the oceans were dominated by rivers. But over time it was recognized that much of the riverine load was deposited in estuaries or on the continental shelves. In contrast, winds can rapidly span great distances to reach even the most remote ocean regions. The transport and deposition of particulate matter (PM) to the oceans depends on many factors including the distribution of sources, physical and chemical properties of the particles, meteorological conditions, and removal mechanisms. Our interest here focuses on particles between about 0.1 and 10 mm in diameter which, because of their small size, have atmospheric lifetimes ranging from days to several weeks. These are commonly referred to as aerosol particles or aerosols. Larger particles are deposited close to their sources and do not contribute substantially to ocean deposition except in some coastal regions. Smaller aerosols carry little mass and, while they are important for other atmospheric issues, they are not particularly important for air/sea chemical exchange. Winds carry billions of tons of PM to the ocean. Some of the PM is emitted by natural processes and some is produced anthropogenically, that is, as a result of human activities. Much PM is emitted directly as primary particles; this includes mineral (soil) dust, organic particles from plants, and emissions from anthropogenic combustion processes (e.g., from industry, homes, and vehicles) and biomass burning,

248

which can be natural (e.g., started by lightning) or anthropogenic (e.g., in clearing land, burning agricultural waste). An important PM fraction – secondary PM – is that produced from gases, natural and anthropogenic, that react in the atmosphere to form particles. One goal of marine scientists is to characterize atmospheric transport and chemical deposition to the ocean and to assess the impact of the air/sea exchange process. This is a difficult task which can only be achieved when we know the kinds of materials deposited and their temporal and spatial variability. Because of the patchy distribution of sources and the relatively short tropospheric residence times of aerosols, PM concentrations over the oceans vary by orders of magnitude in time and space. Here we review the sources and composition of aerosols and the removal mechanisms relevant to deposition issues. We then present estimates of deposition rates of some PM classes to the oceans.

Aerosol Sources, Composition, and Concentrations The composition of PM over the oceans varies greatly due to the myriad sources and variations in their strengths. Soils emit fine mineral particles. Plants produce a wide range of organic particles, ranging from decayed leaf matter, to plant waxes, and condensed organic compounds. Volcanoes sporadically inject many tons of material into the atmosphere, and much of this is deposited in the oceans; but the total amount of PM deposited over time is relatively small compared with other sources. Smelters, power plants, and incinerators emit PM highly enriched with trace metal pollutants. Combustion sources, both natural (wild fires, biomass burning) and anthropogenic (power plants, vehicles, home heating) emit thousands of organic compounds. Combustion processes and the use of fertilizers contribute to the production of nitrogen-rich particles. Pesticides and other synthetic organic compounds are emitted from industrial and domestic sources. Typically the dominant marine aerosol species by mass are: (a) sea salt, produced by breaking waves and bursting bubbles; (b) sulfate, including that from sea salt aerosol and non-sea-salt sulfate (nss-SO4 2 ), the latter of which is both transported from pollution sources on the continents and produced from


ATMOSPHERIC TRANSPORT AND DEPOSITION OF PARTICULATE MATERIAL TO THE OCEANS

gaseous precursors such as dimethyl sulfide (DMS) emitted from the oceans; (c) nitrate, originating from pollution sources on the continents and produced by lightning; (d) ammonium, derived mostly from continental sources but in some areas from ocean sources; (e) mineral dust, from arid lands; (f) organic carbon (OC), largely from anthropogenic and natural sources on the continents; and (g) black carbon (BC), from biomass burning and anthropogenic sources. PM composition is strongly size dependent not only as a result of various production mechanisms but also because size-selective removal occurs during transport. Physical production mechanisms (grinding of rocks, bursting bubbles) normally produce large particles with most of the mass in PM greater than 1-mm diameter (coarse particles). For example, the mass median diameter (MMD: 50% of the mass is greater than the MMD and 50% is less) of dust particles over deserts can be extremely large, many tens or hundreds of micrometers, but over the oceans, it is typically only several micrometers. The MMD of sea salt particles is generally in the range of about 5–10 mm, depending on wind conditions. Other primary particles including soot emitted from smoke stacks, diesel exhaust, and particles shed by plants (e.g., plant waxes, fibers), tending to be in the size range of B0.1–1.0 mm.

Table 1

Gas-phase reactions produce secondary PM ranging in size range of 0.001–0.1 mm diameter. Examples are sulfate particles produced from SO2 emitted from power plants or from the oxidation of DMS emitted from the oceans. Particles in this very fine particle mode are highly mobile. They can rapidly coagulate to form larger particles (typically 0.1–1 mm) or they can diffuse to the surface of cloud or fog droplets or to larger particles (e.g., sea salt, mineral dust). Table 1 presents concentration data for the aerosols that make up most of the PM mass over the oceans; it also includes data for vanadium, which is included as an example of an element strongly affected by pollution sources. The column on the extreme right shows the total aerosol concentration less that of sea salt, so as to better illustrate the impact of transported PM. These data are the product of longterm measurements obtained from a global ocean network. In general, PM concentrations are much higher in the Northern Hemisphere relative to the Southern Hemisphere. Mineral dust shows an extremely wide range of concentrations over the oceans; the highest are over the tropical North Atlantic and the western North Pacific. These reflect the impact of dust transport from North Africa and China, respectively. Dust concentrations in the southern oceans tend to be extremely low due to the

Annual mean aerosol concentrations measured at remote ocean stations Station locationa Lat 1N

North Atlantic Mace Head Bermuda Barbados

Sea salt (mg m 3)

NO3 (mg m 3)

nss-SO4 (mg m 3)

NH4 (mg m 3)

Dust V (mg m 3) (mg m 3)

Total b (mg m 3)

Total SSc (mg m 3)

19.8

4.1

7.2

3.0

15.5

4.1

49.5

29.8

13.8 15.1

0.3 0.4

0.5 0.5

0.8 0.0

0.7 0.7

0.2 0.3

15.4 16.7

1.6 1.6

Lon 1E

North Pacific Western Pacific Cheju, 33.5 126.5 Korea Central Pacific Midway 28.2 177.4 Oahu 21.36 157.7 53.5

9.9

14.1

1.5

2.0

0.9

0.5

0.9

19.0

4.9

32.3 13.2

64.9 59.4

13.7 16.5

1.1 0.5

2.2 0.8

0.3 0.1

5.6 14.6

1.3 1.9

22.8 32.5

9.2 16.0

South Pacific American 14.3 Samoa

170.6

16.7

0.1

0.4

0.0

0.1

17.2

0.5

62.5

0.3

0.0

0.1

0.5

0.1

Antarctic Mawson

249

67.6

0.0

a

Station Location: negative latitudes – Southern Hemisphere: negative longitudes – Western Hemisphere. Total aerosol: the sum of the major aerosol components – sea salt, soil dust, nss-SO4, NO3, and NH4. c Total SS: Total aerosol minus sea salt. b


250


dearth of strong dust sources combined with the great distances to central ocean regions. The impact of air pollution is evident over much of the Northern Hemisphere. Extremely high NO3 and nss-SO4 2 concentrations are seen in the western Pacific near Asia; these are attributable to continental outflow and exacerbated by limited emission controls. Moderately high pollutant levels are seen over the North Atlantic as well, a result of emissions from North America and Europe. In contrast, the concentrations of NO3 and nss-SO4 2 at American Samoa and the Antarctic stations Mawson and Palmer are extremely low; these represent conditions that one might expect when pollution impacts are minimal. The larger-scale picture of PM distributions is provided by sensors such as the advanced very high resolution radiometer (AVHRR), which measures solar radiation backscattered to space by PM to estimate aerosol optical thickness (AOT) (Figure 1). There are three characteristics of the global distributions of AOT, all consistent with the data in Table 1. First, the highest AOT (i.e., the greatest column loadings of PM) is found close to the continents. This distribution affirms the fact that over most of the ocean PM is largely the result of material transported from the continents. Second, there are large seasonal differences in PM concentrations due to the seasonality of emissions and meteorology. Third, some continents emit more PM than others, illustrating the large-scale differences in production and transport. Especially notable in satellite images is a large plume of AOT over the tropical Atlantic, extending from the coast of Africa to South America (December– February) and to the Caribbean (June–August). This plume is mainly African dust. A large region of high AOT over the Arabian Sea (June–August) is attributed to dust from Africa and the Middle East. In this same season, a large PM plume seen off the west coast of southern Africa is attributed to smoke from intense biomass burning. Substantial aerosol plumes are also seen over the North Atlantic; these are caused by pollutants from North America and Europe. Large regions of high AOT are seen along the coast of Asia; but the Asian plume is most prominent in spring when large quantities of soil dust mix with pollution aerosols. The attribution of these plumes to these dominant aerosol types is supported by evidence from field studies.

Aerosol Removal Mechanisms Estimating Wet and Dry Deposition

PM is deposited to the ocean by two broadly characterized mechanisms: (1) dry deposition, in which a

particle is transferred directly from the atmosphere to the surface; and (2) wet deposition, in which a particle is first incorporated into a cloud or rain droplet that subsequently falls to the surface. The relative efficiency of the removal processes is dependent on a number of factors, especially the particle-size distribution and the hygroscopic properties of the aerosol. In most ocean regions, wet removal is thought to dominate for most aerosol species. Wet deposition is relatively easy to measure using precipitation collectors, such as automatic bucket systems that open only when precipitation falls. Dry deposition, on the other hand, is much more difficult to collect because this process is affected by a variety of factors, all highly variable: the properties of the aerosol and the water surface, atmospheric stability, relative humidity, wind velocity, etc. Furthermore, while there is a vast quantity of data on wet deposition to continental areas, there is very little for ocean environments. There are some long-term records for selected species in precipitation at a few island stations but there are no matching data for dry deposition. Wet Deposition

Long-term studies show that on average the wet deposition rates of many species are related to their concentrations in the atmospheric aerosol and to rainfall rates. This relationship is expressed in terms of a dimensionless scavenging ratio, S: S ¼ Cp rC1 a

½1

where Cp is the concentration of the substance in precipitation (g kg 1), r the density of air (B1.2 kg m 3), and Ca the aerosol concentration of the species of interest (g m 3). Wet deposition rates depend on the vertical distribution of PM and the type of precipitation event (e.g., frontal, cumulus, and stratus). In practice, comprehensive, long-term, aerosol data are only available from surface sites; consequently one must assume that over the longer term the PM concentrations in surface-level air are correlated with their vertical distributions. Therefore, S is appropriately calculated only when data have been obtained over periods of a year or more. That is, the variability in the aerosol and precipitation events must be smoothed by the averaging process. Typically used values for S fall in the range of 200–1000. Scavenging ratios can be applied to regions where no precipitation data exist using the following



251

(a) Dec., Jan., Feb.

EAOT 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05

Husar, Stove and Prospero, 1996

(b) Jun., Jul., Aug.

EAOT 0.5 0.45 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05

Husar, Stove and Prospero, 1996

Figure 1 The distribution of aerosols over the oceans inferred from aerosol optical thickness estimated (EAOT) by National Oceanic and Atmospheric Administration (NOAA) AVHRR. Aerosol optical thickness is a measure of the attenuation of direct solar radiation at a specific wavelength due to the scattering and absorption caused by aerosols. Large values of optical thickness suggest high concentrations of aerosols. Distributions are shown for the months (a) Dec.–Feb. and (b) Jun.–Aug. Adapted by permission of American Geophysical Union from Husar RB, Prospero JM, and Stowe LL, Characterization of tropospheric aerosols over the oceans with the NOAA advanced very high resolution radiometer optical thickness operational product, Journal of Geophysical Research, vol. 102(D14), pp. 16889–16909, 1997. Copyright 1997 American Geophysical Union.

expression: Fp ¼ PCp ¼ PSr1 Ca

½2

where Fp is the wet deposition flux (g m 2 yr 1) and P is the precipitation rate (m yr 1), using conversion factors to translate rainfall amounts to the mass of water deposited per unit area. Note


252


that the combined terms PSr 1Ca have a unit of velocity. Dry Deposition

There are no widely accepted methods for directly measuring PM dry deposition to water surfaces. In practice, dry deposition is almost always calculated by assuming that the deposition rate is proportional to PM concentration times a ‘deposition velocity’, vd. The dry PM flux, Fd (g m 2s 1), is given by Fd ¼ vd Ca

½3

where vd is the dry deposition velocity (m s 1) and Ca is the mass concentration of the substance in the atmosphere (g m 3). Deposition velocities have been modeled based on physical principles, and they have been empirically derived by concurrently measuring the concentration of PM species in the atmosphere and the amount deposited to a surrogate surface (typically a flat plate). While there have been some determinations of deposition velocities at continental sites, the data for ocean regions are scant. In eqn [3], vd incorporates all the processes of dry deposition, but it is difficult to accurately parametrize vd for the ambient aerosol because the importance of these processes varies with particle size. For PM between 0.1 and 1.0 mm, dry deposition is inefficient, and wet removal is normally the major sink. Gravitational settling and surface impaction control the dry removal of PM larger than about 1 mm while below about 0.1 mm, Brownian diffusion dominates. Each of these mechanisms depends on wind speed, aerosol hygroscopicity, relative humidity near the surface, and other factors which are poorly characterized. Unfortunately, there is no general agreement on how to resolve these uncertainties. Nonetheless, many estimates of dry deposition make the following assumptions about the dependence on particle size and the uncertainties in the resulting estimated deposition rate:

• • •

submicrometer aerosol particles: 0.001 m s 17 a factor of 3, supramicrometer crustal particles not associated with sea salt: 0.01 m s 17 a factor of 3, giant sea salt particles and materials carried by them: 0.03 m s 17 a factor of 2.

Despite the widespread use of these values, it should be emphasized that they are only estimates and that the error range is, if anything, probably optimistic. For example, wind speed has a great influence on deposition velocities: for PM B0.1–1.0 mm in diameter, the deposition velocity ranges from c. 0.005 cm s 1 at 5 m s 1 to c. 0.1 cm s 1 at 15 m s 1.

Deposition of Aerosols to the Oceans In this section, we present estimates of the deposition of a number of PM species that are potentially important for biogeochemical processes in the oceans. Estimates of deposition to specific locations can be made using the above relationships, assuming that the necessary aerosol concentration data are available. Larger-scale estimates of deposition are best obtained with atmospheric chemical transport models as discussed below. These models are subject to large uncertainties because they generally rely on estimates of aerosol properties and concentrations over the oceans and they incorporate highly parametrized removal schemes. For illustrative purposes, we present results for mineral dust, selected trace elements, and a group of nitrogen-containing species. A wide range of natural and anthropogenic organic species are also transported to the oceans and deposited there. Of these, certain persistent organic pollutants are known to have a harmful impact on marine biological systems. There are, however, relatively little data on the large-scale transport of organics that would enable us to address this issue on a global scale. Consequently, we do not include organic species in this report.

Deposition of Mineral Dust and Eolian Iron

In many ocean regions, primary (photosynthetic) biological production is limited by the classical nutrients such as nitrate and phosphate. In nutrient-rich surface waters, biological activity is usually high which results in high chlorophyll concentrations. But in large areas of the world’s oceans, nutrient concentrations are high, yet chlorophyll remains low which suggests low productivity. These are termed high-nutrient, low-chlorophyll (HNLC) waters; prominent examples include the equatorial Pacific and much of the high-latitude southern oceans. In the 1980s, it was found that primary production in some HNLC regions was limited by the availability of iron, an essential micronutrient in certain enzymes involved in photosynthesis. Remote ocean regions are largely dependent on atmospheric dust for the input of iron. The deposition of this eolian iron and its impact on productivity has important implications for the global CO2 budget and, hence, climate. Increased iron fluxes could conceivably fertilize the oceans, thereby increasing productivity and drawing down atmospheric CO2. In addition, certain nitrogen fixers (e.g., Trichodesmium sp.) have a high iron requirement; an increased eolian iron flux could stimulate the growth



of nitrogen fixers, thereby increasing nitrate levels and further contributing to the CO2 drawdown. Much effort has gone into estimating the temporal and spatial patterns of dust deposition to the oceans. Some studies have used satellite aerosol measurements coupled with network measurements of dust to calculate wet and dry deposition fluxes using scavenging ratios and deposition velocities. Recently, regional- and global-scale models have been developed to provide estimates of dust emissions, transport, and deposition. Dust is generally included as a passive tracer, and its removal is highly parametrized. Dry deposition is calculated using the model’s dust size distribution and size-dependent deposition (see section ‘Dry deposition’ above). Wet removal is also modeled, but the interaction of dust with clouds is not well constrained, partly because aerosol–cloud interactions in general are not well understood, and also because there are few measurements of cloud microphysical measurements in dusty regions. Mineral dust is not readily soluble in water; so some models assume that mineral aerosols do not interact with clouds directly, but rather are scavenged via subcloud removal – hence, simple scavenging ratios are used (see section ‘Wet deposition’ above). As a consequence of these uncertainties, current models show large differences in dust wet deposition lifetimes, ranging from 10 to 56 days. A typical model estimate of dust deposition rates to the oceans is shown in Figure 2. In the tropical North Atlantic, rates typically range from 2 to 10 g m 2 yr 1; over the Arabian Sea, they are as

0.000

0.2

0.5

1

2

253

high as 20 g m 2 yr 1. Over much of the North Pacific, rates are in the range 0.5–1 g m 2 yr 1, increasing to 1–2 g m 2 yr 1 closer to the coast of Asia. Dust deposition rates in Figure 2 tend to mirror the PM distribution shown in Figure 1, which, as previously stated, is largely linked to the presence of dust and, in some regions, smoke from biomass burning. Table 2 shows estimates of deposition rates to the major ocean basins produced by eight commonly used models. There is considerable agreement for the North Atlantic which is heavily impacted by African dust. In contrast, there is a considerable spread in the estimates for other regions, especially the Indian Ocean and South Pacific. Despite these differences, current models yield a reasonable, albeit broad, match with sediment trap measurements in the oceans. These various studies show that North Africa is clearly the world’s most active dust source followed by the Middle East and Central Asia. In effect, these combine to form a global dust belt that dominates transport to the oceans. These sources account for the much greater deposition rates to the northern oceans compared with southern oceans. Nonetheless, there are some substantial and important dust sources in the Southern Hemisphere in Australia, southern Africa, and southern South America. Within these continental regions, certain specific environments are particularly active dust sources, and they are sensitive to changes in climate, especially rainfall and wind speed. The presence of large, deep, alluvial deposits, usually deposited in the Pleistocene or

5

10

20

50

Figure 2 Model estimates of dust deposition rates (units: g m 2 yr 1) to the continents and the oceans. Reproduced by permission of American Geophysical Union from Mahowald NM, Baker AR, Bergametti G, et al., Atmospheric global dust cycle and iron inputs to the ocean. Global Biogeochemical Cycles, vol. 19, GB4025, 2005. Copyright 2005 American Geophysical Union.


254


Table 2

Estimate of mean annual dust deposition to the global ocean and to various ocean basins Annual dust deposition ratea (1012 g yr 1)

Reference

Duce et al. (1991) Prospero (1996) Ginoux et al. (2001) Zender et al. (2003) Luo et al. (2003) Ginoux et al. (2004) Tegen et al. (2004) Kaufman et al. (2005)

GO

NAO

SAO

NPO

SPO

NIO

SIO

910 358 478 314 428 505 422

220 220 184 178 230 161 259 140

24 5 20 29 30 20 35

480 96 92 31 35 117 56

39 8 28 8 20 28 11

100 20 154 36 113 164 61

44 9 12 15

a

GO, Global Oceans; NAO, North Atlantic Ocean; SAO, South Atlantic Ocean; NPO, North Pacific Ocean; SPO, South Pacific Ocean; NIO, North Indian Ocean; SIO, South Indian Ocean. Adapted from Engelstaedter S, Tegen I, and Washington R (2006) North African dust emissions and transport. Earth-Science Reviews 79(1–2): 73–100.

Table 3 Metal

Iron Copper Nickel Zinc Arsenic Cadium Lead

Atmospheric and riverine fluxes of dissolved and particulate trace metals to the oceana Atmospheric transports

Riverine transports

Dissolved transports

AtmDiss

AtmPart

Ratio Diss/Part

RivDiss

RivPart

Ratio Diss/Part

Ratio AtmDiss/RivDiss

3200 30 10 102 4 3 75

28 000 5 16 33 2 1 9

0.1 6.6 0.6 3.1 1.7 4.7 8.3

1100 10 11 6 10 0.3 2

11 000 1 500 1 400 3 900 80 15 1 600

0.100 0.007 0.008 0.002 0.125 0.020 0.001

2.9 3.0 0.9 16.9 0.4 8.7 37.5

Units: 109 g yr 1. Atm, atmospheric; Riv, riverine; Diss, Dissolved; Part, particulate. Adapted from Duce RA, Liss PS, Merrill JT, et al. (1991) The atmospheric input of trace species to the world ocean. Global Biogeochemical Cycles 5: 193–259.

a

Holocene, is a common prerequisite for strong dust sources. Trace Element Deposition

Some oceanographers study the biogeochemical cycling of trace elements and seek to quantify the elements’ oceanic sources and sinks. In regions dominated by mineral dust, the ratios of many trace elements (e.g., Al, Ba, Ca, Cs, Fe, Hf, Mn, Rb, Sc, Ta, Th, and Yb) are similar to those in geological materials such as soils, thus implicating mineral dust as their main source. Several elements (Co, Cr, Eu, Mg, and Na) show slight enrichments over crustal values while others such as As, Cd, Cu, Ni, Pb, Sb, Se, V, and Zn are strongly enriched. Such large enrichments are typically associated with pollution impacts, but emissions from natural sources such as volcanoes can also be responsible.

The impact of trace element deposition on ocean biogeochemistry depends to a great extent on the degree to which the elements are soluble in seawater. Extensive studies of trace element solubilities in various natural aqueous media or in aqueous solutions of similar composition yield a wide range of solubilities; these depend on the types of aerosols used, the exposure times, and other experimental variables, especially pH. Thus it is difficult to convert the estimated air/sea exchange rates into an effective or bioavailable flux of trace elements, and therefore it is difficult to accurately assess the impact of PM deposition on ocean processes. Comparison of Trace Element Transport by Rivers and by the Atmosphere

Rivers carry large quantities of dissolved and suspended materials to the oceans. Table 3 compares the



amounts of selected trace elements carried by rivers with that carried by winds and also the relative amounts of particulate and dissolved or soluble phases in the transported material. Riverine transports of trace elements are overwhelmingly in the particulate phases and the ratio of dissolved to particulate phases ranges from about 0.001 to 0.13. In comparison, the corresponding ratio for atmospheric transport is much larger, ranging from 0.1 to 8.3. Table 3 also shows that the dissolved or soluble inputs to the ocean from the atmosphere exceed those from rivers; in almost all cases the ratio is greater than unity, in some cases much larger. The comparison of river versus atmospheric inputs is based on the measured concentrations in rivers before they reach the oceans. However, much of the material carried by rivers is rapidly deposited when the rivers reach the sea; therefore, the impact of air/ sea exchange on ocean systems is in reality much greater than suggested by Table 3. We emphasize, however, that the data in Table 3 are rather old. Recent work shows that the concentrations of some trace elements have changed significantly over time. For example, during the mid-1900s, aerosol lead greatly increased in response to increasing pollution emissions but in later years concentrations decreased as controls were implemented. In recent decades, other trace elements have increased due to growing industrialization in developing nations. Also, recent research suggests that there is considerably more uncertainty in PM trace metal solubility than shown in Table 3. Nonetheless, one would still expect that the impact of atmospheric transport is much greater than river transport, especially for the open ocean. Nitrogen Deposition

Anthropogenic activities have greatly increased the amounts of nitrogenous materials that enter the atmosphere and find their way into the rivers (Table 4). There is interest in the possible impacts of these materials on the marine environment, especially about chemicals such as nitrate that can affect primary production. This issue is of particular importance in regions where nitrogen is the limiting nutrient, for example, the oligotrophic central oceanic gyres where an enhancement in productivity would increase the drawdown of CO2 and hence affect climate. In coastal waters, atmospheric inputs could contribute to eutrophication although one would expect the inputs from rivers to dominate. There are two broad classes of nitrogen compounds of interest: oxidized and reduced. The most important oxidized species are NO and NO2

Table 4

255

Atmospheric emissions of fixed nitrogen, 1993a

Sources

NOx

NH3

Anthropogenic Biomass burning Agricultural activity Fossil fuel combustion Industry Total anthropogenic

6.4 2.6 20.9 6.4 36.3

4.6 39.7 0.1 2.8 47.2

2.9 5.4 0.8 0.6 0.0 6.8

4.6 0.0 0.8 0.0 5.6 11.0

43.1 5.3

58.2 4.3

Natural Soils, vegetation, and animals Lightning Natural fires Stratosphere exchange Ocean exchange Total natural Grand total Ratio: anthropogenic/natural

Units Tg/g(N) yr 1. Adapted from Jickells TD (2006) The role of air–sea exchange in the marine nitrogen cycle. Biogeosciences 3: 271–280. a

(collectively referred to as NOx) and NOy (termed reactive odd nitrogen) which is comprised of NOx plus the compounds produced from its oxidation, including HNO3 and other compounds. NOx is rapidly oxidized in the atmosphere to a wide range of compounds, many of which are ultimately converted to HNO3 and aerosol NO3 . In the marine boundary layer, HNO3 reacts rapidly with sea salt particles and promptly deposits on the sea surface. Indeed, NO3 is the N-containing compound of greatest interest in terms of impact on the oceans, and it is the N compound most commonly measured and modeled. Table 4 lists the major sources of oxidized and reduced N emitted to the atmosphere. The primary natural sources of NOx are biological fixation and lightning, the latter being rather minor. In modern times, fossil-fuel combustion along with industry and biomass burning dominate the oxidized N cycle. The ratio of anthropogenic NOx emissions to that of natural sources is 2.5 and continues to increase. While most research has focused on inorganic N (IN), there is evidence that organic nitrogen (ON) also may play an important role in marine biogeochemical cycling. However, there are relatively few data on ON compounds and most focus only on dissolved ON. The major reduced nitrogen species (NHx) are aerosol NH4 þ and NH3, the latter being the only gas-phase species that significantly titrates atmospheric acidity. The major natural sources of NH3 (Table 4) include soils, vegetation, and excreta from wild animals. However, the emissions of NH3 to the atmosphere are now dominated by fertilizers and the


256


excreta from dairy and beef cattle. Indeed, the ratio of anthropogenic and natural NH3 emissions is 9. The oceans can also be a source of NH3 under certain conditions in some regions, but the continental sources clearly dominate. Models provide estimates of the present-day atmospheric N fluxes to the oceans and their spatial distribution. Figure 3 presents the deposition rate of

reactive nitrogen NOy þ NHx for the year 2000. As was the case for dust, emissions and deposition in the Northern Hemisphere are much greater than those of the Southern Hemisphere. Deposition rates are extremely high adjacent to the continental coastlines; thus one would expect that the adjacent water bodies would be most heavily impacted. The total N flux to the ocean (NOy and NHx but not including ON) in mg N m−2 6000 3000

60° N

2000 1000 900

30° N

800

Latitude (deg)

700 600 EQ.

500 400 300 200

30° S

100 75 60° S

50 25 10

180° W 150° W 120° W

90° W

60° W

30° W

0° E

30° E

60° E

90° E

120° E 150° E 180° E

Longitude (deg)

Figure 3 Model estimates of the deposition rate of total reactive nitrogen (NOy þ NHx) (units: mg N m 2 yr 1) in the year 2000. Reprinted with permission from Dentener F, Stevenson D, Ellingsen K, et al. (2006) The global atmospheric environment for the next generation. Environmental Science and Technology 40(11): 3586–3594 (doi:10.1021/es0523845). Copyright (2006) American Chemical Society.

Table 5

NOy and NHx deposition for the year 2000

Deposition region

Ocean Coastal ocean NH SH World Ratio: ocean/ world

NOy þ NHx

NHx

NOy Total (Tg(N) yr 1)

Mean rate (mg(N) m2 yr 1)

Total (Tg(N) yr 1)

Mean rate (mg(N) m2 yr 1)

Total (Tg(N) yr 1)

23 4 38 14 52 0.43

61 192 150 54 102

24 4 48 16 65 0.36

63 206

47 8 87 30 117 0.40

126

Adapted from Dentener F, Drevet J, Lamarque JF, et al. (2006) Nitrogen and sulfur deposition on regional and global scales: A multimodel evaluation. Global Biogeochemical Cycles 20: GB4003 (doi:10.1029/2005GB002672).



2000 was 46 Tg N yr 1 of which 8 Tg N yr 1 is deposited to the coastal ocean (Table 5). The deposition of oxidized forms (NOy) is essentially equal to that of reduced forms (NHx). The ocean N deposition amounts to c. 40% of global emissions. The total reactive N transport by rivers is about 48 Tg N yr 1, essentially the same as the atmospheric source. However, there is evidence that fluvial nitrogen inputs to the oceans are denitrified on the shelf and that the shelf region is a sink rather than a source of nitrogen for the open oceans. Thus it appears that air/sea transfer is the major source of N transported to the open ocean. ON compounds could also be playing a significant role in total N fluxes. Studies from many different environments suggest that ON constitutes about a third of the total atmospheric reactive nitrogen. Thus, ON could add significantly to the total global flux to the oceans, conceivably raising the total to about 69 Tg N yr 1.

Conclusions It is now recognized that atmospheric transport plays a central role in ocean biogeochemical processes. There is increased interest in the chemically coupled ocean/atmosphere system, how this system has changed over time, and how it might respond to global change. Although many models are currently focusing on this question, the development of these models is handicapped by the dearth of measurements over many ocean regions. It remains a formidable challenge to the community to carry out the necessary measurements over such large ocean areas.

See also Anthropogenic Trace Elements in the Ocean. Atmospheric Input of Pollutants. Iron Fertilization. Metal Pollution. Nitrogen Cycle. Biological Pump and Particle Fluxes. Trace Element Nutrients.

Further Reading Arimoto R, Kim YJ, Kim YP, et al. (2006) Characterization of Asian dust during ACE-Asia, global and planetary change. Monitoring and Modelling of Asian Dust Storms 52(1–4): 23--56.

257

Arimoto R, Ray BJ, Lewis NF, Tomza U, and Duce RA (1991) Mass-particle size distributions of atmospheric dust and the dry deposition of dust to the remote ocean. Journal of Geophysical Research – Atmospheres 102(D13): 15867--15874. Boyd PW, Watson A, Law CS, et al. (2000) A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization of waters. Nature 407: 695--702. Dentener F, Drevet J, Lamarque JF, et al. (2006) Nitrogen and sulfur deposition on regional and global scales: A multimodel evaluation. Global Biogeochemical Cycles 20: GB4003 (doi:10.1029/2005GB002672). Dentener F, Stevenson D, Ellingsen K, et al. (2006) The global atmospheric environment for the next generation. Environmental Science and Technology 40(11): 3586--3594 (doi:10.1021/es0523845). Duce RA, Liss PS, Merrill JT, et al. (1991) The atmospheric input of trace species to the world ocean. Global Biogeochemical Cycles 5: 193--259. Engelstaedter S, Tegen I, and Washington R (2006) North African dust emissions and transport. Earth-Science Reviews 79(1–2): 73--100. Harrison SP, Kohfeld KE, Roeland C, and Claquin T (2001) The role of dust in climate today, at the Last Glacial Maximum and in the future. Earth-Science Reviews 54: 43--80. Husar RB, Prospero JM, and Stowe LL (1997) Characterization of tropospheric aerosols over the oceans with the NOAA advanced very high resolution radiometer optical thickness operational product. Journal of Geophysical Research 102(D14): 16889--16909. Jickells TD (2006) The role of air–sea exchange in the marine nitrogen cycle. Biogeosciences 3: 271--280. Jurado E, Jaward F, Lohmann R, et al. (2005) Wet deposition of persistent organic pollutants to the global oceans. Environmental Science and Technology 39(8): 2426--2435 (doi:10.1021/es048599 g). Mahowald NM, Baker AR, Bergametti G, et al. (2005) Atmospheric global dust cycle and iron inputs to the ocean. Global Biogeochemical Cycles 19: GB4025 (doi:10.1029/2004GB002402). Parekh P, Follows MJ, and Boyle EA (2005) Decoupling of iron and phosphate in the global ocean. Global Biogeochemical Cycles 19: GB2020 (doi:10.1029/ 2004GB002280). Prospero JM (1996) The atmospheric transport of particles to the ocean. In: Ittekkot V, Scha¨fer P, Honjo S and Depetris PJ (eds.) Particle Flux in the Ocean. SCOPE Report 57, pp. 19--52. Chichester: Wiley. Wesely ML and Hicks BB (2000) A review of the current status of knowledge on dry deposition. Atmospheric Environment 34(12–14): 2261--2282.


AUTHIGENIC DEPOSITS G. M. McMurtry, University of Hawaii at Manoa, Honolulu, HI, USA Copyright & 2001 Elsevier Ltd.

usually lithified sediment or basalt substrate of a seamount or island slope to create ‘pseudo’ or ‘seamount nodules’ that are otherwise indistinguishable from crustal pavements.


Ferromanganese Nodules

Introduction There are many kinds and forms of authigenic mineral deposits in the deep-sea. Here we concentrate on the more abundant and, in some cases, potentially economic types, namely ferromanganese deposits, phosphorites, marine barite, and the authigenic silicate minerals in deep-sea sediments. Hydrothermal deposits and a class of sedimentary iron and manganese oxides associated with seafloor hydrothermal activity – the metalliferous sediments found as a basal sequence in the deep-sea sedimentary column and in deep basins adjacent to active spreading centers – are described elsewhere in the encyclopedia. Ferromanganese deposits are here divided into nodules and crusts on the basis of morphology and environment of formation. The sources of the major and minor metal enrichments in these deposits are contrasted between continental weathering and submarine volcanism, with important enrichment influences from the marine biosphere. Marine barite has a similar mixed origin with strong biological influence. Knowledge of the origins of submarine phosphorites has been greatly aided by studies of recent seafloor deposits. There are a variety of authigenic silicate minerals in deep-sea sediments. The more abundant zeolites and clays such as smectite form mainly from diagenetic alteration of metastable volcanic glasses, but many other reaction paths are known and suggested.

Ferromanganese Deposits Marine ferromanganese deposits are complex assemblages of authigenic minerals and detrital components. They are broadly characterized by shape and by environment of formation as deep-sea nodules and crusts. Ferromanganese nodules are commonly found on the abyssal seafloor at and near the top of the underlying pelagic sediment cover. Crusts are usually found on the steep slopes of islands and seamounts but are also found with the pelagic sediment on the deep-sea floor where sedimentation is extremely low. Occasionally, nodules may grow to coalesce into a pavement or crust of ferromanganese deposits. Conversely, crusts can grow around small fragments of the

258

Deep-sea ferromanganese nodules are classified as three major types on the basis of their origin: hydrogenetic, early diagenetic, and mixed hydrogenetic–early diagenetic. In this same scheme, crusts are described as primarily hydrogenetic, with comparatively minor diagenetic and hydrothermal occurrences (Figure 1). Hydrogenetic nodules obtain their ferromanganese oxides as precipitates from sea water. However, they are rarely if ever pure and also contain a variety of detrital minerals and biogenic debris from the pelagic sediment and fallout from the overlying water column and atmosphere. Early diagenetic nodules obtain their ferromanganese oxides from metals dissolved from the sediment and transported within the pore waters of the upper, peneliquid sediment layer (Figure 1). Because manganese and minor metals such as nickel and copper are more readily mobile in pore waters under suboxic seafloor conditions, diagenetic nodules are enriched in these metals relative to concentrations of iron and cobalt. The hybrid or mixed type nodule has both hydrogenetic and diagenetic components. These nodules are usually found at the sediment–water interface where their outer layers record their most recent position. The upper, relatively smooth hydrogenetic top is composed of iron- and cobalt-rich ferromanganese oxides from the bottom water and the lower, relatively bumpy or ‘botroyidal’ diagenetic bottom is composed of manganese-, nickel-, and copper-rich ferromanganese oxides from the sediment. Attempts have been made to classify nodules on the basis of morphology and to relate their occurrence to the composition of the underlying sediment. Such field classifications were made to predict metal enrichment and genesis, but reliance upon morphological data for other than abundance information from either seafloor photography or remote acoustic imaging has not yet been widely practiced in exploration. Nodule abundance appears to correlate more strongly with siliceous ooze than with red clays, and both siliceous ooze and red clays have higher nodule abundance than does carbonate ooze. Ferromanganese nodules and crusts are mainly composed of three types of manganese oxide minerals: vernadite or d-MnO2; birnessite; and todorokite. These


AUTHIGENIC DEPOSITS

259

Ocean water

Hydrogenetic crusts

Hydrogenetic nodules Range of calcite compensation depth Mixed-type nodules

Early diagenetic nodules

Precipitation of colloidal particles from near-bottom sea water Upward migration of dissolved 2+

2+

metals such as Mn , Cu , 2+ 2+ 2+ Ni , Zn , Mg under pore water conditions

Basaltic substrate rocks

Peneliquid sediment layer

Consolidated pelagic sediment

Figure 1 Schematic figure of the principal types of ferromanganese nodules (diameters vary from 1 to 15 cm) and crusts (thicknesses vary from o1 to 10 cm) distributed in pelagic sediments and on seamounts and island slopes. (After Halbach P (1986) Processes controlling the heavy metal distribution in Pacific ferromanganese nodules and crusts. Geologische Rundschau 75: 235–247.)

minerals are listed in order of decreasing Eh or oxidation potential of the formation environment, and increasing crystal order. In vernadite, all manganese occurs as Mn4þ, which allows for crystal lattice substitution of metals such as Co3þ and Pb4þ. Todorokite contains some manganese as Mn2þ, which allows for crystal lattice substitution of metals such as Cu2þ and Ni2þ. Birnessite is an intermediate form. Iron oxides are primarily found as closely intermingled X-rayamorphous FeOOH and occasionally as goethite or lepidocrocite. Other minerals include detrital quartz, plagioclase feldspar, clay minerals such as smectites, authigenic zeolites, apatite, barite, and calcite. The global distribution of ferromanganese nodules on the seafloor (Figure 2) largely reflects the inverse distribution of thickest sediment cover. Nodules, which accumulate at rates of several millimeters per million years, cannot grow where sediments are accumulating at rates that are greater than 103 times faster (several centimeters per thousand years). Otherwise, rapid burial and dissolution would be their fate. Paradoxically, nodules are found to lie atop sediments that are accumulating at rates up to 103 times faster; this phenomenon has been explained by bottom current winnowing and gentle

bumping or tilting of nodules during the burrowing activities of benthic organisms. Nodules are truly pelagic deposits and are rarely found near the continents or near islands where rapid hemipelagic sedimentation occurs. Exceptions are found on the Blake Plateau off eastern Florida, within the Drake Passage off South America, and near the Cape of Good Hope off Southern Africa, where bottom currents are strong and sweep sediments from these areas (Figure 2). Highly concentrated nodule fields occur between the Clarion and Clipperton fracture zones just north of the equatorial high productivity zone in the Eastern Pacific, just south of this zone in the Indian Ocean, and near the Antarctic Convergence in the Southeast Pacific. Source of major metals The sources of the major metals in ferromanganese nodules and crusts – iron and manganese – are commonly described as products of the weathering of the continents, seafloor hydrothermal activity, and sediment diagenesis. Other than a small but distinct cosmogenic component (see Ferromanganese Crusts below), the iron and manganese must ultimately come from the Earth’s crust and mantle. The


260

AUTHIGENIC DEPOSITS

Areas of nodule coverage

Areas where nodules are particularly abundant

Figure 2 Distribution of ferromanganese nodule abundance on the world ocean floor. (After Cronan, 1980.)

fraction coming from the continents is more geologically recycled than that from the ocean crust, and diagenesis in this regard is clearly not an ultimate source but a process of further recycling of these metals within the sediment column. The relative importance of continental weathering versus seafloor hydrothermal activity as a source of iron and manganese to pelagic sediments and marine ferromanganese deposits has been debated since the 1890s. Early work by able geochemists on both sides of this issue was hampered by a lack of convincing evidence. For example, on the basis of early distribution maps of manganese and other trace metals in surface Atlantic sediments and evidence for iron and manganese enrichments in atmospheric dust, it was suggested that continental weathering and atmospheric transport was an important pathway of these metals to the ocean. However, the manganese distribution in Atlantic sediments can more readily be explained by hydrothermal enrichments from hot springs along the Mid-Atlantic Ridge, unknown at that time but well-known today. Likewise, early arguments for extensive hydrothermal iron and manganese enrichments from volcanic hot springs located along the mid-ocean ridges were compelling, but were not confirmed until the discovery of hot springs there in 1977. Water column trace-metal analytical techniques have improved to the point that a very convincing argument can be made for continental weathering and atmospheric transport as a path for at least some fraction of the iron and manganese in marine ferromanganese deposits. Another approach, which additionally answers the old

question of the relative contributions of these two sources to ferromanganese deposits at any given location of the seafloor, was made for the Pacific (Figure 3) using a ‘Co-chronometer’. This dating method is based upon the inverse relationship to growth rate of some minor metals’ concentrations in hydrogenetic ferromanganese crust deposits, more painstakingly measured by radiometric techniques such as U-series (excess 230Th) or 10Be dating. Using the relatively easy Co-chronometer, hundreds of ferromanganese deposits were measured to construct the detailed map of ferromanganese crust growth rates for the Pacific in Figure 3. The highest growth rates are associated with active spreading along the East Pacific Rise, Juan de Fuca Ridge, and Galapagos Rift, and with submarine arc volcanism in the Mariana Island arc. The deposits in these areas therefore receive most of their iron and manganese from hydrothermal sources. The slowest crust growth rates and highest cobalt enrichments are located near the Mid-Pacific Mountains seamount province west of Hawaii, far from active hot springs and the continents. Source of minor metals Relative to sea water and the Earth’s crust, ferromanganese nodules are enriched in nickel, copper, cobalt, and a host of other minor metals such as the platinum-group elements (PGE) and the rare-earth elements (REE). Concentrations of nickel and copper reach up to 2 wt% in ‘high-grade’ nodules from fields between the Clarion and Clipperton fracture zones, with cobalt concentrations approaching 0.5 wt%. Broadly, nodules from the Pacific appear to be more enriched


AUTHIGENIC DEPOSITS

120˚

140˚

160˚

180˚

160˚

140˚

120˚

100˚

80˚

R (mm/Myr) 60˚

60˚

40˚

40˚

20˚

20˚

0˚

0˚

20˚

20˚

40˚

40˚

60˚

60˚

261

inverse relationship of depth of copper and nickel concentration in Pacific nodules with the sedimentary concentration of calcium carbonate tests from planktonic organisms, with deeper sediment deposits containing lesser CaCO3 because of increased dissolution. Comparative studies of copper, nickel, and cobalt in plankton and the labile or easily mobilized fraction of pelagic sediment and associated deep-sea nodules has also suggested that diagenesis of organic matter in the sediments leads to an enrichment of these metals in the nodules. All told, scavenging by organisms in the surface waters enriches the particle rain in minor metals that originally enter the world’s oceans by winds and rivers, and further concentration of these metals occurs within the surface sediments during suboxic diagenesis, with ferromanganese nodules as the final metals depository. Ferromanganese Crusts

120˚

140˚

160˚

180˚

160˚

140˚

120˚

100˚

80˚

Figure 3 Distribution of ferromanganese crust accumulation rates in the Pacific Ocean. (After Manheim FT and Lane-Bostwick CM (1988) Cobalt in ferromanganese crusts as a monitor of hydrothermal discharge on the Pacific sea floor. Nature 335, 59–62.)

in these metals than those from the Indian and Atlantic Oceans, but metal concentrations can vary widely within basins and nodule fields. For at least the better-studied Pacific, concentrations of nickel and copper show good correlation to the nodule Mn/Fe ratio, suggesting that manganese-rich, todorokitecontaining diagenetic nodules are the major depositories of these metals (Figure 4A, B). There is also a regional trend of nickel and copper enrichments and Mn/Fe ratios in the Pacific that is highest in the equatorial NE Pacific, the Peru Basin, and the SE Pacific near the Antarctic Convergence. The distribution of nodule cobalt concentrations shows little correlation to the Mn/Fe ratio, however, being enriched instead near the Mid-Pacific Mountains and in the South Central Pacific (Figure 4C). The ultimate source for the minor metals in ferromanganese nodules is again the Earth’s crust and mantle, but with minor metals the case for continental weathering versus seafloor hydrothermal activity is stronger. First, hydrothermal ferromanganese crusts, like many rapidly accumulating diagenetic crusts, are known to have very low concentrations of minor metals. Second, studies of copper and nickel distributions in pelagic surface sediments suggest that the highest concentrations are nearest to regions of high surface productivity, implying that plankton are involved in the enrichment process. There is an

Crust deposits differ from nodules in form, occurrence, and composition. Hydrogenetic crusts are principally composed of vernadite with iron oxides and minor detrital mineral and carbonate fluorapatite (CFA) contaminates. They grow at extremely slow rates from less than 1 mm to tens of millimeters per million years. Because the cobalt flux is for the most part invariant throughout the world’s oceans, cobalt concentrations approach 2 wt% on comparatively shallow seamount slopes between 800 and 2000 m water depth within the Western Equatorial Pacific (Figure 3 – compare with Figure 4C) where crust growth rates are slowest. Above approximately 800 m dilution and coverage of crustal pavements by principally carbonate-rich sediments limits their growth, and below 2000 m incorporation of increasing fluxes of seamount and wind-blown detritus and ferromanganese oxides causes crust growth rates to increase, effectively decreasing the cobalt concentration. The oxygen minimum zone (OMZ) presently intersects seamounts and island slopes in the Central Pacific at between 500 and 1500 m water depth. Although not presently anoxic enough to inhibit manganese oxide precipitation in this area, the OMZ was probably more intense in the past, thereby providing a mechanism for both transport of dissolved Mn2þ and inhibition of crust growth. Cobalt concentrations also show a general trend of decreasing values within older crustal layers, suggesting that past seafloor conditions favored increased manganese fluxes and growth. Relative to nodules, crusts are on average enriched in Fe, Ca, P, Ti, Pb, Ce, As, and Pt, as well as cobalt, and are depleted in Si, Al, Ni, Cu, and Zn. Manganese concentrations are similar. The relative enrichments of calcium and phosphorus in crusts reflect the more widespread incorporation of CFA in seamount crusts


262

AUTHIGENIC DEPOSITS

(B)

(A)

>5 2_ 5

> 1% 0.5−1%

1_ 2

0.25−0.5%

0.50% 0.25−0.50% < 0.25% Figure 4 (A) Distribution of Mn/Fe ratios in ferromanganese nodules from the Pacific. (B) Distribution of nickel concentrations in ferromanganese nodules from the Pacific. Copper shows a similar distribution. (C) Distribution of cobalt concentrations in ferromanganese nodules from the Pacific. (After Calvert SE (1978) In: Sea Floor Development: Moving into Deeper Water. London: The Royal Society.)

(see Phosphorites below), whereas the higher silicon and aluminum contents of nodules reflect their origin within loose, aluminosilicate-rich sediments of the abyssal seafloor. The mechanism for cobalt enrichment is oxidative scavenging, whereby dissolved Co2þ is oxidized to Co3þ on the manganese oxide surface. This mechanism also explains the relative enrichments of cerium, and possibly some of the lead and titanium enrichments in crusts. Lower concentrations of iron (and arsenic, an oxyanion in sea water scavenged by positively charged FeOOH; MnO2 surfaces are negatively charged at sea water pH) in nodules reflect a

greater diagenetic component that favors mobilization of manganese over iron. Because of their extremely slow growth in areas relatively free of detrital input, crusts also accumulate large amounts of cosmogenic debris. Enrichments of platinum and PGE metals can be partially explained by this source, but several lines of geochemical evidence indicate that the majority of platinum and the other PGE (Ir, Os, Pd, Rh, Ru) are scavenged from sea water. Oxidation and reduction mechanisms have been proposed for both incorporation and postdepositional remobilization of the PGE in crusts and nodules.


AUTHIGENIC DEPOSITS

Phosphorites Submarine phosphorites are principally composed of CFA or francolite with usually large admixtures of detrital impurities from the sediments in which they form or replace during authigenesis. Concentrations of P2O5 commonly range from 5 to 28 wt%, reaching values of 35 wt%, with minor element enrichments of strontium, yttrium, and the REE. There are three principal types of phosphorite deposits in the world’s oceans: offshore or continental shelf-slope nodules and concretions; massive insular deposits originating from sea bird guano; and ‘seamount’ phosphorites that often associate with ferromanganese crusts as layers, veins, impregnations, and substrate material (Figure 5). Submarine phosphorites were originally thought to be fossil deposits no younger than Miocene age, like those marine phosphorites mined on land, but were subsequently found to be actively forming in shelf-slope areas off Western South America, Southwest Africa, and Australia. Studies of the deposits off Peru and Chile have demonstrated their contemporaneous origin near anoxic sediments where the OMZ impacts the continental shelf-slope there at between 200 and 400 m (Figure 6). Upwelling currents bring waters rich in dissolved PO4 onto the continental slope and shelf, increasing surface productivity and enhancing the reduction potential in sediments underlying the OMZ. Bacterial consumption of organic matter within the sediment releases HPO4 2 that can combine with dissolved Ca2þ, F and CO3 2 in pore waters to form authigenic precipitates of CFA. This process is thought to occur in sediments near the boundaries of the OMZ where the hydrogen ion

Figure 5 Photograph of a ferromanganese crust on a phosphorite substrate recovered from the summit of Schumann seamount, north of Kauai, Hawaii. Note the inclusions of altered basalt clasts from the seamount within the phosphorite, which has likely replaced limestone. (From G. McMurtry and D. L. VonderHaar, unpublished data.)

263

concentrations (low pH) produced during bacterial consumption and in CFA precipitation are diluted by relatively alkaline (high pH) sea waters. Associated sediment diagenetic reactions involving dolomite formation and Mg2þ exchange in clays reduce the concentration of apatite precipitation inhibiting Mg2þ in pore waters, aiding CFA formation. The distribution of phosphorites throughout the Phanerozoic rock record is highly irregular, correlating broadly with periods of warm climate and higher sea level. Off South America and South-west Africa, Useries dating of the deposits has demonstrated their correlation with warm, interglacial periods back to 150 000 years before present. Increased chemical weathering of the continents during warm periods may provide enhanced PO4 fixation into hemipelagic

Figure 6 Schematic figure of the western South American shelf and slope, showing contours of dissolved phosphate and oxygen in the bottom waters and the locations of aerobic and anaerobic sediments, phosphorite and Recent phosphorite deposits. (After Burnett WC, Veeh HH and Soutar A (1980) U-series, oceanographic and sedimentary evidence in support of recent formation of phosphate nodules off Peru. In: Marine Phosphorites, SEPM Special Publ. no. 29, pp. 61–72.)


264

AUTHIGENIC DEPOSITS

sediments. Concurrent encroachment of the OMZ onto the continental shelf-slope during marine transgressions increases the extent of organic matter preservation. When combined, these effects are thought to promote initial phosphorite formation. Subsequent marine regressions may enhance the initial enrichment by winnowing detritus from older phosphatic sediments, producing deposits of economic value. Insular phosphorites are relatively easily understood deposits from guano in the nesting areas of ancient sea birds that created the ore deposit by transporting and further biologically concentrating scarce dissolved phosphate in surface sea waters within the flight radius of the island. The island of Nauru in the central Equatorial Pacific is a prime example. Although some are presently submerged, these deposits have all formed subaerially and are no older than Tertiary. Comparatively less is known about the formation of seamount phosphorites. Substrate samples show cryptocrystalline CFA matrix replacement of preexisting limestones that often contain altered basalt breccia and other evidence of high-energy deposition upon the summits and upper slopes of Cretaceous seamounts in the Pacific (Figure 5). Within ferromanganese crusts, these deposits show the effects of lowered redox conditions (e.g., platinum remobilization) with impregnation and replacement of older ferromanganese oxides. Strontium and oxygen isotope-derived ages of CFA formation appear to center on the Eocene–Oligocene (36 Ma) and Oligocene– Miocene (24 Ma) boundaries, with some evidence for a minor Middle Miocene event at about 15 Ma. Paleotracking of the Pacific Cretaceous seamounts shows that many were close to the equatorial high productivity belt during the late Cretaceous and Paleocene–Eocene periods, where their then-shallower summits and slopes could have intersected the OMZ in a region of equatorial upwelling. Later depositional episodes would require a greatly expanded and more intense OMZ than the present one.

mineral in hydrothermal deposits on land and has more recently been found as a principal component of moderate-temperature hydrothermal chimneys on midocean ridges and volcanically active seamounts. Early work on the distribution of marine barite in deep-sea sediments had difficulty distinguishing the relative importance of hydrothermal sources from continental weathering as cycled through the marine biosphere. Distribution maps of marine barite in Pacific deep-sea sediments and the more quantitative barium accumulation rate (Figure 7) show both an association with the equatorial high productivity zone and the East Pacific Rise–Bauer Basin areas that are heavily influenced by metalliferous hydrothermal deposition. The key question is how much of the relatively dispersed, finegrained (usuallyo2 mm) marine barite particles in deep-sea sediments result from hydrothermal plume fallout and bottom current redistribution of ridge-crest metalliferous sediments versus those carried to the seafloor in the particle rain from surface productivity. The association of marine barite with organic matter complicates the interpretation of occurrence of increased barite deposition found along midocean ridges because greater preservation of organic matter also occurs with increased carbonate sedimentation above the calcite compensation depth. Additionally, basin-scale dispersal of hydrothermal particles appears limited, especially for the relatively dense barite. Studies of marine barite saturation show that barite is below saturation in the water column but rapidly approaches saturation in the pore waters of deep-sea sediments. Discrete barite particles are nevertheless found in the microenvironments of suspended

Marine Barite and Authigenic Silicates Marine Barite

Barite (BaSO4) is a widespread mineral in deep-sea sediments, varying between 1 and 10 wt% on a carbonate-free basis. It is the predominant barium phase in the ocean. BaSO4 is known to compose a solid solution series with SrSO4 as celestobarite in the skeletal portions of some marine organisms (i.e., the Xenophyophoria) and is often found in association with marine organic matter, such as in suspended particles and fecal pellets. Barite is also a well-known gangue

Figure 7 Distribution of barium accumulation rates (units of mg cm2 per 1000 years) in the Pacific sediments. (After Bostro¨m K, Joensuu O, Moore C et al. (1973), Geochemistry of barium in pelagic sediments. Lithos 6: 159–174.)


AUTHIGENIC DEPOSITS

(A)

(B)

(C)

(D)

265

Figure 8 Photograph of barite crystals inside a ferromanganese nodule. (A) Scanning electron micrograph of foraminifera shell walls that have been replaced by barite. (B) Barium elemental X-ray map produced by EMP. (C) Magnification of (A) showing individual barite crystals. (D) Elemental spectrum of the barite crystals: S Ka at 2300 eV; Ba La, b, g at 4460 eV. (After Lalou C, Brichet E, Poupeau G, Romany P and Jehanno C (1979) Growth rates and possible age of a North Pacific manganese nodule. In Bischoff JL and Piper DZ (eds) Marine Geology and Oceanography of the Central Pacific Manganese Nodule Province, pp. 815–834. New York and London: Plenum Press.)

biogenic matter and in fecal pellets that rapidly fall to the seafloor; their dissolution on the seafloor and in transit is likely the dominant control on the barium concentration in the deep ocean. Within the sediment and in ferromanganese deposits, barite can form authigenically as discrete particles and diagenetically as replacement of biogenic skeletal remains (Figure 8). Authigenic Silicates

Zeolites Low-temperature alteration of basalt and metastable volcanic glasses on pillow rims, in layers, and dispersed throughout deep-sea sediments can produce a variety of diagenetic alteration phases. (Low temperature is defined as ranging from modern ambient seafloor conditions, or near 01C to up to

1501C, the low-temperature metamorphism limit.) These include X-ray-amorphous palagonite, the zeolites phillipsite, clinoptilolite, analcite (plus several others of rarer occurrence), smectites, and authigenic K-feldspar (Table 1). Phillipsite is a hydrated potassium- and sodium-rich aluminosilicate that forms elongated and sometimes twinned crystals from 8 to 250 mm length with numerous inclusions that indicate rapid growth. Oxygen isotope values of 34% for phillipsite indicate formation in the marine environment at modern seafloor temperatures. Phillipsite appears metastable on geological timescales, becoming rarer beyond Cenozoic age. The mineral is found globally at the sediment–sea water interface and continues to grow within the sediment column until it dissolves and disappears in the deepest


266

AUTHIGENIC DEPOSITS

Table 1

Summary of observed and suggested diagenetic reactions in deep-sea sediments

Opal-A-quartz Basaltic glass-palagonite Volcanic glass-smectite minerals Volcanic glass-zeolites Basalt-smectite minerals Basalt þ H4SiO4-smectite minerals Basaltic glass þ H4SiO4-phillipsite Poorly cyrstalline smectite-phillipsite Andesitic and rhyolitic glass þ H4SiO4-clinoptilolite Andesitic and rhyolitic glass-smectite minerals Phillipsite þ H4SiO4-clinoptilolite Phillipsite þ smectite þ H4SiO4-clinoptilolite þ palygorskite þ Opal-A þ Al(OH) 4 þ K -clinoptilolite Opal-A þ Al(OH)4 þ Kþ-opal-CT þ clinoptilolite Plagioclase þ Kþ þ H4SiO4-K-feldspar þ Naþ þ Ca2þ -K-feldspar ?-albite Clinoptilolite þ Naþ-analcite þ Kþ þ quartz Clinoptilolite- K-feldspar þ quartz Clinoptilolite þ Naþ-analcite þ K-feldspar þ quartz H4SiO4 þ Mg2þ þ Al(OH) 4 - palygorskite H4SiO4 þ Mg2þ-sepiolite Volcanic glass þ Mg2þ þ H4SiO4-palygorskite Smectite þ Mg2þ þ H4SiO4- palygorskite Clinoptilolite þ palygorskite þ calcite- K-feldspar þ dolomite þ quartz Clinoptilolite þ sepiolite þ calcite- K-feldspar þ dolomite þ quartz Amorphous hydroxides (mainly Fe) þ H4SiO4 þ Mg2þ-Fe-montmorillonite Nontronite þ Mg2þ þ reduced sulfur-saponite þ FeS2 Dissolved silica adsorption by clay minerals Amorphous aluminosilicate reconstitution? Smectite-mixed-layer illite–smectite? After Kastner (1981).

drill holes (usually below 500 m depth). Marine clinoptilolite forms finer grained (o45 mm), platy crystals of relatively silicon-rich hydrated potassiumand sodium-rich aluminosilicate (high Si/Al ratio). Clinoptilolite is more frequently encountered at depths 4100 m in the sediment column and persists to great depths and geological age, suggesting that it is thermodynamically stable in the deep-sea. Analcite is a rarer, sodium-rich zeolite in deep-sea sediments that displays a general trend of increasing abundance with geological age which parallels that of clinoptilolite. Both clinoptilolite and analcite can form directly from alteration of more siliceous volcanic glass (andesite, rhyolite, from active island arcs and explosive continental volcanism), by diagenetic reaction of phillipsite and dissolved silica (clinoptilolite) or Na/K exchange with clinoptilolite (analcite). Palygorskite and sepiolite Authigenic clay minerals (o2 mm particle size) in the deep-sea include the fibrous minerals palygorskite and sepiolite and the smectite family of expandable phyllosilicates. Both palygorskite and sepiolite are rare in recent marine

sediments, occuring more often in Eocene and older sediments. Palygorskite is a hydrous silicate containing Mg, Al and Fe3þ, whereas sepiolite is almost a pure hydrous magnesium silicate. Pore water solutions of alkaline pH with high concentrations of dissolved silica and magnesium favor the formation of both minerals. Fine fibrous textures and overgrowths of siliceous tests and opal-CT attest to their authigenesis. Reactions range from the diagenetic alteration of (mainly silicic and intermediate) volcanic ash, either directly or indirectly via smectite with biogenic silica, formation upon magnesium release after conversion of biogenic opal-A to opal-CT, and reaction of biogenic silica tests with marine pore waters, including hypersaline brines enriched in magnesium (Table 1). Smectites Most clay minerals in deep-sea sediments are detrital phases from continental weathering. These minerals compose the bulk of the nonbiogenic sediments in the o2 mm fraction and include clay-sized quartz and the phyllosilicates kaolinite, illite, chlorite, and smectite. The smectite


AUTHIGENIC DEPOSITS

267

Figure 9 Distribution of montmorillonite (smectite) abundance in the o2 mm fraction of sediments in the world ocean, carbonate-free basis. (After Griffin JJ, Windom H and Goldberg ED (1968) The distribution of clay minerals in the world ocean. Deep-Sea Research 15: 433–459.)

group of expandable clay minerals includes these common end-members: magnesium-rich saponite, found mostly as a diagenetic or low-temperature metamorphic product of basalt; iron-rich nontronite, often found as a low-temperature (o1001C) hydrothermal deposit; and aluminumrich beidellite predominantly derived from volcanic ash alteration on land. The surface distribution of marine smectite in the world’s oceans shows broad areas, such as in the south-eastern Pacific where 470% of the mineral’s abundance is far from land and blankets the active East Pacific Rise spreading

center (Figure 9). Most of this smectite is iron-rich montmorillonite. Up to 50% of this mineral has been described as authigenically formed from the seafloor temperature alteration of volcanic glass, with the remaining 50% from detrital sources. Early oxygen isotope studies of deep-sea montmorillonite suggested that much of it formed pedogenically on land, but iron-rich montmorillonite is not typically found in windblown dust. More recent oxygen isotope work has suggested either a lowtemperature submarine hydrothermal origin for the iron-montmorillonite or formation from low-

Fe2O3

Galapagos nontronite

East Pacific Rise Fe-mont NE Pacific Fe-mont OCP ridge Fe-mont

Galapagos Fe-mont

Detrital Al-mont /beidellite

MgO

Al2O3

Figure 10 Triangle plot of octahedral Mg–Al–Fe composition for marine smectites. (After McMurtry GM, Wang CH and Yeh HW (1983) Chemical and isotopic investigations into the origin of clay minerals from the Galapagos hydrothermal mounds field. Geochimica et Cosmochimica Acta 47: 475–489.)


268

AUTHIGENIC DEPOSITS

temperature hydrothermal alteration of ash within subaerial volcanoes that is periodically deposited in the ocean by explosive volcanism. Nontronite and iron-montmorillinite have also been suggested to form at seafloor temperatures by reaction of hydrothermally derived iron oxides with biogenic silica. One of the difficulties in assigning a precise origin to this mineral class is the possibility of physical admixtures of truly authigenic iron-rich smectites with more aluminum-rich smectites derived from the continents (Figure 10).

Glossary Aluminosilicate A silicate containing aluminum in coordination with oxyhydroxides and/or in substitution for silicon in SiO4 tetrahedra. Authigenesis New origin. Process of formation of new minerals in place. Detrital Formed from detritus, usually from rocks, minerals, or sediments from elsewhere than the depositional site. Diagenesis Changed origin. Recombination or rearrangement of a mineral that results in a new mineral, usually postdepositionally. Hemipelagic Deep-sea sediment that accumulates near the continental margin, so that the sediment contains abundant continentally derived material and rates of sedimentation are high. Hypersaline Excessive salinity, much greater than the normal salinity of sea water. Metalliferous Metal bearing, usually enriched toward economic extraction of the metals. Metastable State of a phase that is stable toward small disturbance, but is capable of reaction if sufficiently disturbed. Pedogenesis Soil origin. Mineral formation within the soil. Pelagic Open ocean environment. A marine sediment with that fraction derived from the continents indicating deposition from a dilute suspension distributed throughout deep-sea water. Phyllosilicate Layered or sheet silicate mineral, formed by sharing three of the four oxygens in neighboring silicon tetrahedra. Plankton Aquatic organisms that drift, or swim weakly. Can be either plants (phytoplankton) or animals (zooplankton). Redox Abbreviation for reduction–oxidation, usually expressed as a potential. Seamount Underwater mountain, 1000 m or higher elevation from seafloor base. Morphology may be peaked or flat-topped, with the latter called guyot.

Suboxic Condition lacking free oxygen, but not extremely reducing. Zeolite Any of the minerals of the zeolite group. Aluminosilicate minerals with an open framework structure that allows for easily reversible hydration, gas adsorption, and either cation or anion exchange.

See also Aeolian Inputs. Clay Mineralogy. Hydrothermal Vent Deposits. Manganese Nodules. Mineral Extraction, Authigenic Minerals. Platinum Group Elements and their Isotopes in the Ocean. Pore Water Chemistry. Rare Earth Elements and their Isotopes in the Ocean. River Inputs. Sediment Chronologies. Sedimentary Record, Reconstruction of Productivity from the. Tracers of Ocean Productivity. Transition Metals and Heavy Metal Speciation. Uranium-Thorium Decay Series in the Oceans Overview.

Further Reading Bentor YK (ed.) (1980) Marine Phosphorites; a Symposium. Oklahoma: SEPM Special Publication no. 29. Burns RG and Burns VM (1981) Authigenic oxides. The Sea, vol. 7, pp. 875--914. New York: Wiley. Chamley H (ed.) (1989) Clay Sedimentology. Berlin: Springer-Verlag. Cronan DS (1974) Authigenic minerals in deep-sea sediments. In: Goldberg ED (ed.) The Sea, vol. 5, pp. 491--525. New York: Wiley. Cronan DS (ed.) (1980) Underwater Minerals. London: Academic Press. Cronan DS (ed.) (2000) Handbook of Marine Mineral Deposits. Boca Raton, FL: CRC Press. Glasby GP (ed.) (1977) Marine Manganese Deposits. Elsevier Oceanography Series. Amsterdam: Elsevier. Glenn CR, Pre´vot-Lucas L, and Lucas J (eds.) (2000) Marine Authigenesis: from Global to Microbial. Oklahoma: SEPM Special publication no. 66. Halbach P, Friedrich G, and von Stackelberg U (eds.) (1988) The Manganese Nodule Belt of the Pacific Ocean: Geological Environment, Nodule Formation, and Mining Aspects. Stuttgart: F. Enke Verlag. Kastner M (1981) Authigenic silicates in deep-sea sediments: formation and diagenesis. In: Emiliani C (ed.) The Sea, vol. 7, pp. 915--980. New York: Wiley. Manheim FT (1986) Marine cobalt resources. Science 232: 600--608. Margolis SV and Burns RG (1976) Pacific deep-sea manganese nodules: their distribution, composition, and origin. Annual Review of Earth and Planetary Science 4: 229--263.


BACTERIOPLANKTON H. W. Ducklow, The College of William and Mary, Gloucester Point, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 217–224, & 2001, Elsevier Ltd.

Introduction Marine bacteria, unicellular prokaryotic plankton usually less than 0.5–1 mm in their longest dimension, are the smallest autonomous organisms in the sea – or perhaps in the biosphere. The nature of their roles in marine food webs and the difficulty of studying them both stem from their small size. A modern paradigm for bacterioplankton ecology was integrated into oceanography only following development of modern epifluorescence microscopy and the application of new radioisotopic tracer techniques in the late 1970s. It was not until a decade later, with the use of modern genomic techniques, that their identity and taxonomy began to be understood at all. Thus we are still in the process of constructing a realistic picture of marine bacterial ecology, consistent with knowledge of evolution, plankton dynamics, food web theory, and biogeochemistry. The lack of bacterioplankton compartments in most numerical models of plankton ecology testifies to out current level of ignorance. Nevertheless, much is now well known that was just beginning to be guessed in the 1980s. Bacterioplankton are important in marine food webs and biogeochemical cycles because they are the principal agents of dissolved organic matter (DOM) utilization and oxidation in the sea. All organisms liberate DOM through a variety of physiological processes, and additional DOM is released when zooplankton fecal pellets and other forms of organic detritus dissolve and decay. By recovering the released DOM, which would otherwise accumulate, bacterioplankton initiate the microbial loop, a complicated suite of organisms and processes based on the flow of detrital-based energy through the food web. The flows of energy and materials through the microbial loop can rival or surpass those flows passing through traditional phytoplankton-grazerbased food chains. For further information on the topics summarized here, the reader may consult the Further Reading, especially the recent book edited by Kirchman.

Identity and Taxonomy Most bacterial species cannot be cultivated in the laboratory and, until the development of cultureindependent genomic methods, the identity of over 90% of bacterial cells enumerated under the microscope was unknown. Only those few cells capable of forming colonies on solid media (agar plates) could be identified by classical bacteriological techniques. However, since the application of molecular genomic methods to sea water samples in the mid-1980s, our understanding of marine bacterial systematics and evolution has undergone a profound revolution. In this approach, plankton samples including bacterioplankton cells are collected and lysed to yield a mixture of DNA strands reflecting the genetic composition of the original assemblage. Then individual genes on the DNA molecules can be cloned and amplified via the polymerase chain reaction (PCR) for further analysis. Theoretically, any gene complex can be cloned, and several major groups of genes have been studied to date – for example, genes controlling specific biogeochemical transformations like ammonium oxidation, nitrogen fixation, sulfate reduction, and even oxidation of xenobiotic pollutant molecules. The most useful and widely studied genes for elucidating evolutionary relationships among bacterioplankton have been the genes coding for small subunit ribosomal RNA (SSU rRNA), because they evolve relatively slowly and their characters have been conserved across all life forms during the course of evolution. By sequencing the base pairs making up individual SSU rRNA molecules, the similarity of different genes can be established with great sensitivity. To date, nearly 1000 individual microbial SSU rRNA genes have been cloned and sequenced, yielding an entirely new picture of the composition of marine communities. The most important aspect of our understanding is that what we term ‘bacterioplankton’ really consists of two of the fundamental domains of life: the Bacteria and the Archaea (Figure 1). Domain Archaea is a group of microbial organisms with unique genetic, ultrastructural, and physiological characters that are about as different, genetically, from the Bacteria as either group is from higher life forms. Members of the Archaea may be typified by organisms from extreme habitats including anaerobic environments, hot springs, and salt lakes, but marine archaeal groups I and II are common in sea water. They make up about 10% of the microbial plankton in the


269

270

BACTERIOPLANKTON

Cytophaga group

lopha

i de p

s 24 r u R3 vo SA erio B/ ct ba

me

tact

Bacteria

s

ne

e og

s

iae on m c io eu um su rad os hilus pn n i er us a t il sp nop ac cc ph m ob oco ydo obiu s lim r r e b n m Fi Dei hla omic myc leyi pallida C ruc cto sta era n a r a Ve Plairellul osph P Is m lla" inutu rix parvice m m h icrot obiu Atop idatus "M Cand Marine Actinobacteria Rubrobacter xylanophilus Nitrospira marina SAR2 Therm 02 omicro bium r Chlo oseum r o H f l The e erp eto xus aura rm s oto Geo ntiac ga tog iphon us aur Aq ma a su a uife ritim bte ntia rran xp cus yro a ea ph ilus n ci

n ra du

0.10 substitutions per nucleotide position

Mag

neto

spir

illum

mag

neto

SA

R1

rio

lus

16 icum Azospirillu m brasilen se

tal lire m du dof ethylo c era x fe troph ens rme us nta ns SAR86 Bathymodiolus thermophilus gi ll sym. Oceanospirillum linum macleodii Alteromonas um SAR11 linar o sa us i r b i t dov sula Rho cap R83 us r e A ng act / S lo dob ter ter Rho c c a ob oba se r o h R yt Er Rho

Proteobacteria

ou Gr

ter

phi

b vi lo el

ac

Bd

ob

Lactococcus lactis Bacillus subtilis Mar Gl Sy ine Picoph oe nec ytopla ob ho nkton coc ac ter c vio us PC lac C6 301 eu s

AR406 Marine Group A/S rme m vibriofo Chlorobiu ica ga lyt lis agi s fr

Cellu

tero Bac e in

ar M

Ge Me thy lo

Cyanobacteria

sm a oc Arc ido ha ea ph Gr ilu Hal ou m pI oba I cte rium hal Halofe obi u rax vo lc anii m Methanosp irillum hun gatei Methanosarcina barkeri

pla

mo

er Th

Su Py Desu lfolobus lf s ro dic urococ olfataric us cus tiu mob m ilis oc cu ltu m

x na te ns us de te pI en p ro rou op filum aG m ae er rmo ch Ar Th The

ii asch jann tae ler s u l c e m o ococ s v s c cu m han ccu occu mici hicu Met o c no o c or p tha erm m f otro Me h T teriu aut o ac ob erm n h ha m t et iu M er ct a ob an h et M

Archaea

Figure 1 Dendrogram showing relationships among the most widespread SSU rRNA gene clusters among the marine prokaryotes (the ‘bacterioplankton’). (Modified after Giovannoni SJ, in Kirchman (2000).)

surface waters of the oceans, and are relatively more numerous at greater depths, where they approach about half the total abundance. Since most of these organisms are known only from their RNA genes

and have never been cultured, their physiology and roles in the plankton are almost entirely unknown. Domain Bacteria contains all the familiar, culturable eubacterial groups and also a large number of


BACTERIOPLANKTON

1000 _1

Knowledge of the nutrition and physiology of naturally occurring bacterioplankton as a functional group in the sea is based partly on laboratory study of individual species in pure culture, but mostly on sea water culture experiments. Traditional laboratory investigations show that bacteria can only utilize small-molecular-weight compounds less than B500 Daltons. Larger polymeric substances and particles must first be hydrolyzed by extracellular enzymes. In the sea water culture approach, samples with natural bacterioplankton assemblages are incubated for suitable periods (usually hours to a few days) while bacterial abundance is monitored, the utilization of various compounds with 14C- or 3Hlabelled radiotracers is estimated, and the net production or loss of metabolites like oxygen, CO2, and inorganic nutrients is measured. Such experiments, combined with size-fractionation using polycarbonate filters with precise and uniform pores of various diameter (0.2–10 mm), revealed that over 90% of added organic radiotracers are utilized by the smallest size fractions (o1 mm). Bacteria are overwhelmingly the sink for DOM in all habitats studied to date. Nutrient limitation of bacterial growth can

_2

Nutrition and Physiology

be identified by adding various compounds (e.g., ammonium, phosphate, or iron salts; monosaccharides and amino acids) singly or in combination to experimental treatments and comparing growth responses to controls. Using this approach, it has been learned that bacteria are effective competitors with phytoplankton for inorganic nutrients, including iron, which bacteria can mobilize by producing ironbinding organic complexes called siderophores. In general, bacterial growth in the sea, from estuaries to the central gyres, tends to be limited by organic matter. Sea water cultures most often respond to additions of sugars and amino acids, with the response sometimes enhanced if inorganic nutrients (including iron) are also added. At larger scales, the ultimate dependence of bacteria on organic matter supply is indicated by significant correlations between bacterial standing stocks or production (see below) and primary production (PP) across habitats (Figure 2). At withinhabitat scales and shorter timescales, significant relationships are less common, indicating time lags between organic matter production and its conversion by bacteria. Such uncoupling of organic matter production and consumption is also shown by transient accumulations of DOM in the upper ocean, where production processes tend to exceed utilization. It is not yet understood why DOM accumulates. Some fraction might be inherently refractory or rendered so by ultraviolet radiation or chemical condensation reactions in sea water. Deep ocean

Bacterial production (mg C m d )

unculturable, previously unknown groups. The main culturable groups include members of the Proteobacteria, marine oxygenic, phototrophic Cyanobacteria, and several other major groups including methylotrophs, planctomycetes, and the Cytophaga– Flavobacterium–Bacteroides group. But the most abundant genes recovered so far are not similar to those of the known culturable species. These include the most ubiquitous of all groups yet recovered, the SAR-11 cluster of the alpha Proteobacteria, which have been recovered from every bacterial clone library yet isolated. It appears to be the most widely distributed and successful of the Bacteria. The photosynthetic Cyanobacteria, including Synecchococcus spp. and the unicellular prochlorophytes, are functionally phytoplankton and they dominate the primary producer populations in the open sea, and at times in coastal and even estuarine regimes. They are treated elsewhere in this encyclopedia, so our discussion here is limited to heterotrophic forms of Bacteria and to the planktonic Archaea, although we cannot specify what many (or most) of them do. Genomic techniques are now being used to investigate bacterial and archaeal species succession during oceanographic events over various timescales, much as phytoplankton and higher organism successions have been observed for a century or more.

271

100

10

1 100

1000

10 000 _2

_1

Primary production (mg C m d ) Figure 2 Bacterial production plotted against primary production for the euphotic zone in several major ocean regimes or provinces. The overall data set has a significant regression, but the individual regions do not. J, Sargasso Sea; K, Arabian Sea; B, equatorial Pacific; n, equatorial Pacific.


272

BACTERIOPLANKTON

DOM has a turnover time of centuries to millennia, and seems to become labile (vulnerable to bacterial attack) when the ocean thermohaline circulation returns it to the illuminated surface layer. Alternatively, bacterial utilization of marine DOM, which generally has a high C:N ratio, might be limited by availability of inorganic nutrients. The latter hypothesis is supported by observations that DOM accumulation tends to be greater in the tropics and subtropics, where nitrate and phosphate are depleted in surface waters. The efficiency with which bacteria convert organic matter (usually expressed in carbon units) into biomass can be estimated by comparing the apparent utilization of individual compounds or bulk DOM with increases in biomass or with respiration. Respiration is usually measured by oxygen utilization but precise new analytical techniques for measuring carbon dioxide make CO2 production a preferable approach. Bacterial respiration (BR) is difficult to measure because water samples must first be passed through filters to remove other, larger respiring organisms, and because the resulting respiration rates are low, near the limits of detection of oxygen and CO2 analyses. It is also not easy to estimate bacterial biomass precisely (see below). The conversion efficiency or bacterial growth efficiency (BGE) is the quotient of net bacterial production (BP) and the DOM utilization: BGE ¼

BP BP ¼ DDOM BP þ BR

½1

Bacteria have rather uniform biomass C:N composition ratios of 4–6. Intuitively, it seems reasonable to expect that they would utilize substrates with high C:N ratios at lower efficiency. Enrichment cultures initiated from natural bacterial assemblages grow in sea water culture in the laboratory on added substances with efficiencies of 30–90%. The BGE is inversely related to the C:N ratio of the organic substrate if just a single compound is being utilized, but when a mixture of compounds is present, as is probably always the case in the environment, there is no discernible relationship between the chemical composition of the materials being used and the BGE. In the open ocean, BGE averages about 10–30%, a relatively low value that has important implications for our understanding and modeling of organic matter turnover and ocean metabolism. At larger scales, BGE appears to increase from B10% to 50% along an offshore-to-onshore gradient of increasing primary productivity, probably reflecting greater organic matter availability. This pattern has been used

to support an argument suggesting that in lake and oceanic systems with the lowest primary productivity, respiration exceeds production; that is, such oligotrophic systems might be net heterotrophic. This possibility has also been supported by results from careful light–dark bottle studies in which oxygen consumption exceeds production. This finding, however, is inconsistent with a large amount of geochemical evidence, for instance showing net oxygen production at the basin and seasonal to annual scale. Resolution of this debate probably rests on improved estimates of BGE. Pure culture, sea water culture, and the latest genomic studies indicate fundamental metabolic and genetic differences among different bacterial populations, which can generally be grouped into two broad classes based on organic matter utilization. Native marine bacteria capable of utilizing DOM at concentrations below 100 nmol l1, termed oligotrophs, cannot survive when DOM is greater than about 0.1–1 mmol l1. Copiotrophic bacteria found in some habitats with higher ambient DOM levels thrive on concentrations far exceeding this threshold. Observations that copiotrophs shrink and have impressive survival capability under severe starvation conditions (thousands of days to, apparently, centuries) led some investigators to suggest that the dominant native marine bacteria are starving (nongrowing) copiotrophs in a survival mode, awaiting episodes of nutrient enrichment. A variable fraction of the total population usually does appear to be dormant, as indicated by autoradiography, vital staining, and RNA probes, but the timescales of the transition from active growth to dormancy and back again are not well defined. Maintenance of dormant cells in a population depends on strong predator preferences for actively growing cells and prey selection against the nongrowing cells. Most oligotrophs so far isolated in the laboratory under stringent low-DOM conditions appear to be unrelated to known bacterial groups.

Bacterial Biomass, Growth, and Production The standing stock of bacteria is still most commonly assessed by epifluorescence microscopy, following staining of the cells with a fluorochrome dye. Flow cytometric determination is gradually taking over, and has several key advantages over microscopy: faster sample processing, improved precision, and discrimination of heterotrophic and phototrophic bacteria. There is a gradient in bacterial abundance proceeding from B1010 cells l1 in estuaries to 109


BACTERIOPLANKTON

cells l1 in productive ocean regimes and 108 cells l1 in the oligotrophic gyres (Figure 3). These horizontal gradients parallel gradients in primary production and organic matter fluxes, suggesting the overall importance of bottom-up controls on bacterial abundance. Chlorophyll a concentrations, indicative of phytoplankton biomass, vary somewhat more widely than bacterial abundance over basin to global scales, but within habitats, the variability of bacterial and phytoplankton biomass is about equal, reflecting the generally close coupling between the two groups and the similarity of removal processes (grazing, viral lysis, unspecified mortality) acting on them. It is more difficult to estimate bacterial biomass, because we cannot measure the mass (e.g., as carbon) directly, and have to convert estimates of cell volumes to carbon instead. The best estimates now indicate 7–15 1015 g C cell1 for oceanic cells and 15–25 1015 g C cell1 for the slightly larger cells found in coastal and estuarine habitats. Thus the biomass gradient is steeper than the abundance gradient because the cells are larger inshore. Table 1 shows data compiled from Chesapeake Bay and the Sargasso Sea off Bermuda, two well-studied sites that illustrate the contrasts in phytoplankton and bacterioplankton from a nutrient-rich estuary to the oligotrophic ocean gyres. Bacterial and phytoplankton biomass are much greater in the estuary, as

10

Cells l

_1

10

9

10

expected. Interestingly, assuming a mean euphotic zone depth of 1 m in the Bay and 140 m off Bermuda, we find that the standing stocks of bacteria in these euphotic zones are B10 and 50 mmol C m2 in the estuaries and open sea, respectively. The oceanic euphotic zone is somewhat more enriched in bacteria than the more productive estuaries. Bacterial and phytoplankton stocks are nearly equal in the open sea, but phytoplankton exceeds bacterial biomass inshore. Carbon from primary producers appears to be more efficiently stored in bacteria in oceanic systems compared to estuarine ones. Bacterial stocks in different environments can be assessed using the relationship Bmax ¼ F=m

10

94

136

30

37

27

162

H

aw ai i Sa rg as so R os s Se G a ul fS Eq tre am ua to ria lP ac N . C A he tla sa nt ic pe ak e Ba y

66

Figure 3 Bacterial abundance in the euphotic zone of several major ocean provinces. The box plots show the median, 10th, 25th, 75th and 90th centiles of the data. The number of samples is listed for each region. There is no statistical difference among the regions except for Chesapeake Bay.

½2

where Bmax is the carrying capacity in the absence of removal, F is the flux of utilizable organic matter to the bacteria, and m is their maintenance efficiency (the specific rate of utilization when all of F is used to meet cellular maintenance costs, with nothing left for growth). The problem is specifying values for F and m. The DOM flux can be evaluated by flow analysis and is about 20–50% of the net primary production (NPP) in most systems. Maintenance costs are poorly constrained and possibly very low if most cells are near a starvation state, but 0.01 d1 is a reasonable value for actively growing cells. Thus for the oligotrophic gyres where the latest NPP estimates are about 200–400 mg C m2 d1, we can calculate that Bmax should be about 4–8 109 cells l1, an order of magnitude greater than observed. Removal processes must maintain bacterial stocks considerably below their maximum carrying capacity. Bacteria convert preformed organic matter into biomass. This process is bacterial production, which can be expressed as the product of the biomass and the specific growth rate (m) BP ¼ dB=dt ¼ mB

8

273

½3

Like biomass, BP cannot be measured directly in mass units. Instead, metabolic processes closely coupled to growth are measured and BP is derived using conversion factors. The two most common methods follow DNA and protein synthesis using (3H)thymidine and (3H)leucine incorporation rates, respectively. The values for the conversion factors are poorly constrained and hard to measure, leading to uncertainty of at least a factor of two in the BP estimates. Few measurements were performed in the open sea before the 1990s. The Joint Global Ocean Flux Study (JGOFS) time-series station at Bermuda is perhaps the best-studied site in the ocean (Table 1).


274

BACTERIOPLANKTON

Table 1

The biomass (B) and production rates (P) of bacterioplankton and phytoplankton at estuarine and open ocean locationsa Biomass (mmol m3)

Production rate (mmol m3 d1)

P/B (d 1)

Location

Phytoplankton

Bacteria

Phytoplankton

Bacteria

Phytoplankton

Bacteria

Chesapeake Bay

5–400 (56)

1–80 (11)

20–47 (33)

0.1–50 (4)

0.07–1.9

0.01–2 (0.34)

Sargasso Sea

0.3–3.2 (1.0)

0.2–0.6 (0.4)

0.06–0.9 (0.3)

0.002–0.07 (0.02)

0.1–1 (0.3)

0.01–0.16 (0.06)

a The values are annual, euphotic zone averages derived from published reports. P/B is the specific turnover rate for the population. The data are presented as ranges with the mean of various estimates in parentheses. Ranges encompass observations and assumptions about conversion factors for deriving values from measurements (see text).

In the open sea, far removed from allochthonous inputs of organic matter, we can compare BP and PP directly, since all the organic matter ultimately derives from the PP. One difficulty is that BP itself is not constrained by PP, since if the recycling efficiency of DOM and the BGE are sufficiently high, BP can exceed PP. BP also commonly exceeds local PP in estuaries, where inputs of terrestrial organic matter are consumed by bacteria. Bacterial respiration, however, cannot exceed the organic matter supply and serves as an absolute constraint on estimates of BP. But as noted above, bacterial respiration is very hard to measure and there are many fewer reliable measurements than for BP itself. BR is usually estimated from the BGE. Rearranging eqn [1], BR ¼

ð1 BGEÞBP BGE

½4

Most commonly, variations of eqn [1] have been used to estimate the total bacterial carbon utilization or demand (BCDBR þ BP) from estimates or assumptions about BP and BGE. Earlier estimates and literature surveys suggested that BP was as high as 30% of PP. Combining this value with a BGE of 20% yields a BCD of 1.5 times the PP. This estimate in itself is possibly acceptable, if recycling of DOM is high, but then eqn [4] yields a BR of 1.2 times the total PP – an impossibility. More recent estimates of BP, typified by the Sargasso Sea data, suggest BP is about 10% of PP in the open sea. Applying this value and the mean BGE for the region (0.14), we find that BR consumes about 55% of the primary production in the Sargasso Sea, still a substantial figure. Similar calculations for other well-studied ocean areas suggest that zooplankton (including protozoans and microzooplankton) and bacteria consume nearly equal amounts of the total primary productivity. These estimates illustrate the biogeochemical importance of bacterioplankton in the ocean carbon cycle: although their growth efficiency is low,

bacteria process large amounts of DOM. DOM produced by a myriad of ecological and physiological processes must escape bacterial metabolism to enter long-term storage in the oceanic reservoir.

Role in Food Webs and Biogeochemical Cycles The process of bacterivory (consumption of bacteria by bacteriovores) completes the microbial loop. Bacterioplankton cells are ingested by a great diversity of predators, but, because of the small size of the prey, most bacteriovores are small protozoans, typically o5 mm nanoflagellates and small ciliates. Bacterial cells only occupy about 107 of the volume of the upper ocean, indicating the difficulty of encountering these small prey. Larger flagellates, small ciliates, and some specialized larger predators can also ingest bacterial prey. The most important of the larger predators are gelatinous zooplankton like larvaceans, which use mucus nets to capture bacterial cells sieved from suspension. But most bacteriovores are also very small. Nanoflagellates can clear up to 105 body volumes per hour, thus making a living from harvesting small, rare bacterial prey, and generally dominating bacterivory in the sea. Protozoan bacterivory closely balances BP in lessproductive oceanic regimes. Most crustacean zooplankton cannot efficiently harvest bacterioplankton unless the latter are attached to particles, effectively increasing their size. Bacterial prey enter marine food webs following ingestion by flagellates, and ingestion of the flagellates by other flagellates, ciliates, and copepods. This means that bacteria usually enter the higher trophic levels after several cycles of ingestion by consumers of increasing size, with attendant metabolic losses at each stage. The microbial loop and its characteristic long, inefficient food chains can be short-circuited by the gelatinous bacteriovores, which package bacterial cells into larger prey.


BACTERIOPLANKTON

Compared to phytoplankton and to bacteriovores, bacteria are enriched relative to body carbon in nitrogen, phosphorus, protein, nucleic acids, and iron. Their excess nutritional content, coupled with the many trophic exchanges that bacterial biomass passes through as it moves in food webs, means that the microbial loop is primarily a vehicle for nutrient regeneration in the sea rather than an important source of nutrition for the upper trophic levels. The main function of bacteria in the microbial loop is to recover ‘lost’ DOM, enrich it with macro- and micronutrients, and make it available for regeneration and resupply to primary producers. Lower bacterial production estimates (see above) would also tend to decrease the importance of bacteria as a subsidy for higher consumers. In estuaries and other shallow near-shore habitats, BP is not as closely balanced by planktonic bacteriovores as in ocean systems. In these productive habitats, bacteria are larger and more often associated with particles, so they are vulnerable to a wider range of grazers. Bacteria can also be consumed by mussels, clams, and other benthic suspension feeders. External subsidies of organic matter mean that BP is higher inshore, so bacteria are a more important food source in coastal and estuarine food webs than in oceanic waters. In these productive systems where bacterial abundance is greater, more of the bacterial stock is also attacked and lysed by viruses, resulting in release of DOM and nutrients instead of entry into food webs. The relative importance of viruses and bacteriovores in removing bacteria is not yet well known, but has important implications for food web structure. Bacteria are the major engines of biogeochemical cycling on the planet, and serve to catalyze major transformations of nitrogen and sulfur as well as of carbon. They participate in the carbon cycle in several ways. Their principal role is to serve as a sink for DOM, and thus regulate the export of DOM from the productive layer. Bacteria also have intensive hydrolytic capability and participate in decomposition and mineralization of particles and aggregates. Bacteria rapidly colonize fresh particulate matter in the sea, and elaborate polymeric material that helps to cement particles together, so they both reduce particle mass by enzymatic hydrolysis and promote particle formation by fostering aggregation. The balance of bacterial activity for forming particles and accelerating particle sedimentation or, in contrast, decomposing particles and reducing it, is not clear. Larvaceans and other giant, specialized bacteriovores, centimeters to meters in size, can repackage tiny bacterial cells into large, rapidly sinking

275

fecal aggregates, thus feeding the ocean’s smallest organisms into the biological carbon pump.

Conclusions Knowledge of the dynamics of bacterioplankton, their identity, roles in food webs and biogeochemical cycles is now becoming better known and integrated in a general theory of plankton dynamics, but these aspects are not yet common in plankton ecosystem models. The differential importance of bacteria in plankton food webs in coastal and oceanic systems might serve as a good test of our understanding in models. The dynamics of DOM are only crudely parametrized in most models, and explicit formulation of bacterial DOM utilization may help in better characterizing DOM accumulation and export. Other interesting problems such as the effects of size-selective predation, bacterial community structure, and species succession are just beginning to be explored. Exploration of marine bacterial communities together with molecular probes and numerical models should lead to a new revolution in plankton ecology.

See also Carbon Cycle. Copepods. Hydrothermal Vent Deposits. Hydrothermal Vent Fauna, Physiology of. Marine Mammal Trophic Levels and Interactions. Microbial Loops. Network Analysis of Food Webs. Primary Production Methods. Primary Production Processes. Protozoa, Planktonic Foraminifera. Protozoa, Radiolarians. Small-Scale Physical Processes and Plankton Biology.

Further Reading Azam F (1998) Microbial control of oceanic carbon flux: the plot thickens. Science 280: 694--696. Carlson C, Ducklow HW, and Sleeter TD (1996) Stocks and dynamics of bacterioplankton in the northwestern Sargasso Sea. Deep-Sea Research II 43: 491--516. del Giorgio P and Cole JJ (1998) Bacterial growth efficiency in natural aquatic systems. Annual Review of Ecological Systems 29: 503--541. Ducklow HW and Carlson CA (1992) Oceanic bacterial productivity. Advances in Microbial Ecology 12: 113--181. Kemp PF, Sherr B, Sherr E, and Cole JJ (1993) Handbook of Methods in Aquatic Microbial Ecology. Boca Raton, FL: Lewis Publishers. Kirchman DL (2000) Microbial Ecology of the Oceans. New York: Wiley. Pomeroy LR (1974) The ocean’s food web, a changing paradigm. BioScience 24: 499--504.


BALEEN WHALES J. L. Bannister, The Western Australian Museum, Perth, Western Australia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 224–235, & 2001, Elsevier Ltd.

Diagnostic Characters and Taxonomy Baleen or whalebone whales (Mysticeti) comprise one of the two recent (nonfossil) cetacean suborders. They differ from the other suborder (toothed whales, Odontoceti), particularly in their lack of functional teeth. Instead they feed on relatively very small marine organisms by means of a highly specialized filter-feeding apparatus made up of baleen plates ‘whalebone’ attached to the gum of the upper jaw. Other differences from toothed whales include the baleen whales’ paired blowhole, symmetrical skull, and absence of ribs articulating with the sternum. Baleen whales are generally huge. In the blue whale they include the largest known animal, growing to more than 30 m long and weighing more than 170 tonnes. Like all other cetaceans, baleen whales are totally aquatic. Like most of the toothed whales, they are all marine. Many undertake very long migrations, and some are fast swimming. A few species come close to the coast at some part of their life cycle and may be seen from shore; however, much of their lives is spent remote from land in the deep oceans. Baleen whale females grow slightly larger than the males. Animals of the same species tend to be larger in the Southern than in the Northern Hemisphere. Within the mysticetes are four families: right whales (Balaenidae, balaenids), pygmy right whales (Neobalaenidae, neobalaenids), gray whales (Eschrichtiidae, eschrichtiids); and ‘rorquals’ (Balaenopteridae, balaenopterids). Within the suborder, 12 species are now generally recognized (Table 1). Right whales are distinguished from the other three families by their long and narrow baleen plates and arched upper jaw. Other balaenid features include, externally, a disproportionately large head (c. one-third of the body length), long thin rostrum, and huge bowed lower lips; they lack multiple ventral grooves. Internally, there is no coronoid process on the lower jaw and cervical vertebrae are fused together. Within the family are two distinct groups: the bowhead (Balaena mysticetus) of northern polar waters (formerly known as the ‘Greenland’ right

276

whale) and the three ‘black’ right whales (Eubalaena spp.) of more temperate seas (so called to distinguish them from the ‘Greenland’ right whale). (Figure 1). All balaenids are robust. Pygmy right whales (Capnea marginata) have some features of both right whales and balaenopterids. The head is short (c. one-quarter of the body length), although with an arched upper jaw and bowed lower lips, and there is a dorsal fin. The relatively long and narrow baleen plates are yellowish-white, with a dark outer border, quite different from the all-black balaenid baleen plates. Internally, pygmy right whales have numerous broadened and flattened ribs. Gray whales (Eschrichtius robustus) are also somewhat intermediate in appearance between right whales and balaenopterids. They have short narrow heads, a slightly arched rostrum, and between two and five deep creases on the throat instead of the balaenopterid ventral grooves. There is no dorsal fin, but a series of 6 to 12 small ‘knuckles’ along the tail stock. The yellowish-white baleen plates are relatively small. Balaenopterids comprise the five whales of the genus Balaenoptera blue, B. musculus; fin, B. physalus; sei, B. borealis; Bryde’s, and minke, B. acutorostrata & B. bonaerensis; whales and the humpback whale (Megaptera novaeangliae). All have relatively short heads, less than a quarter of the body length. In comparison with right whales, the baleen plates are short and wide. Numerous ventral grooves are present, and there is a dorsal fin, sometimes rather small. Internally, the upper jaw is relatively long and unarched, the mandibles are bowed outward, and a coronoid process is present; cervical vertebrae are generally free. All six balaenopterids are often known as ‘rorquals’ (said to come from the Norse ‘whale with pleats in its throat’). Strictly speaking, the term should probably be applied to the five Balaenoptera species, recognizing the rather different humpback in its separate genus, but many authors now use it for all six balaenopterids. Baleen whales are sometimes called ‘great whales.’ Despite their generally huge size, some of the species are relatively small, and it seems preferable to restrict the term to the larger mysticetes (blue, fin, sei, Bryde’s, humpback) together with the largest odontocete (the sperm whale, Physeter macrocephalus) (Figure 2). In a recent review of the systematics and distribution of the world’s marine mammals, Rice (1998)


BALEEN WHALES

Table 1

277

Mysticetes (baleen whales)

Family

Genus

Species

Balaena Eubalaena

B. mysticetus E. glacialis

Balaenidae

E. australis E. japonica Neobalaenidae Neobalaena Caperea marginata Eschrichtiidae Eschrichtius E. robustus Balaenopteridae Megaptera M. novaeangliae BalaenopteraB. acutorostrata

B. bonaerensis B. edeni B. borealis B. physalus B. musculus

Subspecies

Common name

Maximum length (m)

Right whales Bowhead whale 19.8 North Atlantic right 17.0 whale Southern right 17.0 whale North Pacific right 17.0 whale Pygmy right whales Pygmy right whale 6.4 Gray whales Gray whales 14.1 ‘Rorquals’ Humpback whale 16.0 Common minke whale B. a. N. Atlantic minke 9.2 acutorostrata whale B. a. N. Pacific minke ? scammoni whale B. a. subsp. Dwarf minke whale? Antarctic minke whale Bryde’s whale Sei whale Fin whale Blue whale Blue whale

10.7 14.0 17.7 26.8

B. m. 26.0 musculus B. m. indica Great Indian ? rorqual B. m. Pygmy blue whale 24.4 brevicauda B. m. ‘True’ blue whale 30.5 intermedia

has drawn attention to the derivation of the Latin word Mysticeti and clarified the status of a variant, Mystacoceti. He describes the former as coming from Aristotle’s original Greek mustoketos, meaning ‘the mouse, the whale so-called’ or ‘the mousewhale’ (said to be an ironic reference to the animals’ generally vast size). Mystacoceti means ‘moustache whales,’ and although used occasionally in the past (and more obviously appropriate for whales with baleen in their mouths), it has been superseded by Mysticeti. The 12 species in Table 1 differ somewhat from those listed by Rice. Some authors disagree with his use of the genus Balaena for Eubalaena and his preference for the single species glacialis rather than

Generalized distribution

Circumpolar in the Arctic Temperate–Arctic Temperate–N. Atlantic Temperate N. Pacific

Temperate, Southern Hemisphere only

North Pacific–Arctic Worldwide Worldwide Temperate–Arctic Temperate–Arctic Temperate–subantarctic, Southern Hemisphere only Temperate–Antarctic Circumglobal, tropical–subtropical Worldwide, largely temperate Worldwide Worldwide N. Atlantic, N. Pacific N. Indian Ocean Southern Hemisphere, temperate– subAntarctic Southern Hemisphere, temperate– Antarctic

the three species, Eubalaena glacialis, E. japonica and E. australis (the North Atlantic, North Pacific and southern right whales). While acknowledging the need for further investigation, they refer to present-day biologists’ usage, and genetic information, in preferring a separation between Northern and Southern Hemisphere animals and in recognizing two species in the Northern Hemisphere; Eubalaena is, however, the only mysticete genus where one or more separate species is recognized in each hemisphere. Rice also distinguishes between two species of Bryde’s whale: Balaenoptera edeni and B. brydei. The taxonomic status of these ‘inshore, smaller’ and ‘offshore, larger’ forms has yet to be determined and here they are subsumed within B. edeni. In the case


278

BALEEN WHALES

Bowhead

Right

Pygmy right

Gray

Blue

Minke

Humpback

Figure 1 Lateral profiles of representative baleen whales, with a human figure, to scale.

of the blue whale, Rice’s inclusion of a northern Indian Ocean form (B. m. indica, referred to by Rice as ‘the great Indian rorqual’) has been followed. Similarly, his listing of three subspecies of minke whale, including the Southern Hemisphere dwarf minke, which has yet to be formally described, has been retained. However, other subspecies, e.g., two sei whales and two fin whales, have not been included.

Baleen plates

Tongue

Distribution and Ecology Habitat

Baleen fringes Figure 2 Head of a right whale, showing the arrangement of the filter-feeding apparatus (from Bonner, 1980).

In addition the subspecies listed in the previous section, many stocks or populations have been


BALEEN WHALES

recognized, some mainly for management purposes, based on more or less valid biological grounds. Some significant examples include the following. Bowhead whales In addition to the currently most abundant population (the Bering-Chukchi-Beaufort Seas stock), four others are recognized: Baffin Bay/ Davis Strait, Hudson Bay, Spitzbergen, and Okhotsk Sea. Right whales In the North Atlantic species, two populations are currently recognized, western and eastern, with calving grounds off the southeastern United States and northwestern Africa. The latter may now represent only a relict population(s). In the North Pacific species, the current view is that there well may once have been two or more stocks, based on feeding ground information: at least one now centered in summer on the Sea of Okhotsk and another, although possibly not now a functioning unit, summering in the Gulf of Alaska. In the southern right whale, there are several populations, defined by currently occupied calving grounds, but these cover only a proportion of the many areas known from historical whaling records to have once been occupied by right whales. Up-todate information is available on presumed discrete populations off eastern South America, South Africa, southern Australia, and sub-Antarctic New Zealand. Gray whales A western North Atlantic population may have persisted until the 17th or 18th centuries, but is now extinct. The species now survives only in the North Pacific, where, in addition to a flourishing ‘Californian’ stock, wintering on the coast of Baja California, and summering in the Bering Sea, animals are now being reported from a remnant western stock, summering in the northern Okhotsk Sea. Humpback whales In the North Atlantic, two major populations are recognized: one based on animals wintering in the West Indies and the other, now possibly only a relict population, wintering around the Cape Verde Islands. In the North Pacific, three discrete wintering grounds have been recorded: around the Bonin, Mariana, and Marshall Islands in the west; around the Hawaian Islands in the center; and off Mexico in the east. In the Southern Hemisphere, seven populations have been postulated. Six are well defined, based on calving (wintering) grounds either side of each continent (one off eastern Australia is closely related to animals wintering off Fiji and Tonga), and a possible seventh in the central Pacific. In the northwest Indian

279

Ocean, there seems to be a separate population where animals have been reported present throughout the year. Baleen whales thus occupy a wide variety of habitats, from open oceans to continental shelves and coastal waters, from the coldest waters of the Arctic and Antarctic, through temperate waters of both hemispheres and into the tropics. Most specialized is the bowhead, Balaena, restricted to the harsh cold and shallow seas of the Arctic and sub-Arctic. The black right whales (Eubalaena) are more oceanic and prefer generally temperate waters, but come very close to coasts in winter to give birth, particularly in the Southern Hemisphere. Once believed not to penetrate much further south than the Antarctic convergence (c. 50– 551S), there have been recent records in the Antarctic proper, south of 601S. Whether this is a new phenomenon is unclear: a report by Sir James Clark Ross of many ‘common black’ (i.e., right) whales in the Ross Sea (eastern Antarctic) at 631S in December 1840 was discounted when their presence there later that century could not be confirmed. It has been suggested that the currently greatly reduced population of the western North Atlantic right whale, now wintering off the south eastern United States and summering in coastal waters north to the Bay of Fundy (c. 451N), may represent the peripheral remnant of a more widely distributed stock, formerly summering north to Labrador and southern Greenland, i.e., to at least 601N (Figure 3). The pygmy right whale (Caperea) is restricted to Southern Hemisphere temperate waters, between about 301 and 521S; it can be found coastally in winter in some areas, and all-year round in others. Gray whales (Eschrichtius) are the most obviously coastal baleen whales. The long coastal migration of the ‘Californian’ stock, from Mexico to Alaska, supports a major whale-watching industry from December to April. In spring the animals migrate through the Bering Strait into the more open waters of the Bering Sea, but still favoring more shallow waters. Among the balaenopterids, fin and sei whales are probably the most oceanic, with the former penetrating into colder waters than the latter in summer. Blue whales can be found closer inshore, but are often associated with deep coastal canyons, e.g., off central and southern California. The Southern Hemisphere pygmy blue whale (subspecies B. m. brevicauda) has been regarded as restricted to more temperate waters than the ‘true’ blue whale (B. m. intermedia), not often being found much beyond 551S. The most coastal balaenopterid is the humpback (Megaptera), with long migrations between


280

BALEEN WHALES

Breeding cycle Month

Zone of migrations

Antarctic feeding zone

20˚ Mean latitude south

Probable range of breeding grounds

1st year A M

2nd year

Impregnation

Calving Pregnent

C

I 20˚

Pregnant in opposite phase

30˚ 40˚ 50˚

Mother and calf

Mother and calf in opposite phase

30˚

Weaning Resting

60˚

2-year cycle of adult female

60˚ Latitude of pack ice

Gestation

Lactation

Resting

Fetal growth

Neonatal growth (suckling)

Independent feeding

Birth Conception

70˚ Feeding

Feeding

14 12 10 8 6 4 2 0

40˚ 50˚

70˚

Length of fetus and calf (m)

New cycle

J J A S O N D J F M A M J J A S O N D J F M A M J J Winter Summer Winter Winter Summer

Weaning

Accelerated growth

14 12 10 8 6 4 2 0

Grow th of fetus and calf

J J A S O N D J F M A M J J A S O N D J F M A M J J

Figure 3 ‘Typical’ life cycle of a southern baleen whale (as modified by Bonner, 1980, from Mackintosh, 1965).

temperate/tropical breeding grounds and cold-water feeding grounds. In the Southern Hemisphere, much of its journey occurs along the east and west coasts of the three continents. In the Northern Hemisphere, humpbacks are rather more oceanic, but still coastal at some stage in their migration: in the North Pacific they can be found wintering off the Hawaiin Islands and summering off Alaska and in the western North Atlantic they winter in the Caribbean and summer between New England, the west coast of Greenland, and Iceland. Minke whales are wide ranging, from polar to tropical waters in both hemispheres. In the Southern Hemisphere they can, with blue whales, be found closest to the ice edge in summer. Elsewhere they can often occur near shore, in bays and inlets. Their migrations are less well defined and predictable than other migratory balaenopterids; in some regions they are present year-round. The most localized balaenopterid is Bryde’s whale. It is the only species restricted entirely to tropical/ warm temperate waters and probably does not undertake long migrations. The two forms – inshore and offshore, in several areas – can differ in their movements. Off South Africa, for example, the inshore form is thought to be present throughout the year, whereas the offshore form appears and disappears seasonally, presumably in association with movements of its food, shoaling fish. Food and Feeding

Although they include the largest living animals, baleen whales feed mainly on very small organisms

and are strictly carnivorous, feeding on zooplankton or small fish. In ‘filter-feeding’ – sieving the sea – baleen whales are quite different from toothed whales, where the prey is captured individually. Filter feeding has been described as requiring, in addition to a supply of food in the water, three basic features: a flow of water to bring prey near the mouth; a filter to collect the food but allow water to pass through; and a means of removing the filtered food and conveying it to the stomach for digestion. Baleen whales meet those requirements by (a) seeking out areas where their food concentrates, (b) either swimming open-mouthed through food or gulping it in, (c) possessing a highly efficient filter formed by the baleen plates, and (d) forcing the water containing the food out through the baleen plates and then transfering the trapped food back to the gullet and hence to the stomach. In the latter the tongue is presumed to be involved; in balaenopterids the process is aided by the distensible throat. While all baleen whales possess a filter based on baleen plates, two rather different systems – essentially ‘skimming’and ‘gulping’ – have evolved to filter a large volume of water containing food. Each relies on a series of triangular baleen plates, borne transversely on each upper jaw. The inner, longer border (hypotenuse) of each plate bears a fringe of fine hairs, forming a kind of filtering ‘doormat.’ Quite unrelated to teeth (which appear as early rudiments in the gums of fetal baleen whales), baleen is closest in structure to mammalian hair and human fingernails. In the right whales, filtration is achieved with very long and narrow plates in the very large mouth, itself carried in the very large head. The plates, up to 4 m long in bowheads and


BALEEN WHALES

2.7 m in the right whales, are accommodated in the mouth by an arched upper jaw and are enclosed in massively enlarged and upwardly bowed lower lips. There is a gap between the rows at the front of the mouth, and the whole arrangement allows the whale to scoop up a great quantity of water while swimming slowly forward. In balaenopterids, with their much smaller heads, the baleen plates are shorter and broader and the rows are continuous at the front. Taking in a large volume of water and food is usually achieved by swimming through a food swarm and gulping, while simultaneously enlarging the capacity of the mouth greatly by extending the ventral grooves and depressing the tongue. The two systems allow, on the one hand, the relatively slow-swimming balaenids to concentrate their rather sparse slow-swimming food over a period, and on the other, the faster-swimming balaenopterids to take in large amounts of their highly concentrated fast-swimming prey over a shorter time. Typically, baleen whales feed on zooplankton, mainly euphausiids or copepods, swarming in polar

Table 2

Baleen Whale Food Items

Species

Subspecies

or subpolar regions in summer. That is particularly so in the Southern Hemisphere, where the summer distributions of several balaenopterids depend on the presence of Euphausia superba known to whalers by the Norwegian word ‘krill’ (see Krill) in huge concentrations in the Antarctic. In the Northern Hemisphere, with a more variable availability of food, balaenopterids are more catholic in their feeding. Humpbacks and fin whales, for example, feeding almost exclusively on krill in the south, commonly take various species of schooling fish in the north. The variety of organisms taken by the various species in different regions is listed in Table 2. While most feeding occurs in colder waters, baleen whales may feed opportunistically elsewhere. All baleen whales but one, the gray whale, feed generally within 100 m of the surface and, consequently, unlike many toothed whales, do not dive very deep or for long periods. Gray whales feed primarily on bottom-living organisms, almost exclusively amphipods, in shallow waters.

Common name

Food items Northern Hemisphere

B. mysticetus

Southern Hemisphere

Bowhead whale

Mainly calanoid copepods; euphausiids; occasional mysids, amphipods, isopods, small fish North Atlantic right whale Calanoid copepods; euphausiids Southern right whale Copepods; postlarval Munida gregaria; Euphausia superba Pygmy right whale Calanoid copepods

E. glacialis E. australis Caperea marginata E. robustus

Gray whale

M. novaeangliae

Humpback whale

B. acutorostrata B. a. acutorostrata B. a. scammoni

N. Atlantic minke N. Pacific minke

B. bonaerensis B. edeni

B. a. subsp.

Dwarf minke Antarctic minke Bryde’s whale

B. borealis

Sei whale

B. physalus

Fin whale

B. musculus

281

B. m. musculus B. m. indica B. m. intermedia

Blue whale Great Indian rorqual ‘True’ blue

B. m. brevicauda

Pygmy blue

Gammarid amphipods; occasional polychaetes Schooling fish; euphausiids

E. superba (Antarctic); euphausiids, postlarval M. gregaria, occasional fish (ex-Antarctic)

Schooling fish; euphausiids Euphausiids; copepods; schooling fish

Pelagic crustaceans, including euphausiids Schooling fish Schooling fish; squid; euphausiids; copepods Euphausiids ?Euphausiids; copepods


?Euphausiids, schooling fish E. superba Schooling fish; euphausiids Copepods, including Calanus and E. superba E. superba (Antarctic); other euphausiids (ex-Antarctic)

E. superba (Antarctic); other euphausiids (ex-Antarctic) Euphausiids, mainly E. vallentini

282

BALEEN WHALES

The baleen plate structure, particularly the inner fringing hairs, to some extent mirrors the food organisms taken or (in the case of E. superba) different size classes. Thus there is some correlation between decreasing size of prey and fineness of baleen by species, i.e. gray, blue, fin, humpback, minke, sei, and right whales. Where food stocks are very dense, e.g., around sub-Antarctic South Georgia, fin, blue, and sei whales may all overlap in their feeding on E. superba. Baleen whale food consumption per day has been calculated as some 1.5–2.0% of body weight, averaged over the year. Given that feeding occurs mainly over about 4 months in the summer in the larger species, the food intake during the feeding season has been calculated at some 4% of body weight per day, c. 4000 kg per day for a large blue whale. To survive the enormous drain of pregnancy and lactation, it has been calculated that a pregnant female baleen whale needs to increase its body weight by up to 65%. The ability to achieve such an increase in only a few months’ feeding indicates the great efficiency of the baleen whales’ feeding system. Predators and Parasites

Apart from humans, the most notable baleen whale predator is the killer whale (Orcinus orca). Minke whales have been identified as a major diet item of some killer whales in the Antarctic. Killer whale attacks have been reported on blue, sei, bowhead, and gray whales, although their frequency and success are unknown. Humpbacks often have killer whale tooth marks on their bodies and tail flukes. Humpback and right whale calves in warm coastal waters are susceptible to attack by sharks. There are anecdotal reports of calving ground attacks on humpbacks by false killer whales (Pseudorca crassidens). A form of harassment, only recently described, occurs on right whales on calving grounds off Peninsula Valdes, Argentina. Kelp gulls have developed the habit of feeding on skin and blubber gouged from adult southern right whales’ backs as they lie at the surface. Large white lesions can result. The whales react adversely to such gull-induced disturbance and calf development may be affected. External parasites, particularly ‘whale lice’ (cyamid crustaceans) and barnacles (both acorn and stalked) are common on the slower-swimming more coastal baleen whales, such as gray, humpback, and right whales. In the latter, aggregations of light-colored cyamids on warty head callosities have facilitated research using callosity-pattern photographs for individual identification. External parasites are much less common on the faster swimming species,

although whale lice have been reported on minke whales (in and around the ventral grooves and umbilicus); the highly modified copepod Penella occurs particularly on fin and sei whales in warmer waters. The commensal copepod Balaenophilus unisetus often infests baleen plates in such waters, especially on sei and pygmy blue whales. A variety of internal parasites have been recorded, although some baleen whales seem less prone to infection than others. They appear, for example, to be less common in blue whales, but prevalent in sei whales. Records include stomach worms (Anisakis spp.), cestodes, kidney nematodes, liver flukes, and acanthocephalans (‘thorny-headed’ worms) of the small intestine. The cold water diatom Cocconeis ceticola often forms a brownish-yellow film on the skin of blue and other baleen whales in the Antarctic. Because the film takes about a month to develop, its extent can be used to judge the length of time an animal has been there. Its presence led to an early common name for the blue whale ‘sulphur bottom.’ For many years the origin of small scoop-shaped bites on baleen whale bodies in warmer waters remained a mystery until they were found to be caused by the small ‘cookie-cutter’ shark, Isistius brasiliensis. Some species are highly prone to such attacks. In Southern Hemisphere sei whales the overlapping healing scars can impart a galvanized-iron sheen to the body.

Life History Behavior

Sound production Unlike toothed whales, baleen whales are not generally believed to use sound for echo location, although bowheads, for example, are thought to use sound reflected from the undersides of ice floes to navigate through ice fields. However, sound production for communication, for display, establishment of territory, or other behavior, is well developed in the suborder. Blue whales produce the loudest sustained sounds of any living animal. At up to nearly 190 decibels, their long (half-minute or more), very low frequency (o20 Hz) moans may carry for hundreds of kilometers or more in special conditions. Fin whales produce similarly low (20Hz) pulsed sounds. Minke whales also produce a variety of loud sounds. Right whales produce long low moans; bowhead sounds, recorded on migration past hydrophone arrays in nearshore leads, have been used in conjunction with sightings to estimate population size off northern Alaska.


BALEEN WHALES

Southern right whales, at least, seem to use sound to communicate with their calves. Humpbacks produce the longest, most complex sound sequences in ‘songs’, described as an array of moans, groans, roars, and sighs to high-pitched squeaks and chirps, lasting 10 or more minutes before repetition, sometimes over hours. It seems that only the adult males sing, generally only in or close to the breeding season. In any one breeding season, all the males sing the same song, changing slightly over successive seasons. Different populations have different songs; so much so, for example, that those off western Australia have a distinctly different song – less complex, less ‘chirpy’ – than that heard on breeding grounds separated by the Australian continent, off the east coast. ‘Songs’ may also be heard in migrating humpbacks, but less so on the cold-water feeding grounds, where if they occur at all, they appear generally only as ‘snatches’ or isolated segments. Swimming and migration With their stream-lined bodies, rorquals include the fastest-swimming baleen whales. Sei whales have been recorded at around 35 knots (more than 60 km/h) in short bursts; minke and fin whales are also known as fast swimmers, the latter up to 20 knots (37 km/h). Blue whales are among the most powerful swimmers, able to sustain speeds of over 15 knots (28 km/h) for several hours. On migration, humpbacks and gray whales average about 4 knots (8–9 km/h) and bowheads only about 2.7 knots (5 km/h). Migration speeds for southern right whales are not known, but medium range coastal movements off southern Australia indicate 1.5–2.3 knots (2.7–4.2 km/h) over 24 h, for cow/calf pairs. Baleen whales undertake some of the longest migrations known. Gray whales cover some 10 000 nautical miles (18 000 km) on the round trip between the Baja California breeding grounds and the Alaskan feeding grounds, among the longest migrations of any mammal. Southern Hemisphere humpbacks may cover as much as 501 of latitude either way between breeding and feeding grounds, a round trip of some 6000 nautical miles (11 000 km). Not all baleen whale migrations are so well marked. The biannual movements of Bering Sea bowheads are governed by the seasonal advance and retreat and of the sea ice, which varies from year to year. Although Southern Hemisphere blue and fin whales all feed extensively in the Antarctic in summer, the locations of their calving grounds are not known. Sei whale migrations are relatively diffuse and can vary from year to year in response to changing environmental conditions. By comparison, Bryde’s whales hardly

283

migrate at all, presumably being able to satisfy both reproductive and nutritional needs in tropical/warm temperate waters. Even among such migratory animals as humpbacks, it may be that not all animals migrate every year; studies off eastern Australia indicate that a proportion of adult females may not return to the calving grounds each year, and individuals have even been reported in summer further north. However, Southern Hemisphere migrating humpbacks show segregation in the migrating stream: immatures and females accompanied by yearling calves are in the van of the northward migration, followed by adult males and nonpregnant mature females; pregnant females bring up the rear. A similar pattern occurs on the southward journey, with cow/calf pairs traveling last. Very similar segregation is recorded among migrating gray whales. Baleen whale migrations have generally been regarded as taking place in response to the need to feed in colder waters and reproduce in warmer waters. Explanations for such long-range movements have included direct benefits to the calf (better able to survive in calm, warm waters), evolutionary ‘tradition’ (a leftover from times when continents were closer together), and the possible ability of some species to supplement their food supply from plankton encountered on migration or on the calving grounds. Corkeron and Connor (1999) have rejected these explanations, suggesting that there may be a major advantage to migrating pregnant female baleen whales in reducing the risk of killer whale predation on newborn calves in low latitudes. It cites in its favor the greater abundance of killer whales in higher latitudes, that their major prey (pinniped seals) is more abundant there, and that killer whales do not seem to follow the migrating animals. Social activity Large aggregations of baleen whales are generally uncommon. Even on migration, in those species where well-defined migration paths are followed (e.g., gray whales and humpbacks), individual migrating groups are generally small, numbering only a few individuals. It has been stated that predation is a main factor in the formation of large groups of cetaceans, e.g., open-ocean dolphins. Given the large size of most adult baleen whales, predation pressure is low and group size is correspondingly small. Blue whales are usually solitary or in small groups of two to three. Fin whales can be single or in pairs; on feeding grounds they may form larger groupings, up to 100 or more. Similarly, sei whales can be found in large feeding concentrations, but in groups of up to only about six elsewhere. The same is true for minke whales, found in concentrations on the


284

BALEEN WHALES

feeding grounds, but singly or in groups of two or three elsewhere. Social behavior has been studied most intensively in coastal humpbacks, e.g., on calving grounds. Male humpbacks compete for access to females by singing and fighting. The songs seem to act as a kind of courtship display. Males congregate near a single adult female, fighting for position. Such aggression can involve lunging at each other with ventral grooves extended, hitting with the tail flukes, raising the head while swimming, fluke and flipper slapping, and releasing streams of bubbles from the blowhole. As a result of such encounters, individuals can be left with raw and bleeding wounds caused by the sharp barnacles. Among southern right whales, similar ‘interactive’ groups are often observed on the coastal calving grounds in winter, involving a tight group with up to seven males pursuing an adult female, but not generally resulting in wounded animals. As for humpbacks, it is not yet certain whether such behavior results in successful mating, although, at least in such right whale groups, intromission is often observed. Feeding balaenopterids have often been reported as circling on their sides through swarms of plankton or fish. It has been suggested that gray whales feed on their right sides, as those baleen plates are more worn down, presumably through contact with the seabed. The most remarkable behavior, however, is reported from humpbacks. In the Southern Hemisphere, on swarms of krill, they may feed in the same ‘gulping’ way as other balaenopterids. In the Northern Hemisphere, two methods are commonly reported: ‘lunging’ and ‘bubble netting.’ In the former, individuals emerge almost vertically at the surface with their mouths partly open, closing them to force the enclosed water out through the baleen. In the latter, an animal circles below the food swarm; as it swims upward, it exhales a series of bubbles, forming a ‘net’ encircling the prey. It then swims upward through the prey with its mouth open, as in lunging.

22 meters, and live for possibly 80–90 years. Adult female blue whales give birth every 2–3 years, with pregnancy lasting some 10–11 months. Other balaenopterids follow the same general pattern. Mating takes place in warm waters in winter, with birth following some 11-months later. A 7to 11-month lactation period may be followed by a year ‘resting’, or almost immediately by another pregnancy. Most adults are able to reproduce from between 5 and 10 years of age and reach maximum growth after 15 or more years. The smallest balaenopterid, the common minke whale, is born after a pregnancy of some 10 months, at a length of just under 3 meters. Weaning occurs at just under 6 meters, after 3–6 months. The adult female can become pregnant again immediately following birth, but the resulting short calving interval is generally uncommon in baleen whales: 2–3 years is the norm, although female humpbacks can achieve a similar birth rate, enabling their stocks to recover rapidly after depletion (see Further Reading section). Right whales follow a similar general pattern, but there are some differences. In right whales, gestation lasts about 11 months and weaning for about another year. Females are able to reproduce successfully from about 8 years (there are records of successful first pregnancies from 6 years), but the calving interval is usually a relatively regular 3 years. For bowheads, it has been reported, rather surprisingly, that while growth is very rapid during the first year of life (from c. 4.5 meters), it may be followed by a period of several years with little or no growth. Sexual maturity occurs at 13–14 meters; at the reduced growth rate, that would not be reached until 17–20 years. Similarly, evidence shows considerable longevity in this species: stone harpoon heads found in harvested whales and last known to be used off Alaska early this century suggest that individual animals can be at least 100 years old.

Population Status Growth and Reproduction

Young baleen whales, particularly the fetus and the calf, grow at an extraordinary rate. In the largest species, the blue whale, fetal weight increases at a rate of some 100 kg/day toward the end of pregnancy. The calf’s weight increases at a rate of about 80 kg/day during suckling. During that 7-month period of dependence on the cow’s milk, the blue whale calf will have increased its weight by some 17 ton and increased in length from around 7 to 17 meters. Blue whales attain sexual maturity at between 5 and 10 years, at a length of around

For centuries, baleen whales have borne the brunt of human greed, for products and profit. Only the sperm whale, largest of the toothed whales, has rivaled them as a whaling target. Black right whales (Eubalaena) were taken in the Bay of Biscay from the 12th century, with the fishery extending across the North Atlantic by the 16th century. Attention then shifted to the Greenland whale (Balaena) near Spitzbergen and later off southern and western Greenland. Both species’ numbers were reduced to only small remnants, and in several areas (e.g., Spitzbergen and Greenland for Balaena and the


BALEEN WHALES

northeast Atlantic and the North Pacific for Eubalaena) the stocks were virtually exterminated. That destruction was undertaken using the ‘old’ whaling method, with open boats and hand harpoons, on the ‘right’ species – ‘right’ because they were relatively easy to approach, floated when dead, and provided huge quantities of products [oil for lighting, lubrication, and soap and baleen (‘whalebone’) for articles combining flexibility with strength, such as corset stays, umbrella spokes, fishing rods]. Development of the harpoon gun and steam catcher, from 1864, increased the rate of catching greatly, but also allowed attention to turn to the largest baleen whales, the blue and fin whales, whose size, speed, and tendency to sink when dead had prevented capture by the old methods. From its beginning in the North Atlantic, then, by the end of the century, in the North Pacific, ‘modern’ whaling’s next and most intensive phase moved south, first in 1904 at South Georgia in the South Atlantic, just within the Antarctic zone. Initially on humpbacks [up to 12 000 were taken in one year (1912), leading to very rapid stock decline] and then on blue and fin whales, southern whaling based on such land stations – in the Antarctic in summer and the tropics in winter – was overtaken from the late 1920s by the great development of pelagic whaling using floating factory ships. Huge annual Southern Hemisphere catches resulted – a maximum of over 40 000 in 1931 – averaging around 30 000 animals per year in the later 1930s and again after World War II until 1965. Whereas blue whales had been the preferred target in the 1930s, their great reduction in numbers led to a shift in attention progressively over the years to fin whales, to sei whales in the 1940s, and finally to minke. With depletion of stocks and more stringent conservation measures (killing of humpbacks, blue, and fin whales was banned from the mid-1960s, even though some illegal catching continued until the early 1970s or even later), catches fell to between 10 000 and 15 000 per year in the 1970s. The ‘old’ whaling story had virtually repeated itself – enormous reductions through overfishing of one species or stock leading to exploitation of other species and stocks until, apart from minke whales, only remnants were left. Since 1989, a moratorium on all commercial whaling has eliminated that pressure, with the exception of limited whaling carried out under exemption for scientific research, and since 1993, limited commercial catching of minke whales in the eastern North Atlantic. Some ‘aboriginal’ whaling has also continued in the Northern Hemisphere, on bowheads off northern Alaska, on gray whales in the Bering Sea, on fin and minke whales off Greenland, and on humpbacks in the Caribbean.

285

Despite the great scale of the kill in ‘old’ and ‘modern’ whaling, no whale species has become extinct through whaling, although a number of individual stocks have been reduced greatly; at least one, the North Atlantic gray whale, has disappeared within the past 200–300 years. In its most recent (1996) ‘Red List’ of threatened animals, the World Conservation Union (IUCN) (Table 3) includes no baleen whale species or stocks as either extinct or critically threatened (the latter within the threatened category). Within the threatened category, seven taxa – three species, one subspecies and three stocks – are listed as endangered (E), four taxa – one species and three stocks – are vulnerable (V). Six taxa – two species, one subspecies, and three stocks – are listed as at lower risk (LR), and two taxa – one species and one subspecies – as data deficient (DD). Those species under greatest current threat (E) are the Northern Atlantic and North Pacific right, sei, and fin whales, together with the ‘true’ blue subspecies, two of the five bowhead stocks (Okhotsk Sea, Spitzbergen), and the northwest Pacific gray. Next most threatened (V) are the humpback, two bowhead stocks (Hudson Bay, Baffin Bay/Davis Strait), and the North Atlantic blue. At lower risk (LR) are the southern right and Antarctic minke, one bowhead stock (Bering-Chukchi-Beaufort Seas), the North Atlantic minke, northeast Pacific gray, and North Pacific blue; all but one are further qualified as conservation dependent (cd, not vulnerable because of specific conservation efforts). The exception is the North Atlantic minke, listed as near threatened (nt, not conservation dependent but almost qualifying as vulnerable). The two taxa for which insufficient information is currently available (DD) are Bryde’s whale and the pygmy blue. The International Whaling Commission’s Scientific Committee, responsible for the assessments of such stocks’ current status, has reported encouraging recent reversals of stock decline for some stocks of some species. One, the northeast Pacific gray whale, has recovered under protection from commercial whaling (but with aboriginal catches up to some 150 per year) to at or near its ‘original’ (prewhaling) state (c. 26 000 animals). Similarly, the northwest Atlantic humpback and several Southern Hemisphere humpback populations have been showing marked increases. The latest estimate of the North Atlantic stock, some 10 600 animals in 1993 (cf. 5500 in 1986), must reflect some population growth in the intervening period, whereas two Southern Hemisphere stocks (off eastern and western Australia) have been increasing steadily, at 10% or more per year since the early 1980s. Indeed, in all areas where surveys have been undertaken recently on Southern Hemisphere humpback populations they


286

BALEEN WHALES

Table 3

IUCN ‘Red List’ Categories for Baleen Whales (1996)

Species

Subspecies

Common name

Category EX

EW

Threatened CR

B. mysticetus E. glacialis* E. australis Caperea marginatae E. robustus M. novaeangliae B. acutorostrata

B. B. B. B. B.

bonaerensis edeni borealis physalus musculus

B. a. acutorostrata B. a. scammoni B. a. subsp.

B. B. B. B.

m. m. m. m.

musculus indica intermedia brevicauda

Bowhead whale North Atlantic right whale Southern right whale Pygmy right whale Gray whale Humpback whale N. Atlantic minke N. Pacific minke Dwarf minke Antarctic minke Bryde’s whale Sei whale Fin whale Blue whale Great Indian rorqual ‘True’ blue Pygmy blue

LR

EN

VU

*a

*b

*

(cd)c

*

(cd)

*

(cd)g

*

(nt)

DD

NE

*d

*f

*

(cd) * * * *h

*

(cd)i

* *

*

Includes E. japonica. Okhotsk Sea bowhead whale, Spitzbergen bowhead whale. b Hudson Bay bowhead whale, Baffin Bay/Davis Strait bowhead whale. c Bering-Beaufort-Chuckchi Seas bowhead whale. d North Atlantic and North Pacific northern right whales. e Pygmy right whale removed from 1996 Red List. f Northwest Pacific gray whale. g Northeast Pacific gray whale. h North Atlantic blue whale. i North Pacific blue whale. Categories: Ex, extract; EW, extinct in the wild; CR, current risk; EN, endangered; VU, vulnerable; LR, lower risk; DD, data deficient; NE, not evaluated; cd, conservation dependent; nt, not threatened. a

have been shown to be undergoing some recovery. Three southern right whale stocks (off eastern South America, South Africa, and southern Australia) have been increasing since the late 1970s at around 7–8% per year, although at some 3000, 3000, and 1200 animals, respectively, all are still well below their ‘original’ stock size. Even the ‘true’ blue whale, whose future has been of considerable concern, with estimates for the late 1980s at fewer than 500 animals for the whole Antarctic, has shown recent encouraging signs. Based on a series of Antarctic sightings cruises, mainly for minke whales but including other large whales, the most recent calculations (admittedly using only small absolute numbers sighted) show that the population must have been increasing since the last estimate (1991). As yet the analyses do not permit a firm conclusion on the number now present, although it seems likely to be more than 1000. The one species or stock for which there is now very great concern is the North Atlantic right whale. At very low absolute abundance (only some 300

animals), not recovering despite protection from whaling in the 1930s, now even decreasing through a reduced survival rate and an increase in calving interval, and subject to increasing removals from ship strikes and fishing gear entanglement, the only way to ensure the species survival is to reduce such anthropogenic mortality (ship strike and entanglement) to zero. While research on mortality reduction measures should be pursued, immediate management action is urgently needed. It has been calculated that the great reduction of baleen whales by whaling, for the Antarctic to around one-third of original numbers and one-sixth in biomass, must have left a large surplus of food – some 150 million tonnes per year – available for other consumers, such as seals, penguins, and fish. (In a different way, earlier whaling in the North Atlantic, particularly on right whales, is believed to have influenced the spread of one sea bird – the fulmar – by providing food in the form of discarded whale carcasses.) In response to an increase in


BALEEN WHALES

available food, there may well have been increases in growth rates, earlier ages at maturity, and higher rates of pregnancy in some baleen whale species. However, the evidence is equivocal, as it is for competition between individual whale species. For some, e.g., right whales and sei, it has been suggested that an increase in one (right whales) could be inhibited by competition with another (sei whales). In the North Pacific, both sei and right whales can feed on the same prey – copepods – and sei whales can at times be ‘skimming’ feeders, like right whales. However, evidence that they actually compete on the same prey, in the same area, at the same time, and even on the same prey patch is lacking. Similarly, there has been much debate and speculation on whether the recovery of the Southern Hemisphere ‘true’ blue whale has been inhibited by an apparent increase in Antarctic minke whales. In that case, there may in fact be very little direct competition for food where the common prey is not limited in abundance (as in the Antarctic) and is available in large patches. The well-authenticated increases in the substantial annual rates for several stocks of Southern Hemisphere humpbacks and right whales and the possibility of at least a limited increase in numbers for the ‘true’ blue whale suggest that such competition is unlikely, at least where, as in the Antarctic, food supplies are abundant.

Glossary Baleen Plates of keratin hanging transversely in the roof of the mouth of baleen whales, forming the ‘baleen apparatus’ for filter feeding on surface plankton; formerly known as ‘whalebone’ but bearing no resemblance to true bone. Ventral grooves A series of parallel grooves or plates running longitudinally on the undersurface of the throat and chest region in balaenopterid

287

whales, allowing great expansion of the mouth during feeding. Krill Planktonic shrimp-like crustaceans of the genus Euphausia, particularly the Antractic Euphautia superba.

See also Bioacoustics. Copepods. Krill. Marine Mammal Migrations and Movement Patterns. Marine Mammal Overview. Marine Mammal Social Organization and Communication. Marine Mammal Trophic Levels and Interactions. Marine Mammals, History of Exploitation.

Further Reading Bonner WN (1980) Whales. Poole, Dorset; Blandford Press. Corkeron PJ and Connor RC (1999) Why do baleen whales migrate? Marine Mammal Science 15: 1228--1245. Harrison R and Bryden MM (eds.) (1988) Whales, Dolphins and Porpoises. Hong Kong: International Publishing Corporation. IUCN (1996) 1996 IUCN Red List of Threatened Animals (plus annexes). Gland, Switzerland: IUCN. IUCN (1996) 1996 IUCN Red List of Threatened Animals (plus annexes). Gland, Switzerland: IUCN. Laws RM (1977) Seals and Whales of the Southern Ocean. Philosophical Transactions of the Royal Society of London, Series B 279, 81--96. Leatherwood S and Reeves RR (1983) The Sierra Club Handbook of Whales and Dolphins. San Francisco: Sierra Club Books. Rice DW (1998) Marine Mammals of the World:: Systematics and Distribution, Special Publication Number 4. Lawrence, Kansas: Society for Marine Mammalogy.


BALTIC SEA CIRCULATION W. Krauss, Institut fu¨r Meereskunde an der Universita¨t Kiel, Kiel, Germany Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 236–244, & 2001, Elsevier Ltd.

Introduction The Baltic Sea consists of a number of sub-basins connected by straits and deep channels. They have an important influence on the current system. A map of the Baltic Sea with topographic subdivision and bottom topography is shown in Figure 1. The North Sea, via Skagerrak and Kattegat, is connected to the Baltic Proper by three pathways, called the Danish Straits (Figure 2). The eastern route, the Sound, is only 2 km wide at its narrowest location, shallow (sill depth 8 m) and only about 55 km long. It is the shortest inflow route for saline water. The central pathway is about 180 km long and about 13 m deep on the average, 25–30 m along the axis. It consists of the Great Belt and the Fehmarnbelt and is terminated at the east by the Darss sill (18 m deep), the shallowest sill for the main inflow. The third connection is the Little Belt, having a cross section of only 16 000 m2 compared to 255 000 m2 of the Great Belt and 80 000 m2 of the Sound. Therefore, about 70% of the in- and outflows occur through the Great Belt; the Little Belt is negligible. Moving from west to east, the Arkona Basin (45 m) is connected to the Bornholm Basin (95 m) by the Bornholm Channel. The Stolpe Channel (20 km wide by the 60 m isobath) allows inflow into the Gdansk Basin (110 m) and the Eastern Gotland Basin with a maximum depth of 250 m. Further to the north the Faro¨ Deep is followed by the Northern Basin (200 m) which extends to the east towards the Gulf of Finland as a deep channel with decreasing width and depth. By contrast, the Gulf of Bothnia is separated from the deeper layers of the Baltic Proper by the Aland Sea with its numerous islands. The Gulf of Bothnia consists of two basins, the Bothnian Sea and the Bothnian Bay. The deep basins and the connecting channels are important for the water exchange between the North Sea and the Baltic Sea and determine the current structure. Due to the large river runoff, the Baltic Sea is the largest brackish water body of the world with a volume of 21 000 km3. The mean annual total contribution of all rivers for the period 1950–1990 was

288

446 km3 year1. This would correspond to a sealevel rise of 1.18 m year1 if the Baltic Sea were not connected to the North Sea. An additional surplus of fresh water results from the difference of precipitation minus evaporation, which amounts to about 60 km3 year1. The river runoff shows both a seasonal cycle and interannual variations. The lowest and highest annual values differ from the mean value by 27% and þ 22%, respectively. The seasonal runoff varies between 25 000 m3 s1 in spring and 12 000 m3 s1 in winter. As a consequence of this freshwater supply the salinity in the eastern and northern parts of the Baltic Sea is reduced to about 4 PSU (practical salinity units), increasing to about 8 PSU in the central and western Baltic Sea. Due to the associated density difference saline water penetrates from the Kattegat (30–34 PSU) through the Danish Straits and the channels into the Baltic Sea and yields a strong halocline, which separates the water of the deeper layers from the brackish upper water masses (Figure 3). As a consequence vertical mixing is strongly reduced. The Baltic Sea extends from about 541N to 661N thus ranging from mild and humid to a subarctic climate. Frequent and complex synoptic-scale cyclonic activity and subsynoptic-scale depressions are characteristic, leading to a high variability of the prevailing westerly winds. The wind fields produce a highly variable current system, especially in the upper layers, superimposed on the weak baroclinic flow field induced by the salinity differences. Beneath the wind-mixed layer bottom topography has a strong steering influence. The wind-induced currents and the associated sea-level variations may drastically change, when parts of the Baltic Sea are icecovered. Sea ice occurs every year in the Baltic Sea. Under normal winter conditions about 45% of the Baltic Sea is covered with ice and the ice season lasts about 6 months in the northern parts. Severe winters may lead to almost total ice coverage. Ice first appears in the innermost parts of the Bothnian Bay during mid-November. In a normal winter, the entire Bothnian Sea, Aland Sea, Gulf of Finland and the northernmost part of the Baltic Proper are covered by ice. In severe ice-winters the Kattegat, the Belt Sea, the Sound and large parts of the Baltic Proper are also ice covered. Wind and thermohaline forcing determine the currents of the Baltic Sea, strongly influenced by bottom topography and ice coverage. A large


BALTIC SEA CIRCULATION

66°

10°

15°

30°

25°

20°

289

66°

14 N

N

13 62°

62°

12 11 10 1 58°

7

2

3

58°

9

8

5

6

4

54°

54° 10°

15°

20°

25°

E

30°

Figure 1 Skagerrak (1), Kattegat (2) and the subareas of the Baltic Sea: Beltsea (3), Arkona Basin (4), Bornholm Basin (5), Gdansk Basin (6), Eastern (7) and Western Gotland Basin (8), Gulf of Riga (9), Northern Basin (10), Gulf of Finland (11), Aland Sea (12), Bothnian Sea (13) and Bay of Bothnia (14). Depth contours: thin broken line 40 m, heavy dotted line 100 m, heavy full line 200 m. The location of the section of Figure 3 is also shown (full line with open dots).

number of current measurements has been made in the past decades. However, due to the high variability of the wind-induced currents and the extensive fishing activities, which make it impossible to install observational systems in some areas, it was not possible in the past to derive a consistent circulation pattern from observations. In recent years, three-dimensional models, combined with data assimilation, have improved in such a way that both the mean circulation and its variability can now be described with sufficient accuracy.

The Estuarine Circulation In a stratified sea it is generally not possible to separate wind-induced currents properly from

thermohaline ones. Along the coasts and the bottom slopes wind produces up- and downwelling and thus inclinations of the density surfaces which may largely amplify the wind effects. However, some insight is gained by considering the horizontal pressure gradients, which result from the mean surface inclination and the horizontal density differences. They produce an estuarine type of circulation. Figure 3 shows the mean salinity section from the Kattegat along the main axis of the Baltic Sea into the Gulf of Finland. Except at the level of the summer thermocline (about 25–30 m depth) the density distribution is similar. Near the bottom the density decreases from about 25 kg m3 in the Kattegat to 9 kg m3 in the Gotland Basin and less than 4 kg m3 in the interior of the Gulf of Finland.


290


57° 06'

9°

E

10°

12°

14°

15° 57° 06'

Kattegat

20

m

m 40

N

N

Sweden

Little Be

lt

Denmark

56°

Hälsingborg

Helsingör

t Be

lt

Kopenhagen

Sound

SS

56°

Klagshamn

G r ea

Drogden

Bornholm

55°

55°

Arkona basin 40 m

Fe h

ma

rn

DS

20 m

Be

lt

Germany 54°

54° 9°

E

14°

12°

10°

15°

Figure 2 The Danish Straits, Kattegat and Arkona Basin. SS, Samsoe Sill, DS, Darss Sill. The Belt Sea is the area between the Kattegat and Arkona Basin.

Because of the river runoff and the low salinity in the eastern and northern Baltic Sea the sea level is higher by 0.3 m in the interior of the Gulf of Finland and the Gulf of Bothnia compared to that in the Belt Sea. This sea level inclination produces a pressure gradient which drives the upper layers out of the Baltic Sea. Due to the Coriolis force this outflow is Great Belt

Kattegat

0 m 20

30 32

33

40 60 80

18

26 31

16

14

20 30 28 >

Darss 12 10

Bornholm Basin

>

Gotland Basin

7.5

6.5 < 7.5

26 >

7 7.5 10

9

8

16 >

180

5.5

5

4 4.5

< 3 2

0 m 20 40 60 100

120 160

6

80

100 140

Gulf of Finland

9 8

12 14

34

concentrated on the Swedish coast. On average about 1250 km3 year1 leave the Baltic Sea through the Danish Straits. As compensation about 740 km3 year1 of saline, dense and oxygen-rich water penetrates in the deeper layers through the Danish Straits into the Arkona Basin and from there through the Bornholm Channel into the Bornholm Basin.

>10

Mean salinity section August

120 140

12

160 >

200

180 200

Figure 3 Mean salinity (practical salinity units) distribution (1902–1956) along a section from the Kattegat to the Gulf of Finland (for location see Figure 1).


291


9˚

10˚

12˚

11˚

13˚

14˚

16˚

15˚

18˚

17˚

20˚

19˚

22˚

21˚

23˚

24˚

25˚

27˚

26˚

28˚

29˚

30˚

66˚

66˚ 20m 40 m 100m 200m

6 61˚

61˚ 6 5 60˚

>8

8

6

6

10

34 32

5.5
32 28 56˚

30 12

26

55˚

24

28 18 26 14 28

26 20 16 8

28

12 >

20

18 20

18 16

12

8

10

56˚

8

12

16 14

22

54˚

10 12 10

55˚

Bottom salinity August

50% of the softbottom, deep-sea meiofauna (Table 1). They are also often a major component of the macrofauna.

Table 1

339

In the central North Pacific, for example, foraminifera (mainly komokiaceans) outnumber all metazoans combined by at least an order of magnitude. A few species are large enough to be easily visible to the unaided eye and constitute part of the megafauna. These include the tubular species Bathysiphon filiformis, which is sometimes abundant on continental slopes (Figure 3). Some xenophyophores, agglutinated protists that are probably closely related to the foraminifera, are even larger (up to 24 cm maximum dimension!). These giant protists may dominate the megafauna in regions of sloped topography (e.g., seamounts) or high surface productivity. In well-oxygenated areas of the deepseafloor, foraminiferal assemblages are very species rich, with well over 100 species occurring in relatively small volumes of surface sediment (Figure 4). Many are undescribed delicate, soft-shelled forms. There is an urgent need to describe at least some of these species as a step toward estimating global levels of deep-sea species diversity. The common species are often widely distributed, particularly at abyssal depths, although endemic species undoubtedly also occur.

The percentage contribution of foraminifera to the deep-sea meiofauna at sites where bottom water is well oxygenated

Area

Depth (m)

Percentage of foraminifera

Number of samples

NW Atlantic Off North Carolina Off North Carolina Off Martha’s Vineyard

500–2500 400–4000 146–567

11.0–90.4 7.6–85.9 3.4–10.6

14 28 4

NE Atlantic Porcupine Seabight Porcupine Abyssal Plaina Madeira Abyssal Plaina Cape Verde Abyssal Plaina Off Mauretania

1345 4850 4950 4550 250–4250

47.0–59.2 61.8–76.3 61.4–76.1 70.2

8 3 3 1 26

461N, 16–171W

4000–4800

4–27 0.5–8.3

9

Indian Ocean NW Arabian Seab

3350

54.4

Pacific Western Pacific Central North Pacific

2000–6000 5821–5874

36.0–69.3 49.5

11 2

Arctic

1000–2600

14.5–84.1

74

Southern Ocean

1661–1680

2.2–23.7

2

a

1

Data from Gooday AJ (1996) Epifaunal and shallow infaunal foraminiferal communities at three abyssal NE Atlantic sites subject to differing phytodetritus input regimes. Deep-Sea Research I 43: 1395–1421. b Data from Gooday AJ, Bernhard JM, Levin LA and Suhr SB (2000) Foraminifera in the Arabian Sea oxygen minimum zone and other oxygen-deficient settings: taxonomic composition, diversity, and relation to metazoan faunas. Deep-Sea Research II 47: 25–54. Based on Gooday AJ (1986) Meiofaunal foraminiferans from the bathyal Porcupine Seabight (northeast Atlantic): size structure, standing stock, taxonomic composition, species diversity and vertical distribution in the sediment. Deep-Sea Research 35: 1345– 1373; with permission from Elsevier Science.


340

BENTHIC FORAMINIFERA

the sea floor for periods of months. They include various undescribed matlike taxa and branched tubular forms, as well as a variety of small coiled agglutinated species (many in the superfamily Trochamminacea), and calcareous forms. Role in Benthic Communities

Figure 3 Bathysiphon filiformis, a large tubular agglutinated foraminifer, photographed from the Johnson Sealink submersible on the North Carolina continental slope (850 m water depth). The tubes reach a maximum length of about 10 cm. (Photograph courtesy of Lisa Levin.)

Numbers per 25 cm 2

120 100 80 60 40 20 0 Species ordered by rank

Figure 4 Deep-sea foraminiferal diversity: all species from a single multiple corer sample collected at the Porcupine Abyssal Plain, NE Atlantic (4850 m water depth), ranked by abundance. Each bar represents one ‘live’ (rose Bengal-stained) species. The sample was 25.5 cm2 surface area, 0–1 cm depth, and sieved on a 63 mm mesh sieve. It contained 705 ‘live’ specimens and 130 species.

Foraminifera are also a dominant constituent of deep-sea hard-substrate communities. Dense populations encrust the surfaces of manganese nodules as well as experimental settlement plates deployed on

The abundance of foraminifera suggests that they play an important ecological role in deep-sea communities, although many aspects of this role remain poorly understood. One of the defining features of these protists, their highly mobile and pervasive pseudopodial net, enables them to gather food particles very efficiently. As a group, foraminifera exhibit a wide variety of trophic mechanisms (e.g., suspension feeding, deposit feeding, parasitism, symbiosis) and diets (herbivory, carnivory, detritus feeding, use of dissolved organic matter). Many deep-sea species appear to feed at a low trophic level on organic detritus, sediment particles, and bacteria. Foraminifera are prey, in turn, for specialist deep-sea predators (scaphopod mollusks and certain asellote isopods), and also ingested (probably incidentally) in large numbers by surface deposit feeders such as holothurians. They may therefore provide a link between lower and higher levels of deep-sea food webs. Some deep-sea foraminifera exhibit opportunistic characteristics – rapid reproduction and population growth responses to episodic food inputs. Wellknown examples are Epistominella exigua, Alabaminella weddellensis and Eponides pusillus. These small (generally o200 mm), calcareous species feed on fresh algal detritus (‘phytodetritus’) that sinks through the water column to the deep-ocean floor after the spring bloom (a seasonal burst of phytoplankton primary production that occurs most strongly in temperate latitudes). Utilizing energy from this labile food source, they reproduce rapidly to build up large populations that then decline when their ephemeral food source has been consumed. Moreover, certain large foraminifera can reduce their metabolism or consume cytoplasmic reserves when food is scarce, and then rapidly increase their metabolic rate when food again becomes available. These characteristics, together with the sheer abundance of foraminifera, suggest that their role in the cycling of organic carbon on the deep-seafloor is very significant. The tests of large foraminifera are an important source of environmental heterogeneity in the deep sea, providing habitats and attachment substrates for other foraminifera and metazoans. Mobile infaunal species bioturbate the sediment as they move through it. Conversely, the pseudopodial systems of



foraminifera may help to bind together and stabilize deep-sea sediments, although this has not yet been clearly demonstrated. Microhabitats and Temporal Variability

Like many smaller organisms, foraminifera reside above, on and within deep-sea sediments. Various factors influence their overall distribution pattern within the sediment profile, but food availability and geochemical (redox) gradients are probably the most important. In oligotrophic regions, the flux of organic matter (food) to the seafloor is low and most foraminifera live on or near the sediment surface where food is concentrated. At the other extreme, in eutrophic regions, the high organic-matter flux causes pore water oxygen concentrations to decrease rapidly with depth into the sediment, restricting access to the deeper layers to those species that can tolerate low oxygen levels. Foraminifera penetrate most deeply into the sediment where organic inputs are of intermediate intensity and the availability of food and oxygen within the sediment is well balanced. Underlying these patterns are the distributions of individual species. Foraminifera occupy more or less distinct zones or microenvironments (‘microhabitats’). For descriptive purposes, it is useful to recognize a number of different microhabitats: epifaunal and shallow infaunal for species living close to the sediment surface (upper 2 cm); intermediate infaunal for species living between about 1 cm and 4 cm (Figure 5); and deep infaunal for species that occur at depths

0

20

40

341

down to 10 cm or more (Figure 6). A few deep-water foraminifera, including the well-known calcareous species Cibicidoides wuellerstorfi, occur on hard substrates (e.g., stones) that are raised above the sediment–water interface (elevated epifaunal microhabitat). There is a general relation between test morphotypes and microhabitat preferences. Epifaunal and shallow infaunal species are often trochospiral with large pores opening on the spiral side of the test; infaunal species tend to be planispiral, spherical, or ovate with small, evenly distributed pores. It is important to appreciate that foraminiferal microhabitats are by no means fixed. They may vary between sites and over time and are modified by the burrowing activities of macrofauna. Foraminiferal microhabitats should therefore be regarded as dynamic rather than static. This tendency is most pronounced in shallowwater settings where environmental conditions are more changeable and macrofaunal activity is more intense than in the deep sea. The microhabitats occupied by species reflect the same factors that constrain the overall distribution patterns of foraminifera within the sediment. Epifaunal and shallow infaunal species cannot tolerate low oxygen concentrations and also require a diet of relatively fresh organic matter. Deep infaunal foraminifera are less opportunistic but are more tolerant of oxygen depletion than are species living close to the sediment–water interface (Figure 6). It has been suggested that species of genera such as Globobulimina may consume either sulfate-reducing bacteria or labile organic matter released by the metabolic

Number of specimens 0 20

0

10

Depth in sediment (cm)

0 1 2 3 4 (A)

5 0

(B) 0

5

5

10

(C) 5

0 1 2 3 4 5

(D)

(E)

(F)

Figure 5 Vertical distribution patterns within the top 5 cm of sediment of common foraminiferal species (‘live’, rose Bengal-stained specimens) in the Porcupine Seabight, NW Atlantic (511360 N, 131000 W; 1345 m water depth). Based on 463 mm sieve fraction. (A) Ovammina sp. (mean of 20 samples). (B) Nonionella iridea (20 samples). (C) Leptohalysis aff. catenata (7 samples). (D) Melonis barleeanum (9 samples). (E) Haplophragmoides bradyi (19 samples). (F) ‘Turritellella’ laevigata (21 samples). (Amended and reprinted from Gooday AJ (1986) Meiofaunal foraminiferans from the bathyal Porcupine Seabight (northeast Atlantic): size structure, standing stock, taxonomic composition, species diversity and vertical distribution in the sediment. Deep-Sea Research 35: 1345– 1373; permission from Elsevier Science.)


342

BENTHIC FORAMINIFERA _1

Depth in sediment (cm)

0

100 200 300 400 500

0

0

1

1

2

2

3

3

4

4

5

5

6

6

7

7

8

8

9

9

10

(A)

0

Oxygen (µmol l ) 50 100 150 200 250 0

20

30

40

50

1.25 2.25 3.25 4.25 5.25 6.25 7.25 _1

Nitrate (µmol l )

10 Number of stained foraminifera

10

0.25

8.25 9.25

0 20 40 60 80100120140160

(B)

Intermediate and

(C) deep infauna, number of specimens

Figure 6 Vertical distribution of (A) total ‘live’ (rose Bengal-stained) foraminifera), (B) pore water oxygen and nitrate concentrations, and (C) intermediate and deep infaunal foraminiferal species within the top 10 cm of sediment on the north-west African margin (21128.80 N, 17157.20 W, 1195 m). All foraminiferal counts based on 4150 mm sieve fraction, standardized to a 34 cm3 volume. Species are indicated as follows: Pullenia salisburyi (black), Melonis barleeanum (crossed pattern), Chilostomella oolina (honeycomb pattern), Fursenkoina mexicana (grey), Globobulimina pyrula (diagonal lines), Bulimina marginata (large dotted pattern). (Adopted and reprinted from Jorissen FJ, Wittling I, Peypouquet JP, Rabouille C and Relexans JC (1998) Live benthic foraminiferal faunas off Cape Blanc, northwest Africa: community structure and microhabitats. Deep-Sea Research I 45: 2157–2158; with permission from Elsevier Science.)

activities of these bacteria. These species move closer to the sediment surface as redox zones shift upward in the sediment under conditions of extreme oxygen depletion. Although deep-infaunal foraminifera must endure a harsh microenvironment, they are exposed to less pressure from predators and competitors than those occupying the more densely populated surface sediments. Deep-sea foraminifera may undergo temporal fluctuations that reflect cycles of food and oxygen availability. Changes over seasonal timescales in the abundance of species and entire assemblages have been described in continental slope settings (Figure 7). These changes are related to fluctuations in pore water oxygen concentrations resulting from episodic (seasonal) organic matter inputs to the seafloor. In some cases, the foraminifera migrate up and down in the sediment, tracking critical oxygen levels or redox fronts. Population fluctuations also occur in abyssal settings where food is a limiting ecological factor. In these cases, foraminiferal population dynamics reflect the seasonal availability of phytodetritus (‘food’). As a result of these temporal processes, living foraminifera sampled during one season often provide an incomplete view of the live fauna as a whole. Environmental Controls on Foraminiferal Distributions

Our understanding of the factors that control the distribution of foraminifera on the deep-ocean floor is very incomplete, yet lack of knowledge has not

prevented the development of ideas. It is likely that foraminiferal distribution patterns reflect a combination of influences. The most important first-order factor is calcium carbonate dissolution. Above the carbonate compensation depth (CCD), faunas include calcareous, agglutinated, and allogromiid taxa. Below the CCD, calcareous species are almost entirely absent. At oceanwide or basinwide scales, the organic carbon flux to the seafloor (and its seasonality) and bottom-water hydrography appear to be particularly important, both above and below the CCD. Studies conducted in the 1950s and 1960s emphasized bathymetry (water depth) as an important controlling factor. However, it soon became apparent that the bathymetric distribution of foraminiferal species beyond the shelf break is not consistent geographically. Analyses of modern assemblages in the North Atlantic, carried out in the 1970s, revealed a much closer correlation between the distribution of foraminiferal species and bottom-water masses. For example, Cibicidoides wuellerstorfi was linked to North Atlantic Deep Water (NADW) and Nuttallides umbonifera to Antarctic Bottom Water (AABW). At this time, it was difficult to explain how slight physical and chemical differences between water masses could influence foraminiferal distributions. However, recent work in the south-east Atlantic, where hydrographic contrasts are strongly developed, suggests that the distributions of certain foraminiferal species are controlled in part by the lateral advection of water masses. In the case of N. umboniferus there is good evidence that the main


Total numbers of 2 living benthic foraminifera (per 10 cm )

0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

(A) 3000

343

Oxygenated layer (cm)


2000

1000 (B) 0 34 5

8 1112 2 3 5 7 9 11

1991

12

4 6

1993

1992

5

9

12

1994

(C) Bolivina spissa

Others Fursenkoina sp.1

Bulimina aculeata

Bolivina pacifica

Globobulimina spp.

Textularia kattegatensis

Chilostomella ovoidea

Uvigerina spp. Figure 7 Seasonal changes over a 4-year period (March 1991 to December 1994) in (A) the thickness of the oxygenated layer, (B) the total population density of live benthic foraminifera, and (C) the abundances of the most common species at a 1450 m deep site in Sagami Bay, Japan. (Reprinted from Ohga T and Kitazato H (1997) Seasonal changes in bathyal foraminiferal populations in response to the flux of organic matter (Sagami Bay, Japan). Terra Nova 9: 33–37; with permission from Blackwell Science Ltd.)

factor is the degree of undersaturation of the bottom water in calcium carbonate. This abyssal species is found typically in the carbonate-corrosive (and highly oligotrophic) environment between the calcite lysocline and the CCD, a zone that may coincide approximately with AABW. Where water masses are more poorly delineated, as in the Indian and Pacific Oceans, links with faunal distributions are less clear. During the past 15 years, attention has focused on the impact on foraminiferal ecology of organic matter fluxes to the seafloor. The abundance of dead foraminiferal shells 4150 mm in size correlates well with flux values. There is also compelling evidence that the distributions of species and species associations are linked to flux intensity. Infaunal species, such as Melonis barleeanum, Uvigerina peregrina,

Chilostomella ovoidea and Globobulimina affinis, predominate in organically enriched areas, e.g. beneath upwelling zones. Epifaunal species such as Cibicidoides wuellerstorfi and Nuttallides umbonifera are common in oligotrophic areas, e.g. the central oceanic abyss. In addition to flux intensity, the degree of seasonality of the food supply (i.e., whether it is pulsed or continuous) is a significant factor. Epistominella exigua, one of the opportunists that exploit phytodetritus, occurs in relatively oligotrophic areas where phytodetritus is deposited seasonally. Recent analysis of a large dataset relating the relative abundance of ‘live’ (stained) foraminiferal assemblages in the north-east Atlantic and Arctic Oceans to flux rates to the seafloor has provided a


344


quantitative framework for these observations. Although species are associated with a wide flux range, this range diminishes as a species become relatively more abundant and conditions become increasingly optimum for it. When dominant occurrences (i.e., where species represent a high percentages of the fauna) are plotted against flux and water depth, species fall into fields bounded by particular flux and depth values (Figure 8). Despite a good deal of overlap, it is possible to distinguish a series of dominant species that succeed each other bathymetrically on relatively eutrophic continental slopes and other species that dominate on the more oligotrophic abyssal plains. Other environmental attributes undoubtedly modify the species composition of foraminiferal assemblages in the deep sea. Agglutinated species with tubular or spherical tests are found in areas where the seafloor is periodically disturbed by strong currents capable of eroding sediments. Forms projecting into the water column may be abundant where steady flow rates convey a continuous supply of suspended food particles. Other species associations may be linked to sedimentary characteristics.

the degradation of organic matter, concentrations of oxygen in bottom water and sediment pore water are inversely related to the organic flux derived from surface production. In the deep sea, persistent oxygen depletion (O2o1 ml l1) occurs at bathyal depths (o1000 m) in basins (e.g., on the California Borderland) where circulation is restricted by a sill and in areas where high primary productivity resulting from the upwelling of nutrient-rich water leads to the development of an oxygen minimum zone (OMZ; e.g., north-west Arabian Sea and the Peru margin). Subsurface sediments also represent an oxygen-limited setting, although oxygen penetration is generally greater in oligotrophic deep-sea sediments than in fine-grained sediments on continental shelves. On the whole, foraminifera exhibit greater tolerance of oxygen deficiency than most metazoan taxa, although the degree of tolerance varies among species. Oxygen probably only becomes an important limiting factor for foraminifera at concentrations well below 1 ml l1. Some species are abundant at levels of 0.1 ml l1 or less. A few apparently live in permanently anoxic sediments, although anoxia sooner or later results in death when accompanied by high concentrations of hydrogen sulfide. Oxygen-deficient areas are characterized by high foraminiferal densities but low, sometimes very low (o10), species numbers. This assemblage structure (high dominance, low species richness) arises because (i) low oxygen

Low-Oxygen Environments

Oxygen availability is a particularly important ecological parameter. Since oxygen is consumed during

0 30

0 10

1 Tetragonostomia rhombiformis 2 Nonion asterizans 3 Trifarina fornasinii 4 Rupertina stabilis 5 Bolivina albatrossi

30

10

20 1 2

Water depth (m)

100

3 4

5 6

1000

7 8 9 0.1

6 Sphaeroidina bulloides 7 Hoeglundina elegans 8 Pyrgo murrhina 9 Eponides tumidulus and Stetsonia arctica

1 10 _ _ Flux of organic carbon (g m 2 y 1)

100

Figure 8 Dominant ‘live’ (rose Bengal-stained) occurrences of foraminiferal species in relation to water depth and flux or organic carbon to seafloor in the North Atlantic from the Guinea Basin to the Arctic Ocean. Each open circle corresponds to a data point. The polygonal areas indicate the combination of water depth and flux conditions under which nine different species are a dominant faunal component. The diagonal lines indicate levels of primary production (10, 30, 100, 300 g m2 y1) that result in observed flux rates. Based on 4250 mm sieve fraction plus 63–250 mm fraction from Guinea Basin and Arctic Ocean. (Reprinted from Altenbach AV, Pflaumann U, Schiebel R et al. (1999) Scaling percentages and distribution patterns of benthic foraminifera with flux rates of organic carbon. Journal of Foraminiferal Research 29: 173–185; with permission from The Cushman Foundation.)



concentration acts as a filter that excludes non-tolerant species and (ii) the tolerant species that do survive are able to flourish because food is abundant and predation is reduced. Utrastructural studies of some species have revealed features, e.g., bacterial symbionts and unusually high abundances of peroxisomes, that may be adaptations to extreme oxygen depletion. In addition, mitachondria-laden pseudopodia have the potential to extend into overlying sediment layers where some oxygen may be present. Many low-oxygen-tolerant foraminifera belong to the Orders Rotaliida and Buliminida. They often have thin-walled, calcareous tests with either flattened, elongate biserial or triserial morphologies (e.g., Bolivina, Bulimina, Globobulimina, Fursenkoina, Loxotomum, Uvigerina) or planispiral/lenticular morphologies (e.g., Cassidulina, Chilostomella, Epistominella, Loxotomum, Nonion, Nonionella). Some agglutinated foraminifera, e.g.,

Table 2

345

Textularia, Trochammina (both multilocular), Bathysiphon, and Psammosphaera (both unilocular), are also abundant. However, miliolids, allogromiids, and other soft-shelled foraminifera are generally rare in low-oxygen environments. It is important to note that no foraminiferal taxon is currently known to be confined entirely to oxygen-depleted environments.

Deep-Sea Foraminifera in Paleo-Oceanography Geologists require proxy indicators of important environmental variables in order to reconstruct ancient oceans. Benthic foraminifera provide good proxies for seafloor parameters because they are widely distributed, highly sensitive to environmental conditions, and abundant in Cenozoic and Cretaceous deep-sea sediments (note that deep-sea

Benthic foraminiferal proxies or indicators (both faunal and chemical) useful in paleo-oceanographic reconstruction

Environmental parameter/property

Proxy or indicator

Remarks

Water depth

Bathymetric ranges of abundant species in modern oceans

Distribution of bottom water masses

Characteristic associations of epifaunal species Abundance of Nuttallides umbonifera

Depth zonation largely local although broad distinction between shelf, slope and abyssal depth zones possible Relations between species and water masses may reflect lateral advection Corrosive bottom water often broadly corresponds to Antarctic Bottom Water Proxies reflect ‘age’ of bottom watermasses; i.e., period of time elapsed since formation at ocean surface Species not consistently associated with particular range of oxygen concentrations and also found in high-productivity areas Transfer function links productivity to test abundance (corrected for differences in sedimentation rates between sites) in oxygenated sediments Assemblages indicate high organic matter flux to seafloor, with or without corresponding decrease in oxygen concentrations

Carbonate corrosiveness of bottom water Deep-ocean thermohaline circulation

Cd/Ca ratios and d13C values for calcareous tests

Oxygen-deficient bottom-water and pore water

Characteristic species associations; high-dominance, low-diversity assemblages

Primary productivity

Abundance of foraminiferal tests 4150 mm

Organic matter flux to seafloor

(i) Assemblages of high productivity taxa (e.g. Globobulimina, Melonis barleeanum)

Seasonality in organic matter flux Methane release

(ii) Ratio between infaunal and epifaunal morphotypes (iii) Ratio between planktonic and benthic tests Relative abundance of ‘phytodetritus species’ Large decrease (2–3%) in d13C values of benthic and planktonic tests


Reflects seasonally pulsed inputs of labile organic matter to seafloor Inferred sudden release of 12C enriched methane from clathrate deposits following temperature rise

346


sediments older than the middle Jurassic age have been destroyed by subduction, except where preserved in ophiolite complexes). Foraminiferal faunas, and the chemical tracers preserved in the tests of calcitic species, can be used to reconstruct a variety of paleoenvironmental parameters and attributes. The main emphasis has been on organic matter fluxes and bottom-water/pore water oxygen concentrations (inversely related parameters), the distribution of bottom-water masses, and the development of thermohaline circulation (Table 2). Modern deep-sea faunas became established during the Middle Miocene (10–15 million years ago), and these assemblages can often be interpreted in terms of modern analogues. This approach is difficult or impossible to apply to sediments from the Cretaceous and earlier Cenozoic, which contain many foraminiferal species that are now extinct. In these cases, it can be useful to work with test morphotypes (e.g., trochospiral, cylindrical, biserial/ triserial) rather than species. The relative abundance of infaunal morphotypes, for example, has been used as an index of bottom-water oxygenation or relative intensities of organic matter inputs. The trace element (e.g., cadmium) content and stable isotope (d13C; i.e., the deviation from a standard 12C : 12C ratio) chemistry of the calcium carbonate shells of benthic foraminifera provide powerful tools for making paleo-oceanographic reconstructions, particularly during the climatically unstable Quaternary period.

0

3000

6000

The cadmium/calcium ratio is a proxy for the nutrient (phosphate) content of sea water that reflects abyssal circulation patterns. Carbon isotope ratios also reflect deep-ocean circulation and the strength of organic matter fluxes to the seafloor. It is important to appreciate that the accuracy with which fossil foraminifera can be used to reconstruct ancient deep-sea environments is often limited. These limitations reflect the complexities of deep-sea foraminiferal biology, many aspects of which remain poorly understood. Moreover, simple relationships between the composition of foraminiferal assemblages and environmental variables are elusive, and it is often difficult to identify faunal characteristics that can be used as precise proxies for paleo-oceanographic parameters. For example, geologists often wish to establish paleobathymetry. However, the bathymetric distributions of foraminiferal species are inconsistent and depend largely on the organic flux to the seafloor, which decreases with increasing depth (Figure 8) and is strongly influenced by surface productivity. Thus, foraminifera can be used only to discriminate in a general way between shelf, slope, and abyssal faunas, but not to estimate precise paleodepths. Oxygen concentrations and organic matter inputs are particularly problematic. Certain species and morphotypes dominate in low-oxygen habitats that also are usually characterized by high organic loadings. However, the same foraminifera may occur in organically enriched settings where oxygen levels are

9000

0

0

0

5 000

5 000

10 000

YD

50

75

10 000

15 000

15 000

20 000

20 000

25 000

25 000

30 000

30 000

35 000

35 000

40 000

40 000

45 000

45 000 Phytodetritus species, Number per gram

25

100

YD H1 H2 H3

H4 Phytodetritus species, Relative abundance

Figure 9 (A) Absolute (specimens per gram of dry sediment) and (B) relative (percentage) abundances of Alabaminella weddellensis and Epistominella exigua (463 mm fraction) in a long-sediment core from the North Atlantic (50141.30 N, 21151.90 W, 3547 m water depth). In modern oceans, these two species respond to pulsed inputs of organic matter (‘phytodetritus’) derived from surface primary production. Note that they increased in abundance around 15 000 years ago, corresponding to the main Northern Hemisphere deglaciation and the retreat of the Polar Front. Short period climatic fluctuations (YD ¼ Younger Dryas; H1–4 ¼ Heinrich events, periods of very high meltwater production) are also evident in the record of these two species. (Reprinted from Thomas E, Booth L, Maslin M and Shackleton NJ (1995). Northeast Atlantic benthic foraminifera during the last 45 000 years: change in productivity seen from the bottom up. Paleoceanography: 10: 545–562; with permission from the American Geophysical Union.)



not severely depressed, making it difficult for paleooceanographers to disentangle the influence of these two variables. Finally, biological factors such as microhabitat preferences and the exploitation of phytodetrital aggregates (‘floc’) influence the stable isotope chemistry of foraminiferal tests. There are many examples of the use of benthic foraminiferal faunas to interpret the geological history of the oceans. Only one is given here. Cores collected at 501410 N, 211520 W (3547 m water depth) and 581370 N, 191260 W (1756 m water depth) were used by E. Thomas and colleagues to study changes in the North Atlantic over the past 45 000 years. The cores yielded fossil specimens of two foraminiferal species, Epistominella exigua and Alabaminella weddellensis, both of which are associated with seasonal inputs of organic matter (phytodetritus) in modern oceans. In the core from 511N, these ‘phytodetritus species’ were uncommon during the last glacial maximum but increased sharply in absolute and relative abundance during the period of deglaciation 15 000–16 000 years ago (Figure 9). At the same time there was a decrease in the abundance of Neogloboquadrina pachyderma, a planktonic foraminifer found in polar regions, and an increase in the abundance of Globigerina bulloides, a planktonic species characteristic of warmer water. These changes were interpreted as follows. Surface primary productivity was low at high latitudes in the glacial North Atlantic, but was much higher to the south of the Polar Front. At the end of the glacial period, the ice sheet shrank and the Polar Front retreated northwards. The 511N site was now overlain by more productive surface water characterized by a strong spring bloom and a seasonal flux of phytodetritus to the seafloor. This episodic food source favored opportunistic species, particularly E. exigua and A. weddellensis, which became much more abundant both in absolute terms and as a proportion of the entire foraminiferal assemblage.

Conclusions Benthic foraminifera are a major component of deepsea communities, play an important role in ecosystem functioning and biogeochemical cycling, and are

347

enormously diverse in terms of species numbers and test morphology. These testate (shell-bearing) protists are also the most abundant benthic organisms preserved in the deep-sea fossil record and provide powerful tools for making paleo-oceanographic reconstructions. Our understanding of their biology has advanced considerably during the last two decades, although much remains to be learnt.

See also Abrupt Climate Change. Anthropogenic Trace Elements in the Ocean. Benthic Organisms Overview. Cenozoic Oceans – Carbon Cycle Models. Deep-Sea Fauna. Floc Layers. Macrobenthos. Meiobenthos. Microphytobenthos. Ocean Carbon System, Modeling of. Phytoplankton Blooms. Primary Production Processes. Radiocarbon. Stable Carbon Isotope Variations in the Ocean. Tracers of Ocean Productivity.

Further Reading Fischer G and Wefer G (1999) Use of Proxies in Paleoceanography: Examples from the South Atlantic. Berlin: Springer-Verlag. Gooday AJ, Levin LA, Linke P, and Heeger T (1992) The role of benthic foraminifera in deep-sea food webs and carbon cycling. In: Rowe GT and Pariente V (eds.) Deep-Sea Food Chains and the Global Carbon Cycle, pp. 63--91. Dordrecht: Kluwer Academic. Jones RW (1994) The Challenger Foraminifera. Oxford: Oxford University Press. Loeblich AR and Tappan H (1987) Foraminiferal Genera and their Classification, vols 1, 2. New York: Van Nostrand Reinhold. Murray JW (1991) Ecology and Palaeoecology of Benthic Foraminifera. New York: Wiley; Harlow: Longman Scientific and Technical. SenGupta BK (ed.) (1999) Modern Foraminifera. Dordrecht: Kluwer Academic. Tendal OS and Hessler RR (1977) An introduction to the biology and systematics of Komokiacea. Galathea Report 14: 165--194, plates 9–26. Van der Zwan GJ, Duijnstee IAP, den Dulk M, et al. (1999) Benthic foraminifers:: proxies or problems? A review of paleoecological concepts. Earth Sciences Reviews 46: 213--236.


BENTHIC ORGANISMS OVERVIEW P. F. Kingston, Heriot-Watt University, Edinburgh, UK Copyright& 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 286–295, & 2001, Elsevier Ltd.

Introduction The term benthos is derived from the Greek word baos (vathos, meaning depth) and refers to those organisms that live on the seabed and the bottom of rivers and lakes. In the oceans the benthos extends from the deepest oceanic trench to the intertidal spray zone. It includes those organisms that live in and on sediments, those that inhabit rocky substrata and those that make up the biodiversity of coral reefs. The benthic environment, sometimes referred to as the benthal, may be divided up into various well defined zones that seem to be distinguished by depth (Figure 1).

Physical Conditions Affecting the Benthos In most parts of the world the water level of the upper region of the benthal fluctuates, so that the animals and plants are subjected to the influence of the water only at certain times. At the highest level, only spray is involved; in the remainder of the region, the covering of water fluctuates as a result of tides and wind and other atmospheric factors. Below the

0m 7.6%

level of extreme low water, the seafloor is permanently covered in water. Water pressure increases rapidly with depth. Pressure is directly related to depth and increases by one atmospheric (101 kPa) per 10 m. Thus at 100 m, water pressure would be 11 atmospheres and at 10 000 m, 1001 atmospheres. Water is essentially incompressible, so that the size and shape of organisms are not affected by depth, providing the species does not possess a gas space and move between depth zones. Light is essential to plants and it is its rapid attenuation with increasing water depth that limits the distribution of benthic flora to the coastal margins. Daylight is reduced to 1% of its surface values between 10 m and 30 m in coastal waters; in addition, the spectral quality of the light changes with depth, with the longer wavelengths being absorbed first. This influences the depth zonation of plant species, which is partially based on photosynthetic pigment type. Surface water temperatures are highest in the tropics, becoming gradually cooler toward the higher latitudes. Diurnal changes in temperature are confined to the uppermost few meters and are usually quite small (31C); however, water temperature falls with increasing depth to between 0.51C and 21C in the abyssal zone. The average salinity of the oceans is 34.7 ppt. Salinity values in the deep ocean remain very near to this value. However, surface water salinity may vary considerably. In some enclosed areas, such as the Baltic Sea, salinity may be as low as 14 ppt, while in

Littoral zone (intertidal) Sublittoral zone (continental shelf)

200 m

15.3%

3000 m 75.9%

Bathyal zone (continental slope)

Abyssal zone

6000 m Hadal zone (ocean trench)

1.2%

To max. known depth, 11 045 m (Challenger Deep) Figure 1 Classification of benthic environment with percentage representation of each depth zone.

348


Salinity fluctuations

Turbulence

Temperature fluctuations

Effects of light

Depth (m)

Periodic exposure to air

BENTHIC ORGANISMS OVERVIEW

Littoral

0 10

Sublittoral

100 1000 10000

Benthic Zone

349

surface turbulence. Farther offshore, and in the deep sea, the seabed is dominated by sediments, hard substrata occurring only on the slopes of seamounts, oceanic trenches, and other irregular features where the gradient of the seabed is too steep to permit significant accumulation of sedimentary material. Thus, viewed in its entirety, the seabed can be considered predominantly a level-bottom sedimentary environment (Figure 3).

Bathyal More or less constant conditions

Abyssal Hadal

Figure 2 Diagram summarizing the influence of physical factors on the benthos. Relative influence of each factor indicated by width of shaded area.

Classification of the Benthos It was the Danish marine scientist, Petersen working in the early part of the twentieth century, who first defined the two principal groups of benthic animals:

• areas of high evaporation, such as the Arabian Gulf, salinities in excess of 50 ppt may be reached. Salinity profiles may become quite complex in estuarine conditions or in coastal regions where there is a high fresh-water run off from the land (e.g., fiordic conditions) (Figure 2). The composition of the benthos is profoundly affected by the nature of the substratum. Hard substrata tend to be dominated by surface dwelling forms, providing a base for the attachment of sessile animals and plants and a large variety of microhabitats for organisms of cryptic habits. In contrast, sedimentary substrata are dominated by burrowing organisms and, apart from the intertidal zone and the shallowest waters, they are devoid of plants. Hard substrata are most common in coastal waters where there are strong tidal currents and

•

the epifauna, comprising all animals living on or associated with the surface of the substratum, including sediments, rocks, shells, reefs and vegetation; the infauna comprising all animals living within the substratum, either moving freely through it or living in burrows and tubes or between the grains of sediments.

According to the great benthic ecologist Gunnar Thorson, the epifauna occupies less than 10% of the total area of the seabed, reaching its maximum abundance in the shallow waters and intertidal zones of tropical regions. However, the infauna, which Thorson believed occupies more than half the surface area of the planet, is most fully developed sublittorally. Nevertheless, the number of epifaunal species is far greater than the number of infaunal species. This is because the level-bottom habitat

100 50

10

20

30

Percent area above given depth 50 40 60 70

80

90

200 500 1000 2000

75 50 Percent at 25 given depth

3000 4000 5000 6000

6000

7000

Rock,gravel and hydrogenous crust 7000 Detrital sand and silt 8000 Reducing clay Oxidizing clay 9000 Siliceous oozes (diatom and radiolarian) Calcareous oozes (pteropod, foraminifera coccolith) 10 000 m

8000 9000 10 000 m

100 50 0

Figure 3 Distribution of sediment types in the ocean. (After Brunn, in Hedgepeth (1957).)


350


Macrofauna

Size (mm)

Meiofauna

Microfauna

Turbellaria Ciliates

0.05

Gastrotricha

0.5

Copepoda

Nematoda

Polychaeta

Nemertini

Amphipoda

Foraminifera

1.0

Echinodermata

1.5

Decapod crustaceans

5.0

0.005

Increasing number of species

Figure 4 Principal taxa associated with each of the major categories of infauna.

provides a more uniform environment than hard substrata, with fewer potential habitat types to support a wide diversity of species. The epifauna of polar regions are not generally as well developed as those in the lower latitudes because of the effects of low temperatures in shallow waters and the effects of ice and meltwater in the intertidal region. The infauna largely escape these effects, exhibiting less latitudinal variation in number of species (Figure 4). The infauna is further classified on the basis of size, the size categories broadly agreeing with major taxa that characterize the groups (Figure 5):

• • •

macrofauna – animals that are retained on a 0.5 mm aperture sieve; meiofauna – animals that pass a 0.5 mm sieve but are retained on 0.06 mm sieve; microfauna – animals that pass a 0.06 mm sieve.

Petersen was also the first marine biologist to quantitatively sample soft-bottom habitats. After examining hundreds of samples from Danish coastal waters, he was struck by the fact that extensive areas of seabed were occupied by recurrent groups of species. These assemblages differed from area to area, in each case only a few species making up the bulk of the individuals and biomass. This contrasted with nonquantitative epifaunal dredge samples taken over the same range of areas, for which faunal lists

might be almost identical. Petersen proposed the infaunal assemblages that he had distinguished as communities and named them on the basis of the most visually dominant animals (Table 1). Following the publication of Petersen’s work in 1911–12, other marine biologists began to investigate benthic infauna quantitatively, and it began to emerge that there existed parallel bottom communities in which similar habitats around the world supported similar communities to those found by Petersen. These communities, although composed of different species, were closely similar, both ecologically and taxonomically, the characteristic species belonging to at least the same genus or a nearby taxon (Table 2). Although the concept of parallel bottom communities appeared to hold good for temperate waters, in tropical regions such communities are not clearly definable, because of the presence of very large numbers of species and the small likelihood of any particular species or group of species dominating. Not all benthic ecologists believe in Petersen’s communities as functional biological units, since it is clear that abiotic factors such as sediment type must play a central role in determining species distribution. Alternative approaches have been proposed in which benthic associations are linked to



High Arctic

Warm temperate

Increasing number of species

Cold temperate

351

Tropical

a

un

fa pi

E

Infauna

90˚

50˚

70˚

30˚

10˚

Decreasing latitude Figure 5 Relationship between numbers of epifaunal/infaunal species and latitude.

substratum (and latitude), suggesting that it is a common environmental requirement that forms the basis of the perceived community rather than an affinity of the members of the assemblage with one another.

Feeding Habits of Benthic Animals Benthic animals, like all other animals, ultimately rely on the plant kingdom as their primary source of food. Most benthic animals are dependent on the rain of dead or partially decayed material (organic detritus) from above, and it is only in shallow coastal waters that macrophytes and phytoplankton are

Table 1

directly available for grazing or filtering by bottomliving forms. Much of this material consists of cellulose (dead plant cells) and chitin (crustacean exoskeletons). Few animals are able to digest these substances themselves and most rely on the action of bacteria to render them available, either as bacterial biomass or breakdown products. Food particles are intercepted in the water before they reach the seabed, or are collected from the sediment surface after they have settled out, or are extracted from the sediment after becoming incorporated into it. These three scenarios reflect the three main types of detrivores: suspension feeders, selective deposit feeders, and direct deposit feeders.

Petersen’s benthic communities

Petersen’s community

Typical species

Substratum

Macoma or Baltic community Abra community Venus community Echninocardium–Filiformis community

Macoma balthica, Mya arenaria, Cerastoderma edule Abra alba, Macoma calcarea, Corbula gibba, Nephtys spp. Venus striatula, Tellina fabula, Echinocardium cordatum E. cordatum, Amphiura filiformis, Cultellus pellucidus, Turritella communis Brissopsis lyrifera, Amphiura chiagei, Calocaris macandreae B. lyrifera, Ophiura Sarsi, Abra nitida, Nucula tenuis Amphilepis norvegica, Chlamys ( ¼ Pecten) vitrea, Thyasira flexuosa Haploops tubicola, Chlamys septemradiata, Lima loscombi Venus gallina, Spatangus purpureus, Abra prismatica

Intertidal mud Inshore mud Offshore sand Offshore muddy sand

Brissopsis–Chiagei community Brissopsis–Sarsi community Amphilepis–Pecten community Haploops community Deep Venus community


Offshore mud Deep mud Deep mud Offshore clay Deep sand

352


Table 2 Examples of parallel bottom communities identified by the Danish ecologist Gunnar Thorson

Table 3 Percentage representation of different feeding types from sandy and muddy sediments

Species

Feeding types

Sandy sediment (o15 silt/clay) (%)

Muddy sediment (490% silt/clay) (%)

Predators/omnivores Suspension feeders Selective deposit feeders Direct deposit feeders

25.3 27.2 41.1

22.4 14.9 50.8

6.3

11.9

Macoma Cardium Mya Arenicola

•

•

•

Genera NE Atlantic

NE Pacific

Arctic

NW Pacific

baltica edule arenaria marina

nasuta corbis arenaria claparedii

calcarea ciliatum truncata marina

incongrua hungerfordi

Suspension feeders may be passive or active. Passive suspension feeders trap passing particles on extended appendages that are covered in sticky mucus and rely on natural water movements to bring the food to them (e.g., crinoid echinoderms). Active suspension feeders create a strong water current of their own, filtering out particles using specially modified organs (e.g., most bivalve mollusks). Selective deposit feeders either consume surface deposits in their immediate vicinity using unspecialized mouth parts, or, where food is less abundant, use extendable tentacles or siphons to pick up particles over a large area (e.g., terebellid polychaetes). Direct deposit feeders indiscriminately ingest sediment using organic matter and microbial organisms contained in the sediment as food. Polychaeta, such as the lugworm, Arenicola, construct L-shaped burrows and mine sediment from a horizontal gallery, reversing up the vertical shaft to defecate at the surface. Such animals play an important role in the physical turnover of the sediments.

Grazing or browsing animals are most common intertidally or in shallow waters. They are mobile consumers, cropping exposed tissues of sessile prey usually without killing the whole organism. They include animals that feed on macroalgae and those that feed on colonial cnidaria, bryozoans, and tunicates such as gastropod mollusks and echinoids. On the level bottom, the tips of tentacles and siphons of infaunal animals are grazed by demersal fish, providing a route for energy transfer from the seabed back into the pelagic system. Predatory hunters are common among benthic epifauna and include crustaceans, asteroid echinoderms, and gastropod mollusks. These are mobile animals that seek out and consume individual prey items one at a time. Although less common, such predators are also found in the infauna, moving through the sediment, attacking their prey in situ (e.g., the polychaete Glycera).

Although, overall, deposit feeders form the largest single group of benthic infauna, the proportion of each trophic group is greatly influenced by the nature of the sediment. Thus in coarser sandy sediments, where water movement is relatively strong, the proportion of suspension feeders increases, while fine silts and muds are usually dominated by deposit feeders (see Table 3).

Spatial Distribution of Benthos Competition between benthic infauna is usually for space. This is because most benthic animals are either suspension or deposit feeders and are competing for access to the same food source. In this respect, benthic communities are similar to those of terrestrial plants since, in both, competition between individuals is for an energy source that originates from above. Indeed, early approaches to the statistical analysis of benthic community structure were often rooted in principles originally developed by terrestrial botanists. Suspension feeding may take place at more than one level. Some species, such as sabellid worms, extend their feeding organs several centimeters into the water column; some keep a lower profile, with short siphons projecting just a few millimeters above the sediment surface, while others have open burrows, drawing water into galleries below the surface. In addition, selective deposit feeders scour the sediment surface for food particles using palps or tentacles that in some species can extend up to a meter. The result is a contagious horizontal distribution of animals (i.e., neither random nor regular) that is maintained primarily by inter- and intra-specific competition for space. Benthic infauna also show a marked vertical distribution. This is more a function of the subsurface physical conditions of the sediments and the need for the majority of species to be in communication with the surface than of competition between the animals.



Petersen chose physically large representatives to describe his benthic communities, primarily because they were easy to identify under field conditions. However, most benthic infaunal species are too small to be recognized with the naked eye, with burrows that penetrate no more than a few centimeters into the substratum. The result is that, for most levelbottom communities, some 95–99% of the animals are located within 5 cm of the sediment surface.

353

100

Pelagic development

90

Nonpelagic development 80

70

60

%

50

40

30

20

Reproduction in Benthic Animals The act of reproduction offers benthic animals, the majority of which are either sessile or very restricted in their migratory powers, an opportunity to disperse and to colonize new ground. It is therefore not surprising that the majority of benthic species experience at least some sort of pelagic phase during their early development. Most invertebrates have larvae that swim for varying amounts of time before settlement and metamorphosis. The larvae, which develop freely in the surface waters of the ocean, either feed on planktonic organisms (planktotrophic larvae) or develop independently from a self-contained food supply or yolk (lecithotrophic larvae). Pelagic development in temperate waters can take several weeks, during which time developing larvae may be transported over great distances. Where it is within the interests of a particular species to ensure that its offspring are not dispersed (e.g., some intertidal habitats), a free-living larval phase may be dispensed with. In this case eggs may develop directly into miniature adults (oviparity) or may be retained within the body of the adult with the young being born fully developed (viviparity). Reproductive strategies such as these are also common in the deepsea and polar regions where the supply of phytoplankton for feeding is unreliable or nonexistent. A good example of a latitudinal trend in this respect was demonstrated by Thorson. Analysing the developmental types of prosobranchs, he was able to show that the proportion of species with nonpelagic larvae decreases from the arctic to the tropics, while the proportion with pelagic larvae increases (Figure 6). For many years deep-sea biologists believed that the energetic investment required to produce large numbers of planktotrophic larvae, and the huge distances required to be covered by such larvae in order to reach surface waters, would preclude such a reproductive strategy for deep-sea animals. However, it is now known that several species of ophiuroids living at depths of 2000–3000 m not only exhibit seasonal reproductive behavior but also produce larvae that feed in ocean surface waters.

10

0 E.Greenland

N. & E. Iceland W. & S. Iceland

Faroes, Shetlands, Orkneys

S. Norway, W S. England, Sweden. Channel Islands Denmark

Canary Islands

Figure 6 Percentage distribution of prosobranchs with pelagic and nonpelagic development in relation to latitude. (Adapted from Thorson (1950).)

Many benthic invertebrates are able to reproduce asexually. For example, polychaetes from the family the Syllidae are able to reproduce by budding; others, such as the cirratulid Dodecaceria or the ctenodrilid Raphadrilus, simply fragment, each fragment growing into a new individual. The ability to switch between sexual and vegetative means of propagation provides the potential for such species to rapidly colonize areas that have been disturbed. Where disturbance is accompanied by organic enrichment, for example, from sewage or paper pulp discharge, huge localized populations may result. These are the socalled opportunistic species that are sometimes used as indicators of pollution. Although planktonic larvae are able to swim, they are very small and, for the most part, are obliged to go where ocean currents take them. The critical time arrives just before the larvae are about to settle. At one time it was thought that the process of settlement was random, with individuals that settled in unfavorable substrata perishing. Although this undoubtedly happens, most species seem to have some sort of behavioral pattern to increase their chances of finding a suitable substratum. The larvae usually pass through one or more stages of photopositive and photonegative behavior. These enable the larvae to remain near the sea surface to feed and then to drop to the bottom to seek a suitable substratum on which to settle. Depending on the species, larvae may cue on the mechanical attributes of the substratum or on its chemical nature. Chemical attraction is also important in gregarious species in which the young are attracted to settle at sites where adults of the same species are already present (e.g., oysters). Most larvae go through a period when, although able to settle


354


permanently, they retain the ability to swim. This allows them to ‘test’ the substratum, rising back into the water and any prevailing currents should the nature of the ground be unsuitable. After settling, larvae may move a short distance, usually no more than a few centimeters. These early stages in the recruitment of benthic organisms are crucial in the maintenance of benthic community structure and it is now believed that it is at this time that the nature of the community is established. It is clear that the vast majority of planktonic larvae never make it to adulthood. Mortality from predation and transport away from a suitable habitat are on a massive scale. To compensate, species with planktotrophic larvae produce huge numbers of eggs (e.g., the sea hare Aplysia californiensis spawns as many as 450 000 000 eggs at one time). This is possible because there is no need for a large, and energetically expensive, yolk; the larvae hatch at an early embryonic stage and rely almost entirely on plankton-derived food for their development. One consequence of this is that the recruitment varies depending on the success of the plankton production in a particular year and the vagaries of local currents. Thus, populations of benthic species that reproduce by means of planktotrophic larvae tend to fluctuate numerically from year to year, with the potential for heavy recruitment when the combination of environmental factors is favorable, or recruitment failure when they are not. Species reproducing by means of nonpelagic larvae or by direct development tend to produce fewer eggs, since there is a large yolk required to nourish the developing embryo. Although annual recruitment is relatively modest for these species, it is less variable between years, producing populations with a greater temporal stability (Figure 7). Because of this, these populations are 400 350

Corbula gibba (long pelagic life) Nucula nitida (very short pelagic life)

Biomass (g m−2)

300 250 200 150 100 50 0 1910 11 12 13

14

15 16

17 18

19

20 21 22 23 1924

Year

Figure 7 Example of two populations of bivalves showing the influence of type of larvae on population stability. (Adapted from Thorson (1950).)

likely to be slow to recover from major natural environmental disturbances (e.g., unusual temperature extremes or physical disturbance) or major pollution events.

Deep-sea Benthos Because of regional differences, it is difficult to define the exact upper limit of the deep-sea benthic environment. However, it is generally regarded as beginning beyond the 200 m depth contour. At this point, where the continental shelf gives way to the continental slope, there is often a marked change in the benthic fauna. In the past there have been many attempts to produce a scheme of zonation for the deep-sea environment. There are three major regions beyond 200 m depth – these are the bathyal region, the abyssal region, and the hadal region (see Figure 1). The bathyal region represents the transition region between the edge of the continental shelf and the true deep sea (the continental slope). Its boundary with the abyssal region has been variously defined by workers. It is believed that the 41C isotherm limits the depth at which the endemic abyssal organisms can survive. Since this varies in depth according to geographical and hydrographical conditions in the abyssal region, it follows that the upper depth limit of the abyssal fauna will also vary. This depth is usually between 1000 and 3000 m. The abyssal region is by far the most extensive, reaching down to 6000 m depth and accounting for over half the surface area of the planet. The hadal zone, sometimes called the ultra-abyssal zone, is largely restricted to the deep oceanic trenches. The composition of the benthos in these trenches differs from that of nearby abyssal areas. The trenches are geographically isolated from one another and the fauna exhibits a high degree of endemism. The deep sea is aphotic and so has no primary production except in certain areas where chemosynthetic bacteria are found. Thus, the fauna of the deep sea is almost wholly reliant on organic material that has been generated in the surface layers of the oceans and has sunk to the seabed. Since the likelihood of a particle of food being consumed on its way down through the water column is related to time, the deeper the water the less food is available for the animals that live on the seafloor. This results in a relatively impoverished fauna, in which the density of organisms is low and the size of most is quite small. Paradoxically, the deep sea supports a few rare species that grow unusually large. This phenomenon is known as abyssal gigantism and is found primarily in crustacean species. The reasons for abyssal



gigantism are not clear, but it is believed to be the result either of a peculiar metabolism under conditions of high pressure or of the slow growth and time taken to reach sexual maturity. Although the number of individuals in the deep sea is quite low, there are many species. This combination of many species represented by few individuals results in high calculated diversity values and has led to the suggestion by some that the biodiversity of the deep sea is comparable to that of tropical rain forests. Accepting that diversity is high in the deep sea, there are several theories that attempt an explanation. The earliest is the stability–time hypothesis put forward by Howard Sanders of the Woods Hole Oceanographic Institution in the late 1960s. This suggests that the highly stable environmental conditions of the deep sea that have persisted over geological time might have allowed many species to evolve that are highly specialized for a particular microhabitat or food source. Another theory, the cropper or disturbance theory, suggests that, as a result of the scarcity of food in the deep sea, none of the animals is a food specialist, the animals being forced to feed indiscriminately on anything living or dead that is smaller than itself. High diversity results from intense predation, which allows a large number of species to persist, eating the same food, but never becoming abundant enough to compete with one another. More recently, Grassle and Morse-Porteus have suggested that a combination of factors might be responsible, including the patchy distribution of organically enriched areas in a background of low productivity; the occurrence of discrete, small-scale disturbances (primarily biological) in an area of otherwise great constancy; and the lack of barriers to dispersal among species distributed over a very large area. Although most animal groups are represented in the deep sea, the fauna is often dominated by Holothuroidea (sea cucumbers) or Ophiuroidea (brittle stars). Crustacea and polychaete worms are also important members of deep-sea communities. For many years it was believed that the deep seafloor was too remote from the surface, and the physical conditions were too constant, for the organisms living there to be influenced by the seasons. Although this may be the case for many deep-sea species, longterm time-lapse photography has shown that cyclical events, such as the accumulation of organic detritus on the deep seabed, do take place. Furthermore, in temperate waters, these appear to correspond with seasonally driven processes such as the spring plankton bloom. Where they occur, these pulses of organic input inevitably influence the deep-sea communities below with the consequence that seasonal

355

life cycles are not uncommon in abyssal animals. Localized areas of organic enrichment can occur in the deep sea as the result of the sinking of large objects such as the carcass of a large sea mammal or fish or waterlogged tree trunk. The surprisingly quick response of deep-sea scavengers to large food-falls such as these, and the frequency with which they have been recorded, has led researchers to believe that they are important contributers to energy flow on the deep-sea floor. Hydrothermal vents are also thought to provide a significant input of energy into the benthic environment. These are a relatively recent discovery and, at first, were thought to be rare and isolated phenomena occurring only on the Galapagos Rift off Ecuador. It is now known that active vents are associated with nearly all areas of tectonic activity that have been investigated in the deep Pacific and Atlantic. These vents provide a nonphotosynthetic source of organic carbon through the medium of chemoautotrophic bacteria. Theses organisms use sulfur-containing inorganic compounds as an oxidizing substrate to synthesize organic carbon from carbon dioxide and methane without the need for sunlight. The chemicals come from hot water that originates deep within the Earth’s crust. Some of the bacteria form dense white mats on the surface of the sediments similar to those of Beggiatoa, an anaerobic bacterium found in anoxic sediments in shallow water; others enter into symbiotic relationships with the bacteria, either hosting them on their gills (e.g., the mussel Bathymodiolus) or within a special internal sac, the trophosome (e.g., Riftia). The relationship is very complex as there has to be a compromise between the anaerobic needs of the bacteria and the aerobic needs of the animals. Nevertheless, the arrangement is very successful and the animals, which sometimes occur in huge numbers, often grow to gargantuan size. There is still much to be learned about these vent communities, which can support concentrations of biomass several orders of magnitude greater than that of the nearby seafloor. It remains a mystery how these vents become populated, since they are known to be transient and variable, probably lasting only decades or less. It has been suggested that pelagic larvae of many vent species may be able to delay settlement for months at a time so as to increase their chances of locating a suitable site.

Glossary Abyssal region That region of the seabed from between 1000 and 3000 m depth reaching down to 6000 m. Atmosphere Measure of pressure (1 atm ¼ 101kPa).


356


Bathyal region Region of the seabed that represents the transition region between the edge of the continental shelf and the true deep sea (the continental slope). Benthal The benthic environment. Benthos Those organisms that live on the seabed and the bottoms of rivers and lakes. Continental shelf Region of the seabed from lowwater mark to a depth of 200 m. Detritivores Animals feeding on organic detritus. Direct deposit feeders Animals indiscriminately ingesting sediment using organic matter and microbial organisms contained in the sediment as food. Epifauna All animals living on or associated with the surface of the substratum. Hadal zone Region of the seabed below 6000 m depth, largely restricted to the deep oceanic trenches. Infauna All animals living within the substratum or moving freely through it. Kilopascal (kPa) Measure of pressure (100 kPa ¼ 1 bar ¼ 0.987 atm). Lecithotrophic larvae Pelagic larvae of marine animals that develop freely in the surface waters of the ocean, developing independently from a selfcontained food supply or yolk. Macrofauna Animals retained on a 0.5 mm aperture sieve. Meiofauna Animals passing a 0.5 mm sieve but retained on a 0.06 mm sieve. Microfauna Animals passing a 0.06 mm sieve. Organic detritus Dead or partially decayed plant and animal material. Oviparity Eggs laid by the adult develop directly into miniature adults. Pelagic larvae Larvae of marine animals that swim freely in the water column. Planktotrophic larvae Pelagic larvae of marine animals that develop freely in the surface waters of the ocean, feeding on planktonic organisms. Selective deposit feeders Animals feeding on surface particles of organic matter or sediment particles supporting a rich bacterial flora.

Suspension feeders Animals feeding on organisms or organic detritus suspended in the water column. Viviparity Young are born fully developed either from eggs retained within the body of the mother (oviparity) or after internal embryonic development.

See also Benthic Boundary Layer Effects. Benthic Foraminifera. Deep-Sea Fauna. Demersal Species Fisheries. Fish: Demersal Fish (Life Histories, Behavior, Adaptations). Grabs for Shelf Benthic Sampling. Macrobenthos. Meiobenthos. Microphytobenthos. Network Analysis of Food Webs. Phytobenthos. Tides.

Further Reading Cushing DH and Walsh JJ (eds.) (1976) The Ecology of the Seas. Oxford: Blackwell Scientific Publications. Friedrich H (1969) Marine Biology: An Introduction to Its Problems and Results. London: Sidgwick and Jackson. Gage JD and Tyler PA (1991) Deep-sea Biology: A Natural History of Organisms at the Deep-Sea Floor. Cambridge: Cambridge University Press. Hedgepeth JW (ed.) (1957) Treatise on Marine Ecology and Paleoecology, Vol. 1, Ecology, The Geological Society of America, Memoir 67. Washington, DC: Geological Society of America. Jøgensen CB (1990) Bivalve Filter-Feeding: Hydrodynamics, Bioenergetics, Physiology and Ecology. Fredensborg, Denmark: Olsen and Olsen. Levington JS (1995) Marine Biology: Function, Biodiversity, Ecology. Oxford: Oxford University Press. Nybakken JW (1993) Marine Biology. New York: HarperCollins College Publishers. Thorson G (1950) Reproduction and larval ecology of marine bottom invertebrates. Biological Reviews 25: 1--25. Webber HH and Thurman HV (1991) Marine Biology. New York: Harper Collins.


BIOACOUSTICS P. L. Tyack, Woods Hole Oceanographic Institution, Woods Hole, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 295–302, & 2001, Elsevier Ltd.

marine mammal. Marine mammal bioacoustics during this period was concerned primarily with identifying which species produced which sounds heard under water. Much of this research was funded by naval research organizations because biological sources of noise can interfere with military use of sound in the sea.

Introduction The term ‘bioacoustics’ has two different usages in ocean sciences. Biological oceanographers use active sonars to map organisms in the sea. Since they use sound to detect marine life, they often call this approach ‘bioacoustics’. The other sense of bioacoustics involves studying how animals use sound themselves in the ocean. This is the kind of bioacoustics covered in this article. Humans are visual animals, and we think of vision as a primary distance sense because light carries so well in terrestrial environments. However, light is useful for vision under the sea only over ranges of tens of meters at best. Sound, on the other hand, propagates extremely well in water – that is why oceanographers so often select sound as a medium for exploring the sea or for communicating under the sea. Sound propagates so well under water that a depth charge exploded off Australia can be heard in Bermuda. Just as we can hear well but emphasize vision, so many marine mammal species see well but emphasize hearing. It is possible to gauge the relative importance of audition versus vision in animals by comparing the number of nerve fibers in the auditory versus the optic nerves. Of all marine mammals, the cetaceans are the most specialized to use sound. Most cetaceans have auditory:optic ratios of fiber counts that are 2–3 times those of land mammals, suggesting that audition is more important than vision. Some cetaceans also use sound to echolocate. Dolphins have a large repertoire of vocalizations spanning frequencies from below 100 Hz to over 100 kHz, and dolphins have evolved high-frequency echolocation similar to some human-made sonars and to the biosonar used by bats. Marine mammals not only hear well, they are also very vocal animals. The sounds of marine mammals are now well known, but the first recordings identified from a marine mammal species were only made in the late 1940s. In the 1950s and 1960s, there was rapid growth in studies of how dolphins echolocate using high-frequency click sounds and of field studies associating different sounds with different species of

Elementary Acoustics Sound consists of mechanical vibrations that propagate through a medium. Sound induces movements or displacements of the particles in the medium. Imagine a small sphere that expands to create a denser area. This compression will propagate as particles are displaced in the direction of propagation. If the sphere then contracts, it can create an area of rarefaction, or lower density, and this also can propagate outward. These compressions or rarefactions can be expressed in terms of particle displacement or as a pressure differential. Now imagine a sound source that creates a series of compressions and rarefactions that propagate through the medium. A source with a purely sinusoidal pattern of compression and rarefaction would produce energy at only one frequency. The frequency of this sound is measured in cycles per second. A sound that takes t seconds to make a full cycle has a frequency f ¼ t1. Older references may refer to frequency in cycles per second, but the modern unit of frequency is the Hertz (Hz) and a frequency of 1000 Hz is expressed as one kiloHertz (1 kHz). If a sound took 1 s for a full cycle, it would have a frequency of 1 Hz. The wavelength of a tonal sound is the distance from one measurement of the maximum pressure to the next maximum. The speed of sound is approximately 1500 m s1 in water, roughly five times the value in air, 340 m s1. The speed of a sound c is related in a simple way to the frequency f and the wavelength l by c ¼ lf. An under-water sound with f ¼ 1 Hz would have l ¼ 1500 m; for f ¼ 1500 Hz, l ¼ 1 m. Not all sounds have energy limited to one frequency. Sounds that have energy in a range of frequencies, say in the frequency range between 2000 and 3000 Hz (2 and 3 kHz), would be described as having a bandwidth of 1 kHz. One can imagine a sound wave as a growing sphere propagating outward from a compression or rarefaction generated by a point source. The initial movement of the source will have transmitted a certain amount of energy to the medium. If none of this


357

358

BIOACOUSTICS

energy is lost as the sound propagates, then it will be evenly diluted over the growing sphere. The acoustic intensity is defined as the amount of energy flowing through an area over a unit of time. As the sphere increases in radius from 1 to r, the surface area increases to 4pr2. The intensity of a sound thus declines as the inverse of the square of the range from the source (r2). A sound in the middle of the ocean can be thought of as spreading in this way until it encounters a boundary such as the surface or seafloor that might cause reflection, or an inhomogeneity in the medium that might cause refraction. One fascinating acoustic feature of the deep ocean is that sound rays propagating upward may refract downward as they encounter warmer water near the surface, and downward-propagating rays will refract upward as they encounter denser water at depth. When one is far from a sound source compared to the ocean depth, the sound energy may be concentrated by refraction in the deep ocean sound channel. This sound can be thought of as spreading in a plane, to a first approximation. In this case, sound intensity would decline as the inverse of the first power of the range, or r1. This involves much lower loss than the inverse square spreading loss in an unbounded medium. Sound spreading is a ‘dilution’ factor and is not a true loss of sound energy. Absorption, on the other hand, is conversion of acoustic energy to heat. The attenuation of sound due to absorption is a constant per unit distance, but this constant is dependent upon signal frequency. While absorption yields trivial effects at frequencies below 100 Hz, it can significantly limit the range of higher frequencies, particularly above 40 kHz or so. A 100 Hz sound can travel over a whole ocean basin with little absorption loss, while a 100 kHz sound would lose half its energy just traveling about 100 m.

36 Hz band and they last several tens of seconds. The pulses of finback whales, Balaenoptera physalus, range roughly between 15 and 30 Hz and last on the order of 1 s. Particularly during the breeding season in mid-latitudes, finbacks produce series of pulses in a regularly repeating pattern in bouts that may last many days. These loud low-frequency sounds appear to be specialized for long-range propagation in the sea. Absorption is negligible at the frequencies of these sounds. While acoustic models predicted that these sounds could be detected at ranges of hundreds of kilometers, it is only recently that this has been confirmed empirically. During the Cold War, the US Navy developed bottom-mounted hydrophones to locate ships and to track them. After the end of the Cold War, these sophisticated systems were made available to biologists, who have worked with Navy personnel to locate and track whales over long ranges, including one whale tracked for more than 1700 km over 43 days (Figure 1). These arrays have proven capable of detecting whales at ranges of hundreds to thousands of kilometers, as was predicted by the earlier acoustic models. The physics of sound can also help explain why dolphins specialize in high-frequency sounds. Dolphins can detect distant objects acoustically by producing loud clicks and then listening for echoes. The clicks used by dolphins for echolocation have been well described. The echolocation clicks of bottlenose dolphins are very short (o100 ms), with a rapid risetime and a relatively broad bandwidth from several tens of kilohertz up to near 150 kHz (Figure 2A).

280°

290°

300° 40°

40°

Relating Acoustic Structure to Biological Function of Marine Mammal Calls Understanding the physics of sound in the sea can help us understand why animals make the kinds of sound they do. For example, the calls of baleen whales are low-frequency because they are adapted for long-range propagation in the deep sea. Large baleen whales have evolved abilities to produce and to hear low-frequency calls well-suited for longrange communication. Blue whales and fin whales produce the lowest-frequency signals of all marine mammals, so low that humans can barely hear them. The long moans of blue whales, Balaenoptera musculus, have fundamental frequencies in the 14–

30°

30°

20°

20°

280°

290°

300°

Figure 1 Track of a calling blue whale, Balaenoptera musculus, as it swam 1700 km over 43 days. The whale was tracked using the Integrated Underwater Sound Surveillance System (IUSS) of the US Navy. (From Figure 4.17 of Au et al. (2000).)


BIOACOUSTICS

Kaneohe Bay SL = 210 _227 dB re 1 μPa 1 0

250 μs Tank SL = 170 _185 dB re 1 μPa

1 0 1.0

22 kHz 58 kHz 100 kHz _ 10 dB _ 20 dB _ 30 dB

250 μs

|U (f )|

30 kHz 108 kHz

40° 30° 20°

0°

Tank

_10° _ 20°

0.5

40 kHz 65 kHz 0

0 0

100 Frequency (kHz)

118 kHz

10°

Kaneohe Bay

(A)

0 dB

359

45 kHz

_ 30°

150 μs

200 (B)

Figure 2 (A) Waveform and spectrum of echolocation clicks of bottlenose dolphins, Tursiops truncatus, in open ocean (Kaneohe Bay) and in a tank. The spectrum of the click from the tank (indicated with a dashed line) has a lower frequency peak at 40 kHz. (B) Beam pattern of Tursiops echolocation clicks. (mPa ¼ micropascal, reference for sound pressure measurements. SL ¼ source level.) ((A) from Figure 9.1, (B) from Figure 9.5 of Au et al. (2000).)

Captive dolphins in a reverberant pool make clicks that are less loud and lower in frequency than dolphins working on long-range echolocation in an open bay. The high-frequency components of these clicks are highly directional. If one moves 10 degrees off the axis of the beam, the click energy is halved and the click contains energy at lower frequencies (Figure 2B). The detection abilities of echolocating dolphins are truly remarkable. For example, trained bottlenose dolphins can detect a 2.54 cm solid steel sphere at 72 m, nearly a football field away. The optimal frequency of a sound used for echolocation depends upon the size of the expected target. Absorption imposes a penalty for higher frequencies, but small targets can best be detected by short-wavelength, or high-frequency, signals. In the nineteenth century, Lord Rayleigh solved the frequency dependence of sound scattering from rigid spherical targets; this is called Rayleigh scattering. A spherical target of radius r reflects maximum energy when the wavelength of the sound impinging on it equals the circumference of the sphere, or when l ¼ 2pr. The echo strength drops off rapidly from signals with wavelength l42pr. Since l ¼ c/f, one can equate the two l terms to get c/f ¼ 2pr. The relationship f ¼ c/2pr can be found by rearranging terms to calculate the optimal frequency for reflecting sound energy off a spherical target of radius r. Higher frequencies than this would still be effective sonar signals, but frequencies below f would show a strong decrease in effectiveness with decreasing frequency. A dolphin echolocating on rigid targets with a ‘radius’ of 0.5 cm should use a frequency fZc/

2pr ¼ 1500/(2p 0.005)B50kHz. This is within the frequency range of dolphin echolocation clicks, which include energy up to about 150 kHz. This upper frequency is appropriate for detecting spherical targets with radii as small as 1.5 mm. The hearing of dolphins is also most sensitive at frequencies of roughly 50–100 kHz. If dolphins have a need to echolocate on rigid targets with sizes in the 1 cm range, that helps explain why their echolocation system emphasizes these high frequencies.

Marine Mammal Hearing In order to detect sound, animals require a receptor that can transduce the forces of particle motion or pressure changes into neural signals. Most mechanoreceptors in animals involve cells with hairlike cilia on their surfaces. As these cilia move, the electric potential between the inside and the outside of the receptor cells changes, and this potential difference modifies the rate of nerve impulses that signal other parts of the nervous system. Terrestrial mammals evolved an ear that is divided into three sections: the outer, middle, and inner ear. The outer ear and middle ear function in terrestrial mammals to transduce airborne sound into vibrations of a fluid the inner ear of mammals which contains the cochlea, the organ in which sound energy is converted into neural signals. Sound enters the cochlea via the oval window and causes a membrane, called the basilar membrane, to vibrate. This membrane is mechanically tuned to vibrate at


BIOACOUSTICS

different frequencies. Near the oval window, the basilar membrane is stiff and narrow, causing it to vibrate when excited with high frequencies. Farther into the cochlea, the basilar membrane becomes wider and ‘floppier’, making it more sensitive to lower frequencies. Sensory cells at different positions along the basilar membrane are excited by different frequencies, and their rate of firing is proportional to the amount of sound energy in the frequency band to which they are sensitive. Marine mammals share basic patterns of mammalian hearing but also have varying adaptations for listening under water as opposed to in air. All marine mammals other than sirenians, the sea otter, and cetaceans spend critical parts of their lives on land or ice and some phocid seals communicate both in air and under water. The relative importance of hearing in air and under water has been compared for three pinniped species whose hearing has been tested in both environments. The California sea lion (Zalophus californianus) is adapted to hear best in air; the harbor seal (Phoca vitulina) can hear equally well in air and under water; and the northern elephant seal (Mirounga angustirostris) has an auditory system adapted for under water sensitivity at the expense of aerial hearing. The eardrum and middle ear in terrestrial mammals functions to efficiently transmit airborne sound to the inner ear where the sound is detected in a fluid. No such matching is required for an animal living in the water, and cetaceans, which are adapted exclusively for listening under water, do not have an air-filled external ear canal. The problem for cetaceans is isolating the ears acoustically, and the inner ear is surrounded by an extremely dense bone that is isolated from the skull. High-frequency sound is thought to enter the dolphin head through a thin section of bone in the lower jaw and is conducted to the inner ear via fatty tissue that acts as a waveguide. Hearing abilities have been tested for those species of marine mammals that can be held in captivity. Figure 3 shows audiograms from a dolphin, porpoise, and several pinnipeds. As discussed above, dolphins have hearing specialized to hear very high frequencies up to ten times the upper limit of human hearing. Seals have less acute hearing than do dolphins and they are less able to hear the highest frequencies. The frequency range of hearing has never been tested in baleen whales. Hearing is usually tested by training an animal, and baleen whales are so big that only a few have been kept for short periods in captivity. However, both their low-frequency vocalizations and the frequency tuning of their cochlea suggest they are specialized for low-frequency hearing.

140

Sound pressure level (dB)

360

130 120 110 100

Dolphin Sea lion

90 80

Seal

70 60

Porpoise

50 40 0.1

1

10 Frequency (kHz)

100

1000

Figure 3 Audiograms from a variety of marine mammals: dolphin Tursiops truncatus; porpoise Phocoena phocoena; sea lion Zalophus californianus; seal Phoca vitulina.

Mammalian hearing is designed to analyze the frequency content of sound. Among mammals, dolphins have extraordinarily good abilities of discriminating different frequencies. They can detect a change of as little as 0.2% in frequency, which is close to the resolution of human hearing.

Vocalizations of Marine Mammals When terrestrial carnivores and ungulates invaded the sea, they encountered new constraints and opportunities for sensing signals. The sirenians, cetaceans, phocid seals, and the walrus (Odobenus rosmarus) evolved specializations for using sound to communicate under water and to explore the marine environment; other taxa, including the otariid pinnipeds, sea otter (Enhydra lutra) and polar bear (Ursus maritimus), vocalize mainly in air. As with hearing, cetaceans show the most elaborate and extreme specializations for acoustic communication under water. The best-known acoustic displays of marine mammals are the reproductive advertisement displays called songs. The songs of humpback whales are the best known advertisement display in the cetaceans, but bowhead whales also sing. Male seals of some species repeat acoustically complex songs during the breeding season. Songs are particularly common among seals that inhabit polar waters and that haul out on ice. The songs of bearded seals, Erignathus barbatus, are produced by sexually mature adult males and are heard most frequently during the peak of the breeding season. Male walruses, Odobenus rosmarus, also produce complex visual and acoustic displays near herds of females during their breeding season. They use their lips to whistle, and also produce loud sounds of breathing that are


BIOACOUSTICS

audible in air when they surface during these displays. When they dive, displaying males produce a series of pulses under water followed by bell-like sounds. Antarctic Weddell seal males repeat under water trills (rapid alternations of notes) during the breeding season. Marine mammals also produce a broad variety of displays, including threat displays and recognition displays used for individual or group recognition.

Mechanisms of Sound Production Most terrestrial mammals produce vocal sounds by vibrating vocal cords in the larynx. It is thought that the polar bear and most pinnipeds make sounds using similar mechanisms. Some adaptations for diving may affect vocalization mechanisms in pinnipeds. Pinnipeds have a more flexible trachea than do terrestrial mammals, so that air inside can compress during a dive, and they have a wider trachea to allow higher rates of air flow. Most pinnipeds can vocalize under water without emitting bubbles; some species have sacs attached to the trachea or upper respiratory sac, but the role of these in vocalization has not been determined. Walruses have

361

many ways of producing sounds. They produce gonglike impulse sounds using specialized pharyngeal sacs, and can even use their lips to whistle in air. Odontocetes have well developed vocal folds in the larynx, but most biologists argue that odontocetes produce sounds as air flows past the nasal plugs or phonic lips in the upper nasal passages (Figure 4A). Mechanisms for sound production must also match the acoustic impedance to the medium of air or sea water, and they may function to direct some sounds in a beam. The beam pattern of dolphin clicks (shown in Figure 2B) stems from a complex interaction of reflection from the skull and air sacs, coupled with refraction in soft tissues (Figure 4A). There is a more detailed model of sound production for sperm whales (Physeter macrocephalus) than for other cetacean species. Sperm whales have a large organ called the spermaceti organ, which lies dorsal and anterior to the skull (Figure 4B). Below the spermaceti organ is the ‘junk’, which is composed of a series of fatty structures separated by dense connective tissue. The primary vocalizations of sperm whales are distinctive clicks comprising a burst of pulses with equally spaced interpulse intervals (IPIs). Bioacousticians suggest that these regular IPIs may result from reverberation within the

Figure 4 Functional anatomy of sound production in two odontocete cetaceans: (A) bottlenose dolphin Tursiops truncatus; (B) sperm whale Physeter macrocephalus. (Adapted from Figures 1.4 and 3.1 of Au et al. (2000).)


362

BIOACOUSTICS

Frequency (Hz) Depth (m)

ES95-1 8 May 1995 2103:27 UT

Dive #29 9 May 95 0002:32 UT

Dive #13

0 20 40 60 80 600 400 200 0 65

2

dB re 1 µPa /Hz

100

Heart rate (beats per min)

70 60

50

40

30 0

5

10 Time (min)

15

0

3

Figure 5 Data on dive profile, acoustic record, and acoustically determined heart rate from a tag on an elephant seal. The acoustic record shows a vessel passing. The closest point of approach occurs at the minimum frequency of the ‘U’ shaped pattern in the spectrogram at about 6 min. The heart rate differs little from the surrounding times or from a quiet period in a later dive from the same seal. (mPa ¼ micropascal, reference for sound pressure measurements.) (Adapted from Figure 8 of Burgess et al. (1998).)

spermaceti organ. The frontal sac at the posterior end of the spermaceti organ has been suggested as a potential reflector of sound and the distal sac as a partial reflector of sound at the anterior end (Figure 4B). The source of the sound energy in the click is thought to come from a strong valve (phonic lips) in the right nasal passage at the anterior end of the spermaceti organ (Figure 4B). This sound production model suggests that some of the energy from the first pulse within the click is transmitted directly into the water. The remaining pulses are hypothesized to occur as some of the sound energy passes through the anterior reflector into the ocean at each reflection there.

Methods for Bioacoustic Research It has been difficult to integrate visual observation of social behavior with patterns of vocalization in submerged mammals because it is difficult to identify which animal within an interacting group produces a sound under water. Biologists studying terrestrial animals take it for granted that they can identify

which animal is vocalizing by using their own ears to locate the source of a sound and then looking for movements associated with sound production. Humans cannot locate sounds under water in the same way that they locate airborne sounds. Furthermore, marine mammals seldom produce visible motions coordinated with sound production under water. It is even more difficult to attempt behavioral observations on marine mammals during a dive when they are out of sight. The need for some technique to track behavior during a dive and to identify which cetacean produces which sound during normal social interaction has been discussed for over three decades. Two different approaches have emerged: (1) passive acoustic location of sound sources using an array of hydrophones; (2) recording information about behavior and sound production by attaching a tag onto the animal. Acoustic location of vocalizing animals is a useful method for identifying which animal is producing a sound. It involves no manipulation of the animals, merely placement of hydrophones near them. In some applications, animals may vocalize frequently enough and be sufficiently separated that source


BIOACOUSTICS

location data may suffice to indicate which animal produces a sound. Tracks of continuously vocalizing finback and blue whales have been made using bottom-mounted hydrophones. Figure 1 shows a 1700 km track of a blue whale that was tracked in the early 1990s by US Navy personnel using arrays of hydrophones initially developed to track submarines. Bottom-mounted recording devices are proving cheaper alternatives for biologists today. Bioacousticians have also developed smaller, portable hydrophone arrays that can be deployed rapidly from a ship or from shore. These arrays have been used to locate vocalizing finback whales, right whales (Eubalaena glacialis), sperm whales, and several species of dolphins. Vertical hydrophone arrays can in some settings be used to calculate the range and depth of vocalizing whales. One classic configuration involves a linear horizontal array of hydrophones that is towed behind a ship. Signal processing techniques allow one to determine what bearing a sound is coming from, and to reconstruct the signal from that bearing. Bioacousticians are only just beginning to explore how to use these techniques in behavioral studies of whales. The second technique does not require locating each animal within a group. If an animal carries a telemetry device that transmits acoustic data recorded at the animal, then the device can record all vocalizations of the animals along with most everything else it hears. This kind of tag can also record depth of dive, movement and orientation of the tagged animal. However, it is difficult to telemeter information through sea water, and marine mammals might sense many of the signals one might want to use for telemetry. These problems with telemetry have led biologists to develop recoverable tags that record data while on an animal, but that need to be recovered from the animal in order for the data to be downloaded. Recently, biologists have had successful programs recovering such tags from many different kinds of marine mammal. Recoverable acoustic tags may have scientific uses well beyond identifying vocalizations. Figure 5 shows acoustic and dive data sampled from an elephant seal. The tag was able to monitor both the acoustic stimuli heard by the whale, and orientation sensors monitored not just the depth of the dive but also movement patterns such as the fluke beat and physiological parameters

363

such as heart rate. This information is useful to determine reactions of marine mammals to man-made noise, an issue of growing concern.

See also Acoustic Scattering by Marine Organisms. Acoustics, Deep Ocean. Baleen Whales. Marine Mammal Social Organization and Communication. Sea Otters. Seals. Sirenians. Sonar Systems. Marine Mammals: Sperm Whales and Beaked Whales.

Further Reading Au WWL (1993) The Sonar of Dolphins. New York: Springer Verlag. Au WWL, Popper AS and Fay R (eds.) (2000) Hearing by Whales and Dolphins. Springer Handbook of Auditory Research Series. New York: Springer Verlag. Burgess WC, Tyack PL, LeBoeuf BJ, and Costa DP (1998) A programmable acoustic recording tag and first results from free-ranging northern elephant seals. Deep-Sea Research 45: 1327--1351. Kastak D and Schusterman RJ (1998) Low-frequency amphibious hearing in pinnipeds: Methods, measurements, noise, and ecology. Journal of the Acoustical Society of America 103: 2216--2228. Medwin H and Clay CS (1998) Fundamentals of Acoustical Oceanography. New York: Academic Press. Miller P and Tyack PL (1998) A small towed beamforming array to identify vocalizing resident killer whales (Orcinus orca) concurrent with focal behavioral observations. Deep-Sea Research 45: 1389--1405. Rayleigh, Lord (1945) The Theory of Sound. New York: Dover. Tyack P (1998) Acoustic communication under the sea. In: Hopp SL, Owren MJ, and Evans CS (eds.) Animal Acoustic Communication: Recent Technical Advances, pp. 163--220. Heidelberg: Springer Verlag. Tyack PL (2000) Functional aspects of cetacean communication. In: Mann J, Connor R, Tyack PL, and Whitehead H (eds.) Cetacean Societies: Field Studies of Dolphins and Whales, pp. 70--307. Chicago: University of Chicago Press. Wartzok D and Ketten DR (1999) Marine mammal sensory systems. In: Reynolds JE III and Rommel SA (eds.) Biology of Marine Mammals, vol. 1, pp. 117--175. Washington, DC: Smithsonian Press.


BIOGEOCHEMICAL DATA ASSIMILATION E. E. Hofmann and M. A. M. Friedrichs, Old Dominion University, Norfolk, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 302–308, & 2001, Elsevier Ltd.

Introduction Data assimilation is the systematic use of data to constrain a mathematical model. It is assumed that the dynamics that are responsible for a particular process or distribution are inherent in the data. By inputting data of various types into a mathematical model, the model, which is a truncated version of the real world, will more accurately stimulate a particular environment or situation. Through data assimilation, the hindcast, nowcast, and/or forecast of the model will be improved. Data assimilation was first used in the 1960s in numerical weather forecasting models, with the goal of providing short-term predictions of meteorological conditions. The use of data assimilation techniques was made feasible by the development of a worldwide atmospheric data network that could provide the measurements needed. Data assimilation provided a methodology for using these observations to improve the forecasting skills of the operational models. Although weather forecasts are now taken for granted, to a large extent the accuracy of these forecasts results from assimilation of meteorological observations. In the 1970s, numerical ocean general circulation models (OGCMs) became an important tool for understanding ocean circulation processes. Initial applications of these models focused on simulation of the large-scale structure of ocean currents. From these simulations, the limitations of the OGCMs were clear. Data assimilation was looked to as an approach for constraining these dynamical models with available data. For example, data assimilation could be used to quantitatively and systematically test and improve poorly known sub-grid-scale parametrizations and boundary conditions that are so abundant in OGCMs. With recent advances in data availability, it is also now feasible to use dataassimilative OGCMs for making forecasts of the ocean state, such as the El-Ninõ–La Ninã cycle in the equatorial Pacific Ocean. Implementing data-assimilative biogeochemical models has been problematic because of the paucity of adequate data. Historically, biological and chemical

364

data were obtained almost exclusively by ship surveys, and thus were extremely limited in both space and time. However, advances in satellite and mooring instrumentation, as well as in the understanding of the structure and function of marine ecosystems, now makes it feasible to begin the development of dataassimilative biogeochemical models. As a result, since the mid 1990s there has been a dramatic increase in the types of data that are input into marine ecosystem models, and the development of robust and varied approaches for assimilating these data. This research provides a framework for future studies of biogeochemical data assimilation and predictive biogeochemical modeling that will inevitably play a major role in the next generation of large interdisciplinary oceanographic observational programs. The following section provides a brief history of how the field of marine biogeochemical modeling has matured as more and more data have become available. This is followed by a description of some data assimilation methods and specific examples of how two of these methods can be used in conjunction with a simple marine ecosystem model. The final section provides a summary.

Biogeochemical Models and Data Availability Mathematical models provide a quantitative framework for investigating processes that are responsible for the biological and chemical distributions that underlie the structure and function of marine ecosystems. Mathematical models were first used to study marine ecosystems in the late 1940s and these models had their basis in the predator–prey models developed in the early 1900s. These early modeling attempts were focused on understanding the processes that allow large blooms of phytoplankton and zooplankton to occur. The models were simple in nature, including only average population characteristics, basic biological processes resulting in plant and animal growth, and interactions at the lowest trophic levels, e.g., primary and secondary producers. Effects of environmental factors such as temperature and circulation, which are important in marine systems, were not explicitly included in these models. The following generation of models included more complex biogeochemical processes, differentiation of species, and coupling of the marine biogeochemical system to circulation models. These more realistic


BIOGEOCHEMICAL DATA ASSIMILATION

models were made possible by the development of large multidisciplinary oceanographic programs in the 1970s, which, for the first time, provided concurrent physical, biological, and chemical measurements that could be input to coupled circulation– biogeochemical models. The resulting models clearly demonstrated the utility of modeling for integrating and synthesizing large multidisciplinary oceanographic datasets. However, more importantly, the realism of the simulated distributions obtained from the coupled modeling efforts helped establish this approach as an important research tool for understanding marine biogeochemical systems. Present-day multidisciplinary oceanographic programs now routinely include a mathematical modeling component. At the time coupled circulation–biogeochemical models were being developed, significant advances were being made in the measurements of biological and chemical distributions in the ocean. In the 1980s, the Coastal Zone Color Scanner satellite was launched, which provided large-scale ocean color distributions, from which phytoplankton chlorophyll distributions, and their evolution over space and time, could be derived. The ocean color data also facilitated making inferences about the relative roles of circulation versus biogeochemical processes in controlling phytoplankton distributions. The availability of large-scale observations of chlorophyll has been enhanced with the subsequent launch of the Sea-viewing Wide Field-of-view Sensor in 1997. Instrumentation capable of providing biological measurements at fine space and time scales, e.g., moored optical and acoustic measurements, now provide in situ observations that can be combined with ocean color to reveal a more complete view of chlorophyll distributions. Also, moored buoy arrays provide concurrent physical data. Thus, high-resolution datasets that can be used to study marine systems are becoming increasingly available. These new, high-quality datasets can now be used to validate coupled circulation–biogeochemical models. In many cases simulated distributions from the models reproduce many of the features seen in the ocean color observations. However, the simulated fields are often unable to reproduce the variability observed on short space (tens of kilometers) and time (days) scales. There are many potential explanations for these discrepancies, including mismatches in the space and time scales that the model resolves versus those resolved by the measurements. There may also be inconsistencies between the model structure and the observations. For instance, specific parameter values or choices for empirical formulations, forcing functions, or initial conditions may be in error.

365

Along with these new types of data comes the possibility of developing a new generation of dataassimilative biogeochemical models that will not only be better able to reproduce the observed variability in biogeochemical fields but may also have the potential for significantly improving the accuracy of model predictions. In the 1990, researchers began to investigate the use of data assimilation as an approach for improving these coupled models, and as an approach for making better use of the many types of environmental and biogeochemical data that are becoming available.

Data Assimilation Methods Many techniques exist for systematically combining data with mathematical models. The development and refinement of many of these techniques have been through the use of meteorological and oceanic general circulation models. These data assimilation methods are just starting to be tested in marine biogeochemical models; however, it is not clear whether the same methods can be used with these multidisciplinary models, since biogeochemical ocean models differ substantially from their physical counterparts. For instance, biological systems have no analogue to the Navier–Stokes equations that form the basis for fluid dynamics. Thus, biogeochemical models are by necessity largely empirical and nonlinear, and abound with poorly known formulations. For example, such models typically include large numbers of parameters that are difficult (in situ growth rates), or even impossible (mortality rates) to measure with current oceanographic instrumentation. The timescales of these models are also typically short, since the model must resolve the rate at which populations double in number, which for most of the abundant phytoplankton species is one day or less. Because of these innate differences between physical models and biogeochemical models, the application of data assimilation techniques to biogeochemical ocean models, and specifically to marine ecosystem models, presents many exciting new challenges. Although considerable effort will undoubtedly be put into developing new assimilation schemes specific to these types of models, the data-assimilative marine ecosystem models that have been developed using existing assimilation methods already show much promise. One of the most straightforward methods that has been used to combine model dynamics with data entails simply replacing the model solution with data whenever such information is available. This technique, referred to as data insertion, integrates the model forward in time until additional observations become available, at which point the model is reinitialized



Numerical Twin Experiments Before the application of data assimilation techniques to marine biogeochemical models becomes routine, a number of methodological questions need to be addressed. For instance, how many data are needed for these studies? What types of sampling strategies are optimal? What level of uncertainty in the data can be tolerated? Are these models too complex, or too simple? What types of data assimilation schemes will work best for these highly nonlinear models? One method for addressing such methodological questions is through the use of identical twin experiments. In an identical twin experiment, the model is initially run using best estimates for the model parameters in order to provide a ‘true’ simulated time-series. This time-series is subsampled to generate a synthetic data set (Figure 1A). The model is then run a second time, using an imperfect parameter set in order to generate a ‘reference’ (no assimilation) time-series. This same imperfect parameter set is used in the third and final model run, True Reference Synthetic data

Concentration

and the process is repeated. A basic assumption underlying this method is that there is adequate knowledge of the governing model dynamics and parameter values. This technique has been used to estimate velocity fields by inserting temperature and salinity data into relatively complex physical oceanographic models. In these analyses, however, model–data inconsistencies caused the resulting simulations to compare poorly with observations. This led to the development of a technique in which the model solution is ‘nudged’ toward observations whenever they become available, instead of being directly replaced by the observations. Although this more gentle method of nudging may provide a significant improvement in simulation skill, like data insertion, it still lacks a means by which information on data uncertainty can be incorporated, and does not provide an estimate of the errors in the resulting solution. More advanced assimilation schemes, such as optimal interpolation and Kalman filtering, have been successfully applied by meteorologists, yet hold little hope for marine ecosystem models because of the inherent nonlinearities of biological systems. Instead, variational schemes, which have recently been applied to nonlinear physical oceanographic systems, may be more applicable to multidisciplinary problems containing biological components. These variational methods of data assimilation, such as the adjoint method and simulated annealing, have their basis in optimization theory and rely on minimizing the differences between observed and simulated quantities, pursuant to predetermined minimization criteria. At the most basic level, these methods can be thought of as nonlinear least-squares analyses, which determine the optimal solution (including parameter values and initial and/or boundary conditions) that maximizes agreement between the model simulation and observations. The adjoint method is a variational scheme that has found considerable success in the field of physical oceanography. Although this method is now also being used in marine ecosystem modeling, the nonlinear nature of these types of models may result in the recovery of suboptimal parameter sets. Simulated annealing is another assimilation scheme that has been used with data-assimilative ecosystem models. Although this method is typically capable of recovering a single optimal parameter set, the stochastic, ‘random-walk’ nature of simulated annealing causes this technique to be considerably less efficient than the adjoint method. As a result, simulated annealing may be computationally too intensive to be of use in large-scale marine biogeochemical assimilation analyses.

(A)

Time True Reference Assimilation Error

Concentration

366

(B)

Time

Figure 1 Schematic illustrating the implementation of an identical twin experiment. The true simulation (thin solid line) represents the solution for one component of the model (e.g., phytoplankton, zooplankton, or nutrient concentration) obtained using the best estimates of the model parameters and initial conditions. A second simulation, using a different parameter set or initial conditions, provides a reference simulation (thick solid line). (A) The true simulation is subsampled to create a synthetic data set. (B) The assimilation of the synthetic data into the reference simulation results in a third model solution (dashed line). The difference between this solution and the true solution (shaded region) is a measure of the error in the data-assimilative solution.



but this time the synthetic data are assimilated into the model (Figure 1B). The success of the assimilation process is judged by the difference between these results and the true simulation, and is typically normalized by the difference between the true and reference simulations. Identical twin experiments are a necessary precursor to true data-assimilative model runs, and have the potential to provide considerable insight into a number of important issues regarding the assimilation process. For instance, they can be used to rigorously compare different assimilation schemes, to determine optimal sampling strategies, and to assess the effects of assimilating observations that are associated with known levels of noise. Furthermore, identical twin experiments can be invoked to determine whether a certain set of model parameters can be estimated independently, and thus whether or not a given model may need to be simplified. Although the utility of identical twin experiments is well accepted within the fields of meteorology and physical oceanography, this approach has only recently been applied to ecosystem modeling analyses. The two examples described below illustrate some of the strengths and weaknesses of this approach for understanding data-assimilative marine biogeochemical models.

367

The accuracy of this basic simulation can be improved by inserting phytoplankton concentrations, such as those derived from ocean color measurements. Results of an identical twin experiment demonstrate that at times when average chlorophyll concentrations are inserted into the box model (e.g., every other day in Figure 3), the error in the simulation of phytoplankton decreases to zero. Nudging yields similar results, except that the error in the phytoplankton field would be reduced to a fixed nonzero value, dependent upon the strength of the nudging. One requirement for data insertion or nudging is that data must be available on timescales coincident with those of the dominant biological processes. Because biological processes, such as phytoplankton growth, have timescales of 1–2 days in many regions of the ocean, data with this level of time resolution are required for data insertion methods to adequately represent the biological dynamics. As illustrated in Figure 3A, the improvement in simulation skill lasts only 1 to 2 days beyond the point at which phytoplankton data are no longer available for insertion. Hence, when using these methods, fully data-assimilative marine biogeochemical models can potentially create a huge demand on data resources. Another primary factor limiting the use of data insertion methods for marine biogeochemical models

Example 1: Data Insertion and Nudging

Concentration

N P Z

Time Figure 2 Schematic of the time evolution of nutrient (N), phytoplankton (P), and zooplankton (Z) concentrations, obtained from a marine ecosystem model simulation.

N P Z

No change

Error

Deterioration

Improvement (A)

T1

T0

Time N P Z

Deterioration No change

Error

The pros and cons of using data insertion or ‘nudging’ to assimilate biogeochemical data can be illustrated using a three-component marine ecosystem box model (nitrogen, phytoplankton, zooplankton). Simulations with a non-data-assimilative version of this model (Figure 2) show the behavior that is expected in this type of marine system. Nitrogen concentrations decrease over time, as nitrogen is used to support a bloom of phytoplankton. Zooplankton, the primary grazer of the phytoplankton, blooms subsequently and results in a decrease of the phytoplankton.

Improvement (B)

T0

T1

Time

Figure 3 Schematics of the change in model solution error obtained in an identical twin experiment, when data are inserted into the nutrient–phytoplankton–zooplankton model (Figure 2A) between (insert eqn) and (insert eqn). (A) Only phytoplankton (P) observations are inserted every other day, and (B) nutrient (N), phytoplankton, and zooplankton (Z) observations are inserted daily. Deterioration represents movement of the data-assimilative solution farther from the true solution; improvement represents convergence of the data-assimilative and true solutions. No change occurs when the data-assimilative model solution remains the same as in the non-data-assimilative (reference) solution.



model is a challenge at best, and in most cases, nearly impossible. For these reasons, the adjoint method, which searches parameter space to find a parameter set that minimizes model–data misfits, holds considerable potential for use in data-assimilative marine biogeochemical models. The utility of the adjoint method can be demonstrated by applying identical twin experiments to the three-component ecosystem model (Figure 2). Because marine biogeochemical models typically contain many parameters that are very highly correlated, it is often not possible to recover a parameter set in its entirety. Therefore, sensitivity or correlation analyses can be performed in order to choose a subset of relatively uncorrelated parameters that will be recovered. If an identical twin experiment is carried out in which phytoplankton data are assimilated every other day using the adjoint method, phytoplankton and zooplankton parameters may be recovered precisely (Figure 4). As a result, the errors in the phytoplankton and zooplankton simulations will be significantly reduced. (In Figure 4B the error in the phytoplankton is shown to be zero for the entire model run). However, if no nutrient data are assimilated, the parameter(s)

N1 Z1 True value P1 P2 (A)

Iterations N P Z

Deterioration No change

Error

is that these models usually consist of many components, all of which must be updated to be in balance with the assimilated data. For instance, although estimates of phytoplankton biomass are improved as a result of the assimilation of phytoplankton data, the accuracy of the other model components is reduced (Figure 3A). The error produced in the other model state variables is variable, and depends upon the level of coupling between the various components of the marine biogeochemical system. This problem may be alleviated if data exist for the other model components, since the assimilation of these additional data (Figure 3B) can substantially reduce errors in the other model components. Since the adjustment timescales for the model components differ (Figure 3B), the insertion of data may be required at varying time intervals for the different model components. Unfortunately, data sufficient to update the other ecosystem model components often do not exist. In this case ad hoc approaches, perhaps based on maintaining ratios between different ecosystem components, can be invoked; however, such approaches also have the potential to introduce errors that may negate the gains made through data assimilation. Thus, data insertion and nudging are easy to implement and in certain instances may improve the accuracy of biogeochemical model predictions. However, many issues remain to be addressed before this method can be used successfully for data-assimilative biogeochemical models. For instance, because data insertion assumes that the input data sets are perfect representations of the real world, model– data inconsistencies can be magnified and can cause model solutions to become dynamically unbalanced. This is especially a problem for simulations of systems in which the circulation is the dominant control on the biogeochemical distributions. Perhaps most importantly, however, neither data insertion nor nudging readily lends itself to improving model parametrizations or model structure.

Normalized parameter value

368

Improvement (B)

Time

Example 2: Adjoint Method

Even a relatively simple marine ecosystem model, such as the model shown in Figure 2, typically contains 10– 20 model parameters that must be specified for a given simulation. This is a crucial aspect of ecosystem modeling, since even small changes in some of these parameters may result in large differences in simulation results. Unfortunately, values for these parameters are often poorly constrained in space and time, and some, such as in situ zooplankton mortality rates, are virtually unknown. Thus, the specification of an optimal parameter set for a given biogeochemical

Figure 4 Schematics illustrating the results of using the adjoint method to assimilate phytoplankton data into the nutrient– phytoplankton–zooplankton model (Figure 2A) in an identical twin experiment. (A) The phytoplankton (P1, P2) and zooplankton (Z1) parameters converge to their true values, but the nutrient parameter (N1) cannot be recovered without the assimilation of nutrient data. (B) Model solution error is greatly reduced for both phytoplankton (P) and zooplankton (Z), but not nutrient (N) concentrations. Deterioration represents movement of the data assimilative solution farther away from the true solution; improvement represents convergence of the data assimilative and true solutions. No change occurs when the data-assimilative model solution remains the same as the non-data-assimilative (reference) solution.



on which nutrient concentration is most highly dependent, e.g., N1 (Figure 4A), will not be recovered, and therefore no significant improvement in the simulation skill of the nutrient component will result (Figure 4B). If synthetic phytoplankton, zooplankton, and nutrient data are all available for assimilation, all parameters may be recovered precisely, and the errors in all model components may be reduced to zero for the entire model run. Research on data-assimilative marine ecosystem models has shown that under certain conditions the results described above are characteristic of those obtained when real data are assimilated using the adjoint method. However, in other instances it is possible that the assimilative model may fail to recover an optimal parameter set. This can occur even if the model has been well tested and calibrated, and implies that the model is in some way inconsistent with the assimilated data set. For instance, changes in plankton community dominance might result in inconsistencies in the model and data that cannot be resolved simply through data assimilation. If this is the case, it may be possible to isolate the specific model assumption(s) that have been violated, e.g., the assumption of a constant species composition, to reformulate the model in a more realistic fashion and to repeat the assimilation analysis in order to test this hypothesis. Sometimes the adjoint method may recover multiple parameter sets, each dependent on the initial choices made for the model parameters. In these situations, rigorous approaches for choosing between the possible parameter sets are required. One approach is to establish a specific uncertainty range, either from experimental or from theoretical considerations, for each parameter that is allowed to vary in the adjoint analysis. Alternatively, the optimal parameter set could be selected on the basis of the ability of each parameter set in simulating an independent data set.

369

model and data are shown to be consistent, the specific mechanisms underlying observed patterns in simulated distributions can be identified. If a model is determined to be inconsistent with observations, it may be possible to isolate the specific model assumption that has been violated, and to reformulate the model in a more realistic fashion. Thus, although the assimilation of data into a marine biogeochemical model cannot necessarily overcome inappropriate model dynamics and structure, it can serve to guide model reformulation. During the 1990s, large interdisciplinary oceanographic programs included model prediction and forecasting as specific research objectives. However, new studies are revealing that much more work needs to be performed before this becomes a realistic and achievable goal. Until high-resolution biological and chemical data are available over large regions of the ocean, and until a much clearer understanding of the intricacies of marine ecosystems is attained, data assimilation in biogeochemical models will be more useful for model improvement and parameter estimation than for model prediction and forecasting. By providing a means for recovering the best-fit set of parameters for a given model, certain assimilation techniques may prove to be a crucial tool for marine biogeochemical modelers. The importance of inclusion of data in all steps of model development and implementation cannot be emphasized enough. It is through model and data comparisons that models are advanced and better observation systems are developed. Therefore, an important aspect of furthering the development of predictive marine biogeochemical models is recognizing the need for interdisciplinary multiscale observational and experimental networks. The availability of such data will necessitate the development of techniques for input of these data into models, and facilitate the development of data-assimilative marine biogeochemical models.

Summary Data assimilation techniques for marine biogeochemical models are just beginning to be explored. Initial results are encouraging and data assimilation approaches, such as adjoint methods, hold great promise for improving the capability of these models. For instance, recent analyses of data-assimilative biogeochemical models demonstrate that the assimilation of biogeochemical data can reduce model– data misfit by recovering optimal parameter sets using multiple types of data. Perhaps even more importantly, these data assimilation analyses can demonstrate whether or not a given model structure is consistent with a specific set of observations. When

See also Data Assimilation in Models. El Ninõ Southern Oscillation (ENSO) Models. Forward Problem in Numerical Models. Inherent Optical Properties and Irradiance. Inverse Models. Moorings. Population Dynamics Models. Primary Production Processes. Regional and Shelf Sea Models. Satellite Remote Sensing: Ocean Color.

Further Reading Fasham MJR and Evans GT (1995) The use of optimization techniques to model marine ecosystem


370


dynamics at the JGOFS station at 471N 201W. Philosophical Transactions of the Royal Society of London, Series B 348: 203--209. Ishizaka J (1993) Data assimilation for biogeochemical models. In: Evans GT and Fasham MJR (eds.) Towards a Model of Ocean Biogeochemical Processes, pp. 295--316. New York: Springer-Verlag.

Lawson LM, Spitz YH, Hofmann EE, and Long RB (1995) Data assimilation applied to a simple predator–prey model. Bulletin of Mathematical Biology 57: 593--617. Matear RJ (1995) Parameter optimization and analysis of ecosystem models using simulated annealing: a case study at Station P. Journal of Marine Research 53: 571--607.


BIOLOGICAL PUMP AND PARTICLE FLUXES S. Honjo, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

A complex ecosystem process known as the ‘biological pump’ efficiently and consistently transports large amounts of carbon molecules in the form of particulate organic carbon (POC) from the epipelagic zone to the deep interior of the World Ocean and further to the abyssal floor. The biological pump begins in the euphotic zone where primary producers sequester dissolved CO2 to produce POC and oxygen through photosynthesis [CO2 þ H2O3CH2O (POC) þ O2], that is, oceanic primary production (PP). POC thus produced is transported to the oceanic interior by the interplay of Earth’s gravity and the deep-ocean ecosystem. Most PP POC is metabolized within the oceanic ecosystem in a reversing of photosynthesis, particularly by the zooplankton and microbial communities, to remineralized CO2 that is dissolved in the oceanic interior. Thus, a huge amount of carbon is stored in the ocean, especially in the mesopelagic zone, which forms one of Earth’s major CO2 sinks. The residence time of remineralized and dissolved CO2 depends on depth; it can be only an instant at the surface and as long as centuries in the deep ocean. The small portion of POC that is not metabolized also sinks and is eventually recycled to CO2, but the deep ocean’s slow rate of ventilation allows dissolved CO2 to reside in the oceanic interior for as long as a millennium. Many phytoplankton and zooplankton secrete calcium carbonate shells in the upper ocean layers and then sink toward the seafloor. This removal of calcium carbonate decreases surface water alkalinity, thereby counteracting the effect of the POC biological pump by rendering the ocean surface less absorbent to atmospheric CO2. However, the partial dissolution of calcium carbonate in the water column and especially in bottom sediments regenerates alkalinity in deep water. This export of CO2 and alkalinity from the surface to deep water by carbonate particles is then balanced by their resupply in the reverse direction as a result of global thermohaline circulation. The biological pump’s effectiveness in removing CO2 from the atmosphere to the deep sea thus depends on the rate of POC export, the ratio of organic carbon to inorganic carbon exported (the so-called ‘rain ratio’ in settling particles), the

residence time of the dissolved metabolic CO2 in a number of ionic forms, and the rate of alkalinity regeneration in the deep sea, which depends on the turnover time of deep and intermediate waters. The spatial and temporal variability of oceanic particle flux can be measured by capturing settling particles with time-series sediment traps (TS traps). Each of these traps provides a horizontal open area (aperture) and a timed sampling mechanism that controls the exposure frequency and the length of collection periods, usually ranging from several days to a month. A sediment trap is not only used to estimate the quantity and rate of particle export but also to investigate the quality of collected particles by employing various laboratory methods including chemical and isotopic analyses and microscopy. Skeletal biomineral particles collected, including coccoliths, diatom frustules, planktonic foraminifera tests, and radiolarian skeletons, carry imprints of the habitat in which they were produced, from the ocean surface to the deep interior. This memory is useful for comparing current and past productivity and ocean and atmospheric conditions as recorded in today’s ocean inhabitants and their fossil counterparts. The majority of POC is hydrated detritus that originates in plankton cell material and microbes. Because of their light specific gravity, they will not sink unless they are aggregated and ballasted by heavier particles. The majority of settling particles are amorphous aggregates known as ‘marine snow’ and fecal pellets produced by micro- and mesozooplankton, including calanoid copepods and salps. Marine snow aggregates are composed of detritus that is agglutinated to a substrate or matrix that is often formed of fibrous matter such as abandoned larvacean nets, degraded gelatinous material, and microbial polymers. The aggregates are highly variable in size in the upper ocean layers, but become more uniform and smaller, about 0.5 mm in diameter, as they settle through the deeper layers. Small fecal pellets are often captured in the matrix as an aggregate settles. Amorphous aggregates are colonized by microbial communities that enhance their food value for zooplankton. Other independently settling particles include hard tissues of zooplankton such as adult foraminiferal tests, pteropod shells, and radiolarian skeletons, but these fast-settling particles do not serve as ballast for POC. The two major ballast materials that cause POC to sink to the oceanic interior are biogenic opal and biogenic CaCO3. The latter is one of the heaviest


371

372

BIOLOGICAL PUMP AND PARTICLE FLUXES

substances commonly produced by the oceanic ecosystem. Coccoliths, the microscopic CaCO3 crystal platelets produced as scales by phytoplankton coccolithophorids, are regarded as ideal ballasts because they are small, occur ubiquitously in the euphotic zone, and are relatively resistant to dissolution. Diatom frustules are composed of biogenic opal (poorly crystallized SiO2). A minor ballasting role is also played by lithogenic particles made up of clays and fragments of rock-forming minerals that originate from aerosols, coastal erosion, river discharge, and resuspension from the seafloor; their contribution as aggregate ballast is limited to hemipelagic areas adjacent to dry land. Aggregates are a primary resource in the food cycle of zooplankton that dwell in subphotic waters and migrate through the mesopelagic zone. These zooplankton prey on other zooplankton and on carcasses and molts, regenerating them into sinking aggregates, while POC as a whole is remineralized to CO2 by zooplankton metabolism. Ballast particles are grazed along with POC but are undigested and thus available as ballast transferring to the newly generated settling aggregates. This vigorous zooplankton activity causes erratic particle flux measurements in data from sediment traps located particularly in the upper mesopelagic zone. However, zooplankton activity diminishes with depth toward the lower mesopelagic zone, approximately 1.5 km deep. Finally, aggregate descent is driven mainly by gravitational settling (‘terminal gravitational transport’) when the diel and nonmigrating zooplankton communities virtually disappear at the boundary between the mesopelagic and bathypelagic zones. The vertical flux of oceanic particles can thus be accessed reliably in the mesopelagic/bathypelagic boundary zone, between 1.5- and 2-km depths. The study of data from TS trap experiments provides an estimate of vertical settling speed for oceanic particles of 100–200 m per day. Thanks to the ballasts and diel vertical migration of zooplankton, POC travels from its production in surface layers to the interior sink and deep ocean floor in time periods ranging from only a few to several weeks. However, because of the complex and dynamic POC cycling process described above, its settling mode changes constantly, often in a short time frame. Since the early 1980s, international programs have deployed more than 410 TS traps at more than 210 stations in the global ocean basins, in major marginal seas, and in large lakes for time periods covering all seasons. The northernmost and southernmost opensea TS moorings with published data sets were located on the Wrangel Abyssal Basin in the High Arctic (811 N, 138.51 E) and on the Ross Sea Shelf near

Antarctica (76.51 S, 1781 E). The annual fluxes of POC range from 605 mmol C m 2 yr 1, about 5 times the global average at a station in the Arabian Sea, to less than 8 mmol C m 2 yr 1 at 2 km, only 7% of the global average, in the ice-covered Wangel Abyssal Plain and the seasonally ice-covered Weddell Sea Station. The minimum POC flux from a lowlatitude open-ocean station, 25–28 mmol C m 2 yr 1, was reported from the Pacific Equatorial Warm Pool (Figure 1). All export flux values in this article are measured at or normalized to 2-km depth where the biological pump is operated under the terminal gravitational settling mode. The annual flux of particulate inorganic carbon (PIC) ranges from 459 mmol C m 2 yr 1 (4 times the global average) observed at the divergent Arabian Sea station to 8 mmol C m 2 yr 1 or less in the seasonally ice-covered Southern Ocean and the High Arctic station. The PIC flux is 8 mmol C m 2 yr 1 at a station in the Central Fram Strait. The smallest carbonate flux observed in non-ice-covered waters is 15 mmol C m 2 yr 1 in the equatorial Western Pacific Warm Pool (WARM) province (Figure 2). The arguably largest biogenic Si flux was 915 mmol Si m 2 yr 1, 8 times the world average, measured in the Antarctic Zone in 1997. Stations in the Aleutian–Bering Sea area exhibited biogenic Si flux of about 800 mmol Si m 2 yr 1. The entire Atlantic Ocean is depleted with regard to biogenic Si export flux. The smallest biogenic Si flux was observed at a North Atlantic Drift station where the flux was only 6 mmol Si m 2 yr 1, 5% of the global average (Figure 3). An estimate of the global POC export rate at the 2-km isobath using the data from 134 TS trap stations is 36 Tmol C yr 1 (1 Tmol ¼ 1012 mol) or 0.43 Gt C yr 1. The annual average POC flux per square meter in the world ocean (42-km deep ocean area) is 120 mmol C m 2 yr 1, that is, 1.4 g C m 2 yr 1. When an estimate of annual global PP, 3 Gmol C yr 1 or 36 Gt C yr 1, is applied (calculated from Behrenfeld and Falkowski), it indicates that about 1.1% of PP is exported to the ocean interior sink and that the majority of photosynthetic carbon is recycled before reaching the bottom of the mesopelagic zone. The difference between two global fluxes (DFm), the ‘export production’ flux of POC passing through the thermocline to the mesopelagic zone and the flux of POC to the mesopelagic/bathypelagic boundary, provides a rough estimate of the global rate of POC metabolized to form the mesopelagic CO2 sink that is a major product of the global biological pump. An estimate of global DFm is then 441 Tmol C yr 1, assuming the global export production is 477 Tmol C yr 1 to the pelagic ocean below 2 km (calculated



373

Organic carbon flux (Fm/b Corg) at 2 km

60° N

30° N

0°

30° S

60° S

0

20

50

100

200

500

1000

2000 20000 mmo| C m−2 yr −1

Figure 1 Global parametrization of POC flux at the 2-km isobath in mmol C m 2 yr 1 based on data sets from 134 globally distributed TS trap stations. From Honjo S, Manganini SJ, Krishfield RA, and Francois R (2008) Particulate organic carbon fluxes to the ocean interior and factors controlling the biological pump: A synthesis of global sediment trap programs since 1983. Progress in Oceanography 76: 217–285.

C in CaCO3 flux (Fm/b Cinorg) at 2 km

60° N

30° N

0°

30° S

60° S

0

20

50

100

200

500

1000

2000 20000 mmo| C m−2 yr−1

Figure 2 Global parametrization of particulate organic carbon (PIC) flux in biogenic CaCO3 at the 2-km isobath in mmol C m 2 yr 1 based on data sets from 134 globally distributed TS trap stations.

from the model by Laws et al.). This means that c. 15% of PP is consumed by the biological pump’s zooplankton ecosystem to form the mesopelagic CO2 sink.

Similarly, it is estimated that 34 Tmol C yr 1 or 0.41 Gt C yr 1 of PIC associated with CaCO3 is exported to the interior of the global ocean. The annual average POC flux per square meter for the world


374


Biogenic silica flux (Fm/b Sibio) at 2 km

60° N

30° N

0°

30° S

60° S

0

20

50

100

200

500

1000

2000 20000 mmo| C m−2 yr −1

Figure 3 Global parametrization of silicon flux in biogenic opal particles at the 2- km isobath in mmol Si m 2 yr 1 based on data sets from 134 globally distributed TS trap stations.

ocean area deeper than 2 km is 112 mmol C m 2 yr 1, that is, 11.2 g C m–2 yr 1. The global ‘rain ratio’ between organic and inorganic carbon in mole flux at 2 km is about 1.1. An estimate of global flux of silicon in the ocean interior is 34 Tmol Si yr 1. It is intriguing that at 2 km the global molar fluxes of POC and its ballast, PIC and biogenic Si, are quite similar. The annual average flux of biogenic silicon flux per square meter is 114 mmol m 2 yr 1 or 6.8 g dehydrated SiO2 m 2 yr 1. Most biogenic opal export resides in diatom frustules. The global average biogenic Si and PIC ratio (in mole) is 1.0. Current export flux data reveal distinct biogeochemical regions where the biological pump depends upon two different biological components, CaCO3-C and SiO2-Si, that significantly affect the efficiency of organic carbon export to the interior of the ocean. POC/PIC and biogenic Si/PIC o1 (in mole ratio) define the ‘carbonate ocean’, and these ratios Z1 define the ‘silica ocean’. The carbonate ocean occupies about 80% of the present world pelagic ocean between two major oceanographic fronts, the North Pacific Polar Front and the Antarctic Polar Front. The silica ocean develops on the pole sides of these two significant oceanographic fronts. The North Pacific Silica Ocean covers the area north of approximately 451 N in the Pacific, including the Bering Sea, the Sea of Okhotsk, and the northern East Sea/Japan Sea. The Antarctic Zone Silica Ocean occupies the area south of the Antarctic Circumpolar

Current. The entire Atlantic, including the Nordic Seas, and the entire Indian Ocean are carbonate ocean, except for the Antarctic Zone. In the silica ocean, CO2 is removed more efficiently because it is usually associated with higher export of organic carbon from the euphotic zone toward the interior sink with a higher rain ratio. Opal-sequestering processes do not affect alkalinity. On the other hand, carbonate particles that are exported to the deep-ocean floor generate alkalinity upon dissolution at the seafloor and thus serve to export alkalinity when deep water is recycled to the surface ocean by global thermohaline circulation, which operates on a millennial timescale.

See also Calcium Carbonates. Carbon Sequestration via Direct Injection into the Ocean. Marine Silica Cycle. Ocean Carbon System, Modeling of. Primary Production Methods. Primary Production Processes.

Further Reading Angel MV (1989) Does mesopelagic biology affect the vertical flux? In: Berger WH, Smetacek VS, and Wefer G (eds.) Productivity of oceans, past and present (Dahlem Konferenzen, 1989). Life Science Research Report 44, pp. 155–173. New York: Wiley.



Behrenfeld MJ and Falkowski PG (1997) Photosynthetic rates derived from satellite-based chlorophyll concentration. Limnology and Oceanography 42: 1--20. Curry WB, Ostermann DR, Guptha MVS, and Ittekkot V (1992) Foraminiferal production and monsoon upwelling in the Arabian Sea: Evidence from sediment traps. In: Summerhayes CP, Prell WL, and Emeis KC (eds.) Geological Society Special Publication 64: Upwelling Systems: Evolution since the Early Miocene, pp. 93--106. London: Geological Society. Fischer G, Wefer G, Romero O, Dittert N, Ratmeyer V, and Donner B (2003) Transfer of particles into the deep Atlantic and the global ocean: Control of nutrient supply and ballast production. In: Wefer G, Mulitza S, and Ratmeyer V (eds.) The South Atlantic in the Late Quaternary: Reconstruction of Material Budgets and Current Systems, pp. 12--46. Berlin: Springer. Francois R, Honjo S, Krishfield R, and Manganini S (2002) Factors controlling the flux of organic carbon to the bathypelagic zone of the ocean. Global Biogeochemical Cycles 16: 1087--1107. Honjo S (1996) Fluxes of particles to the interior of the open oceans. In: Ittekkot V, Schafer P, Honjo S, and

375

Depetris PJ (eds.) Particle Flux in the Ocean, pp. 91--154. New York: Wiley. Honjo S and Doherty KW (1988) Large aperture time-series sediment traps; design objectives, construction and application. Deep-Sea Research 35: 133--149. Honjo S, Manganini SJ, Krishfield RA, and Francois R (2008) Particulate organic carbon fluxes to the ocean interior and factors controlling the biological pump: A synthesis of global sediment trap programs since 1983. Progress in Oceanography 76: 217--285. Laws EA, Falkowski PG, Smith WO, Ducklow H, and McCarthy JJ (2000) Temperature effects on export production in the open ocean. Global Biogeochemical Cycles 14: 1231--1246. Volk T and Hoffert MI (1985) Ocean carbon pumps: Analysis of relative strength and efficiencies of in oceandriven circulation atmospheric CO2 changes. In: Sundquist ET and Broecker WS (eds.) AGU Monograph 32: The Carbon Cycle and Atmospheric CO2: Natural Variation Archean to Present, pp. 99--110. Washington, DC: American Geophysical Union.


BIOLUMINESCENCE P. J. Herring, Southampton Oceanography Centre, Southampton, UK E. A. Widder, Harbor Branch Oceanographic Institution, Fort Pierce, FL, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 308–317, & 2001, Elsevier Ltd.

Introduction Bioluminescence is the capacity of living organisms to emit visible light. In doing so they utilize a variety of chemiluminescent reaction systems. It has historically been confused with phosphorescence and the latter term is still frequently (and erroneously) used to describe marine bioluminescence. Some terrestrial species (e.g., fireflies) have the same ability, but this adaptation has been most extensively developed in the oceans. Bioluminescent species occur in only five terrestrial phyla, and only in one of these (Arthropoda, which includes the insects) are there many examples. In contrast, bioluminescence occurs in 14 marine phyla, many of which include numerous luminescent species (Table 1). All oceanic habitats, shallow and deep, pelagic and benthic, include bioluminescent species, but the phenomenon is commonest in the upper 1000 m of the pelagic environment.

Biochemistry Bioluminescence involves the oxidation of a substrate (luciferin) in the presence of an enzyme (luciferase). The distinctive feature of the reaction is that most of the energy generated is emitted as light rather than as heat. There are many different, and unrelated, kinds of luciferin, and biochemical and taxonomic criteria indicate that bioluminescence has been independently evolved many times. Marine animals are unusual, however, in that many species in at least seven phyla use the same luciferin. This compound is known as coelenterazine because it was first identified in jellyfish (coelenterates) and its molecular structure is derived from a ring of three amino acids (two tyrosines, and a phenylalanine). Nevertheless, many other marine organisms use different luciferins. In some animals (e.g., jellyfish) the luciferin/luciferase system can be extracted in the form of a stable ‘photoprotein’ that will emit light when treated with calcium.

376

Microorganisms Bioluminescent organisms are found in all of the oceans of the world and at all depths. The prevalence of the phenomenon has long been known to seafarers, as the light seen at night in the wake or bow wave of their vessels. Three kinds of single-celled marine organisms include species that produce light, namely bacteria, dinoflagellates, and radiolarians, all with different luciferins. Individual luminous bacteria do not luminesce unless there are a lot of them together – colonies therefore become bright. This is because luciferase production is switched on only by the accumulation in the environment of a critical concentration of a chemical released by the bacteria (an autoinducer). Luminous bacteria are to be found free in the ocean but are more commonly encountered as glowing colonies on either marine snow or fecal pellets, or, as luminous symbionts, in the light organs of some fish and squid (see below). There are many species of luminous dinoflagellates and they are the usual cause of sea surface luminescence, visible in the bow wave or wake of a boat or the turbulence caused by a swimmer, whether man, fish, or dolphin. They can accumulate in dense ‘blooms,’ some dense enough to be recognized as red tides, and individual dinoflagellates flash when subject to sufficient shear force (e.g., in turbulence). Because they live close to the surface, their light would be invisible by day. In fact most species have a circadian rhythm that conserves the luminescence by turning it off during the day. These organisms, and probably the radiolarians too, defend themselves against planktonic predators by their flashing, which has the added ‘burglar alarm’ benefit of alerting larger predators to the presence of the original grazer.

Plankton Other common planktonic luminous organisms are copepod and ostracod crustaceans, cnidarians (jellyfish and siphonophores) and comb jellies. Copepods are in effect the insects of the sea and are the commonest planktonic animals. Many species are luminous. Most of them do not flash but have glands on their limbs or bodies from which they squirt gobbets of luminous secretion into the water as a defensive distraction. Ostracods, though less abundant, also produce luminous droplets from groups of gland cells. Usually this is a defense, but the males of some shallow-water species of Vargula swim up off


BIOLUMINESCENCE

Table 1

Representative examples of bioluminescent marine organisms


377

378

BIOLUMINESCENCE

the bottom to signal to the females. They encode a luminous message in the combination of the frequency of their light puffs, their swimming trajectory, and the timing of their displays. The displays are equivalent to complex smoke signals, or skywriting, using light. Occasionally both copepods and ostracods may swarm in such numbers that their secretions light up the wave crests or the entire ocean surface. The luciferin of Vargula (previously named Cypridina) was the first to be identified and is a tripeptide similar to coelenterazine, but made up of three different amino acids. Certain other ostracods use coelenterazine instead. Copepods and ostracods, like bacteria, dinoflagellates, and most other marine organisms, produce blue or blue-green luminescence (Table 1). These wavelengths penetrate oceanic water best, so they are visible at the greatest range. Many cnidarians and comb jellies also produce blue light, but in a few the luminescence is a vivid green. These animals have incorporated a green fluorescent protein into

the luminous cells, or photocytes. The energy from the luciferin–luciferase reaction is transferred to the fluor and is therefore made visible as green light. Some species of jellyfish, siphonophores, and comb jellies can not only flash but also pour out a luminous secretion. The secretion may include scintillating particles, which flash independently in the water. In other species of cnidarians the light-emitting cells (photocytes) are situated all over the surface of the body and a stimulus can set off one or more waves of light that may circle over the surface for several seconds. None of these animals has image-forming eyes, so their bioluminescent displays must be aimed at other animals, probably as a defense against predators or simply to protect their very fragile tissues from accidental damage by a blundering contact. There are many luminous worms, though most of them spend their time on the sea floor. Syllid worms (fireworms) come to the surface in shallow waters for a luminous mating display, whose timing is linked to (B)

(A)

Pigment cup Photocytes Epidermis

(C)

(D)

Specular reflector

(E)

Diffuse reflector

Figure 1 The effects of pigment and reflectors on light emission from photophores: (A) point source emission of a group of photocytes or bacteria is isotropic; (B) pigment cup restricts the solid angle of emission, but absorbs some of the light; (C)–(E) reflectors of different geometries provide a more efficient emission, whether they are specular (C, D) or diffuse (E). Arrows indicate possible ray paths. (From Herring (1985) with permission.)


BIOLUMINESCENCE

379

(B)

(A)

Lens

Lamellar ring

(D)

(C) Absorption and interference filters

(E)

(F)

Light guides

Light pipe

Figure 2 Effects of accessory optical structures in photophores: (A) lens alone; (B) lens and lamellar ring (e.g., euphausiid shrimp); (C) pigment filter; (D) interference filter; (E) light guide diffuser (e.g., some squid); (F) light pipe (e.g., some anglerfishes). (From Herring (1985) with permission.)

the phase of the moon. They have a greenish light, while the pelagic worm Tomopteris is very unusual in producing yellow light (Table 1). Scale worms when attacked can shed their scales, which then flash independently. A similar tactic is used by luminous brittlestars; when grasped they shed their arm tips, leaving them to flash and writhe in the predator’s grip, like the lizard that sheds its tail. Many other echinoderms (relatives of brittlestars) are bioluminescent, including sea cucumbers, sea stars and sea lilies. Most of these live on the deep-sea floor and, like the jellies, lack image-forming eyes. Other bottom-living luminous animals include species of

sea-spiders, acorn worms, snails and clams, as well as cnidarians such as sea pens and gorgonians. In the plankton and the nekton (those animals that can swim reasonably well) are many other luminous animals, including arrow worms and Pyrosoma. The latter forms a cylindrical colony of sea-squirt-like individuals, each of which has two patches of luminous cells. The cells contain bacteria-like organelles, which are uniquely intracellular. The colonies will respond to illumination by producing a slow glow of several seconds duration, and are often seen at night from the decks of ships. Only among the crustaceans, fish, and squid are the photocytes frequently


380

BIOLUMINESCENCE

associated with accessory optical structures, including reflectors, lenses, collimators, light guides, and filters (Figure 1 and Figure 2). The result is a complex light organ or photophore. Photophores have not been developed in luminous amphipods nor in the mysid Gnathophausia, but those in euphausiid and many decapod shrimps are very elaborate structures. In these animals the photophores are located on the underside of the body and eyestalks and provide a ventral illumination. Predators from below would normally see the shrimp as a silhouette against the dim downwelling daylight but, by emitting light of the same color and intensity as the daylight, the shrimp matches the background, a tactic known as counterillumination camouflage. If the shrimp were to change its orientation in the water, tilting up or down, its luminous output would no longer match the background. All euphausiids and some decapods get over this problem by rotating the photophores in the plane of pitch so that they remain directed vertically downwards and maintain the camouflage. Many deep-sea decapod shrimps (and the mysid Gnathophausia) will squirt an intense cloud of luminescence into the water if they are startled and then disappear into the surrounding darkness. Some of the species living in the upper 1000 m have both squirted luminescence and ventral photophores. The color of light from the two sources is slightly different; the photophores necessarily match the spectral content of daylight, but the squirts are rather bluer and of broader bandwidth.

Squid and Octopods At least one squid (Heteroteuthis) also produces a squirt of luminescence. It is not luminous ink but material from a special luminous gland. This squid

can also produce a steady glow from within the gland. The complexity of photophores in different squid is quite remarkable; a single individual may have several different types on different parts of the body. Many of them are for counterillumination camouflage, being typically located beneath the eye, and sometimes under the liver, two opaque structures that need to be camouflaged. The photophores are able to match the intensity of downwelling light over a considerable range. Other squid have photophores in or on the arms and/or tentacles, sometimes with specialized photophores right at the tips. As they become mature, the females of some squid develop large photophores at the tips of certain arms, presumably as a signal for the males. Females of some pelagic octopods develop an analogous sexual photophore, in the form of a luminous ring round the mouth, as they become ripe, and lose it again when they have spawned. Deep-water octopods may have lights on the arms instead of suckers. Some shallow squids culture luminous bacteria (Photobacterium fischeri) in large paired ventral photophores. Bacteria from the female are shed into the water around the egg masses and reinfect the newly hatched larvae, which have special structures for acquiring the symbionts from the water.

Fishes The variety of photophores in squid is exceeded only by those in fishes. Several groups of fish use luminous bacterial symbionts as their source of light. Shallowwater species (e.g., ponyfish and pinecone fish) utilize bacteria (Photobacterium leiognathi and P. fischeri, respectively) that grow best at warm temperatures. Deep-sea fishes (e.g., rattails and spookfish) have a different symbiont (P. phosphoreum) that does better in colder water. All these fishes have photophores

(A)

(B)

(C)

Chromatophores

Rotation

Shutter

Open

Occluded

Figure 3 Three means whereby a photophore can be occluded: (A) chromatophores; (B) rotation; (C) shutter. (From Herring (1985) with permission.)


BIOLUMINESCENCE

that open into the gut; their symbionts are extracellular and can be grown in laboratory cultures. It is assumed that the symbionts are somehow selected from the normal gut flora. Two particular families of fishes, the shallow-water flashlight fishes and deepsea anglerfishes, have photophores that do not open to the gut, though, like all the bacterial light organs of squid and other fishes, they do open to the sea water via pores. The bacteria of these two groups of fishes are also extracellular but cannot yet be cultured. They do not belong to any known species, though they are closely related to the other symbionts. It is not known how they are reacquired in each generation. Bacteria glow continually, so these photophores have to be occluded to turn the light off (Figure 3). Most fish do not use bacteria but use their own luciferin/luciferase system. There are a few

(B)

(A)

Raised or lowered at constant speed

Rotating baffles

Fiberoptic array

PMT

Detection chamber

Pump Enclosed / pumped

Photomultiplier tube (PMT)

(F)

(E)

(D)

exceptions, which cannot make the luciferin but have to have it in their diet, like a vitamin. The bestknown is the midshipman fish Porichthys, which has numerous, complex, ventral photophores. It uses Vargula luciferin, and if deprived of dietary Vargula it does not luminesce. The luminescence returns if it is fed either whole Vargula or the pure luciferin. Populations of Porichthys that have no Vargula in their region are nonluminescent, even though they have photophores. The mysid Gnathophausia seems to have a similar dietary requirement, in this case for the luciferin coelenterazine. Other fishes probably synthesize their own luciferin. Their photophores can be extremely elaborate and a single fish may have thousands of tiny simple photophores, as well as a much smaller number of large complex ones. Most of those fishes in the upper 1500 m have counterillumination camouflage

(C)

Raised at constant speed

Static

381

Towed systems

Rotating light baffle Low-drag entrance

Stimulation grid

Pump

Figure 4 Various bioluminescence bathyphotometer designs. (A) Open field detectors designed to measure downwelling irradiance also measure bioluminescence stimulated by motion of the detector system. (B) An early sounding bathyphotometer that was raised at constant speed. Water was entrained by the upper funnel and bioluminescence was primarily triggered by turbulent flow at the exit baffle. (C) A refinement of the device in (B), equipped with entry and exit baffles that also provide excitation as water is entrained by raising or lowering. (D) Generic sketch of a low-volume enclosed and pumped bathyphotometer in which excitation is provided by pump impeller. Detector chamber volume about is 50 ml with indeterminate flow path and maximum flow rate of 1 liter s 1. This device could be used in either a moored or profiling configuration. (E) Generic towed system with excitation provided by entry baffle and flow provided either by forward motion or pump downstream from detector chamber. (F) More recent design of a high-flow-rate (up to 44 s 1), large inlet bathyphotometer (12 cm ID) with a large volume detection chamber (>11 litres) and hydrodynamically defined excitation using a grid at the inlet. (Adapted with permission from Case JF, Widder EA, Bernstein SA et al. (1993) Assessment of marine bioluminescence. Naval Research Reviews 45: 31–41.)


382

BIOLUMINESCENCE

photophores along the ventral surface of the body; the shallower species (e.g., hatchetfishes) cover the whole ventral surface with large photophores; the deeper ones (dragon fishes) have fewer, smaller, ventral photophores. In the large family of lanternfishes shallow-living and deep-living species have equivalent differences in the size and number of their ventral photophores. Many stomiiform fishes have a large postorbital photophore, behind or under each eye, very similar in position to the bacterial photophore of flashlight fishes. Both kinds of fish probably use them to illuminate prey in the surrounding water, and both can hide the white reflective surface of the photophore by rotating it or drawing a fold of black skin over its aperture. Stomiiform males usually have much larger postorbital photophores than females. Male and female lanternfishes have special sexually dimorphic photophores on the tail or head in addition to the ventral camouflage ones. Male anglerfishes have no photophores; the female’s bacterial ones can be very complex, with light pipes transmitting the light from the bacterial core to quite distant apertures. The lights are presumed to act as lures, perhaps both for prey and for males. Many stomiiform fishes also have long and complex luminous barbels, whose function is also assumed to be that of a lure, perhaps mimicking particular kinds of luminous plankton. Almost all of these animals produce blue luminescence, but there are a very few remarkable deep-sea fish that produce both blue and red light (Malacosteus, Pachystomias, Aristostomias). They have the usual complement of body photophores, including a blueemitting postorbital photophore, but they also have a suborbital red-emitting one. The red-emitting photophores contain large amounts of red fluorescent material and it is presumed that this acts as a fluor, rather like the green fluorescent protein of some jellyfish. The red light will be invisible to most other animals in the deep sea, which have only blue-sensitive visual pigment, but these fishes also have a redsensitive visual pigment. They have in effect a private wavelength, either for communication or, like a sniperscope, for illuminating prey.

Measurements of Bioluminescence Some of these organisms are the main contributors to the ‘stimulable bioluminescent potential’ of the water, i.e., the maximum amount of light that can be produced by turbulence in the water. Stimulated bioluminescence is most obvious in the wakes and bow waves of ships, but measurements of its vertical and horizontal distribution can give a quick indication of the planktonic biomass as well as an

indication of the signal a fish shoal or a submarine might produce as it travels through the waters. Oceanographic measurements of bioluminescence were first made in the 1950s when sensitive light meters, lowered into the depths to measure the penetration of sunlight, recorded flashes of luminescence. Later, when it became apparent that it was actually the movement of the light meter that was stimulating the bioluminescence, detector systems known as bathyphotometers were developed. These instruments have taken a variety of forms, with the most common design elements being a light detector viewing a light-tight chamber through which water is drawn either by movement of the bathyphotometer or by a pump (Figure 4). Light is stimulated as the bioluminescent organisms in the water experience turbulence, which is generated as the water passes through one or more constrictions or is stirred with a pump impeller. Units of measurements depend on the method of calibration and the residence time of the luminescent organism in the chamber. When residence times are short compared to the duration of the flash, the amount of light measured is a function of the detection chamber volume, so the light measured by the light detector (in photons s 1 or watts) is divided by the chamber volume and reported as photons s 1 per unit volume or watts per unit volume. On the other hand, when the residence time is long enough for an entire flash to be measured, the light measured is a function of the volumetric flow rate (volume s 1) through the chamber rather than the chamber volume and the light measured must be divided by flow and reported as photons per unit volume. Bathyphotometers come in a variety of configurations, including profiling systems, towed systems, and moored systems. The ‘stimulable bioluminescence potential’ measured with a given bathyphotometer will depend on the organisms it samples. Low-flow-rate systems with small inlets will preferentially sample slow swimmers such as dinoflagellates, while higher flow rates and larger inlets will also sample zooplankton such as copepods and ostracods. Bathyphotometer measurements of stimulated bioluminescence have been made in most of the major oceans of the world. These measurements have generally been made in the upper 100 m of the water column at night. There is considerable seasonal variability in the amount of light measured, with average values ranging from approximately 109 to 1011 photons l 1. There is also a pronounced diel rhythm of stimulable bioluminescence, with the photon flux measured in surface waters being greatly reduced or absent during the day. This is a consequence of the circadian rhythm of stimulable


BIOLUMINESCENCE

bioluminescence found in many dinoflagellates, as well as of diel vertical migration, which results in many luminescent species of plankton and nekton moving into surface waters only at night. In most cases where the organisms responsible for the stimulable bioluminescence potential have been sampled, they have been found to be primarily dinoflagellates, copepods, and ostracods. Euphausiids too may be significant sources of bioluminescence in the water column but will only be sampled by very high-flow-rate systems. Gelatinous zooplankton, such as siphonophores and ctenophores, represent another potentially significant source of bioluminescence but are often overlooked because they are destroyed by the nets and pumps that oceanographers generally depend on for sampling the water column. All these organisms represent significant secondary producers and measurement of their bioluminescence provides a rapid means of assessing their distribution patterns, in the same way that fluorescence measurements have provided valuable information on the fine-scale distribution patterns of primary producers. As with fluorescence measurements, the primary method used to determine which organisms are responsible for the light emissions has been to collect samples from regions of interest with nets or pumps. More recently there has also been some progress in developing computer image recognition programs that can identify luminescent organisms by their unique bioluminescent ‘signatures.’ Potential identifying properties of the light emissions include ntensity, kinetics, spatial pattern, and spectral distribution. Flash intensities are highly variable; while a single bacterium may emit only 104 photons s 1 a single dinoflagellate can emit more than 1011 photons s 1 at the peak of a flash (approximately 0.1 mW). Some of the brightest sources of luminescence are found among the jellies; some comb jellies, for example, have been found to emit more than 1012 photons s 1. Flash durations are also highly variable and can be tens of milliseconds (e.g., the flash from the ‘stern chaser’ light organs on the tail of a lantern fish) to many seconds (e.g., in many jellyfish). The vast majority of planktonic organisms such as dinoflagellates, copepods, and ostracods, have flash durations of between 0.1 and 1 s. The number of flashes that a single organism can produce depends on the amount of luminescent material that is stored and the manner and rate of excitation. While some organisms produce only a flash or two in response to prolonged stimulation, others may respond with tens to hundreds of flashes until their luminescent chemical stores are exhausted and/or their excitation pathways are fatigued. Full recovery

383

of luminescent capacity can occur in a matter of hours to days depending on the availability of substrates for resynthesis of the luminescent chemicals. Spatial patterns of bioluminescence vary from essentially point sources for the smaller plankton to highly identifiable outlines and/or species-specific photophore patterns for many of the nekton. As indicated earlier, most marine bioluminescence is blue; however, there are often subtle differences in spectral distributions that could aid in identifications.

Bioluminescent Phenomena Sometimes the bioluminescent plankton are responsible for dramatic surface phenomena. Luminescent wave crests have already been noted, but occasionally the sea may appear to be glowing uniformly. This ‘milky sea’ phenomenon has been described as like ‘sailing through a field of snow’ and is particularly common in the north-west Indian Ocean at the time of the south-west monsoon. It is probably the result of luminous bacteria growing on an oily surface scum. Other luminous phenomena include erupting balls of light exploding at the surface (probably fish schools coming up through dense luminous plankton and scattering at the surface) and, most dramatic of all, ‘phosphorescent wheels.’ These appear first as parallel bands of light racing across the sea surface and then change to become vast rotating wheels whose spokes may appear to extend to the horizon and which travel past the vessel at 50–100 km h 1! They occur only in less than 200 m of water and are most frequent in the Arabian Gulf. Explanations invoke stimulation of the surface bioluminescent plankton either by the ships engines or by seismic activity in the region. Neither alternative is wholly convincing.

Applications of Bioluminescence Bioluminescence plays a major role in the ecology of the ocean at all depths. Its quantification and distribution can provide oceanographers with a rapid biological marker for the proximity of physical features such as fronts and eddies, as well as an indication of the presence of particular species in the zooplankton and nekton communities. Aerial surveys with intensified videocameras have been used to find near-surface shoals of commercial fishes in several parts of the world, and in time of war (hot or cold) can monitor the night-time movements of surface vessels, torpedoes and submarines. More profitably, the use of bioluminescence has extended well beyond the oceans and into less obvious fields such as biomedical assays, pollution monitoring, and


384

BIOLUMINESCENCE

neuromuscular and developmental physiology. Bioluminescent systems extracted from marine organisms are now used widely as intracellular markers whose light emission signals a particular biochemical event or the presence of potentially damaging radicals such as active oxygen. Photoproteins extracted from jellyfish have provided much of the information on the role of intracellular calcium. The green fluorescent protein, also from jellyfish, is widely used as an intracellular marker. These systems have been cloned and manipulated genetically to extend their biomedical usefulness. The genes controlling the bioluminescence of marine bacteria have also been identified and cloned. They and the jellyfish genes can be inserted into other organisms as ‘reporter’ genes. These ‘report’ on the activation of other genes, to which they are attached, by causing light emission that can easily be monitored. Changes in the light emission of cultures of bioluminescent marine bacteria or dinoflagellates are also used to monitor a wide range of toxic pollutants. The bioluminescence that plays such an important part in the ecology of the oceans now has a plethora of other uses in the terrestrial world.

See also Cephalopods. Copepods. Crustacean Fisheries. Deep-Sea Fishes. Fish Larvae. Fish Migration,

Vertical. Gelatinous Zooplankton. Krill. Mesopelagic Fishes. Plankton Viruses. Protozoa, Planktonic Foraminifera. Protozoa, Radiolarians.

Further Reading Buskey EJ (1992) Epipelagic planktonic bioluminescence in the marginal ice zone of the Greenland Sea. Marine Biology 113: 689--698. Harvey EN (1952) Bioluminescence. New York: Academic Press. Hastings JW and Morin JG (1991) Bioluminescence. In: Prosser CL (ed.) Neural and Integrative Animal Physiology, pp. 131--170. New York: Wiley-Liss. Herring PJ (1977) Bioluminescence in marine organisms. Nature, London 267: 788--793. Herring PJ (ed.) (1978) Bioluminescence in action. London: Academic Press. Herring PJ (1985) How to survive in the dark: bioluminescence in the deep sea. In: Laverack MS (ed.) Physiological Adaptations of Marine Animals, pp. 323--350. Cambridge: The Company of Biologists. Lapota D, Geiger ML, Stiffey AV, Rosenberger DE, and Young DK (1989) Correlations of planktonic bioluminescence with other oceanographic parameters from a Norwegian fjord. Marine Ecology Progress Series 55: 217--227. Widder EA (1999) Bioluminescence. In: Archer SN (ed.) Adaptive Mechanisms in the Ecology of Vision, pp. 555--581. Leiden: Kluwer Academic Publishers.


BIO-OPTICAL MODELS A. Morel, Universite´ Pierre et Marie Curie, Villefranche-sur-mer, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 317–326, & 2001, Elsevier Ltd.

Introduction The expression ‘bio-optical state of ocean waters’ was coined, in 1978, to acknowledge the fact that in many oceanic environments, the optical properties of water bodies are essentially subordinated to the biological activity, and ultimately to phytoplankton and their derivatives. More recently the adjective bio-optical has been associated with nouns like model or algorithms. At least two meanings can be distinguished under the term ‘bio-optical model.’ A bio-optical model can designate a tool used to analyze, and then to predict, the optical properties of biological materials, such as phytoplanktonic or heterotrophic unicellular organisms, the most abundant living organisms in the ocean. Such models are based on various fundamental theories of optics which apply to a single particle, and make use of a set of rigorous equations. The optical properties which can be ‘modeled’ belong to the category of the inherent optical properties (IOP, see Radiative Transfer in the Ocean). Defined at the level of a single cell, the extension of IOPs to a collection of cells (a population) or to an assemblage of populations is straightforward from conceptual and numerical viewpoints. The computation of IOPs are carried out by using some physical characteristics of the organisms, or of the population (such as cell size, size distribution, chemical composition which governs the complex index of refraction). Bio-optical models can also refer to various ways of describing and forecasting the ‘bio-optical state’ of the ocean, namely the optical properties of a water body as a function of the biological activity within this water. Both the IOPs and the apparent optical properties (AOPs) of the water are aimed at in such approaches. In contrast to the first kind of theoretical models, these models are essentially empirical, descriptive, and actually derived from field measurements. They initially rest on observations of some regular variations in the oceanic optical properties along with its algal content in ‘Case 1 waters’ (see Table 1). The chlorophyll concentration, [Chl], is commonly used as an index to quantify the algal

content, and more generally the bio-optical state of ocean water. Once identified, and if recognized as statistically significant, such empirical relationships (between optical properties and [Chl] can be inverted, and thereafter used as predictive tools or model. It is worth remarking that regular trends generally vanish in so-called Case 2 waters (Table 1). Indeed, in these waters the optical properties are no longer influenced just by phytoplankton and related particles, as they are in Case 1 waters. They are also, and independently, determined by other substances of terrestrial origin, notably by sediments and colored dissolved (organic) matter, carried from land into coastal zones and not correlated to [Chl]. Therefore, bio-geo-optical models, that might be developed and locally useful in such areas, are not of general applicability. The two kinds of models are not disconnected. To the extent that the IOPs at the level of particles are additive, the first models, in principle, may be utilized to reconstruct the IOPs of a water body containing any assemblage of organisms and other (living or detritus) biogenic particles. Then these bulk IOPs can be combined through the radiative transfer equation (RTE) with the appropriate boundary conditions (the illumination conditions at the surface and the reflectance properties of the bottom, in particular), with a view to computing the AOPs at various depths within the water column. In this way, the result of the second category of models, the descriptive models, can be understood or interpreted. Because empirical bio-optical models generally refer to the trophic level, depicted by [Chl], they are in essence restricted to upper oceanic layers, where the photosynthetic activity takes place, where the vegetal biomass is confined, and [Chl] is measurable. In addition, a considerable effort in developing biooptical models originates from the need to interpret the satellite ocean color data in terms of chlorophyll concentration, which is only detectable in the upper oceanic layer. Possible relationships between optical properties and heterotrophic activity (bacterial abundance), or particulate organic carbon and minerogenic contents in the interior of the ocean, are not examined here. Finally, it must be added that the spectral domain encompassed by bio-optical models is that of visible (or photosynthetic) radiation, namely the 400– 700 nm domain, occasionally slightly extended toward the near infrared and near ultraviolet regions.


385

386

BIO-OPTICAL MODELS

Table 1

Concepts and quantities used in bio-optical models

Case 1/Case 2 water. Case 1 waters are those waters in which phytoplankton and their accompanying and covarying retinue of material are the principal agents responsible for the variations in optical properties of the water bodies. The accompanying material includes living heterotrophic organisms, such as bacteria or virus, various debris of biological origin, and dissolved organic matter excreted by organisms or liberated by decaying detritus. Such waters are typical of the open ocean, far from land influence. Conversely, Case 2 waters are influenced not only by unicellular algae and related particles or substances, but also by other optically significant components, from terrestrial origin, such as inorganic and organic particles in suspension, yellow substances resulting from land drainage, and sediments resuspended from bottom Quantity

Units

Symbol

Absorption coefficient Scattering coefficient Volume scattering function Back-scattering coefficient Back-scattering efficiency (the ratio bb =b) Attenuation coefficient ðc ¼ a þ bÞ Chlorophyll-specific (absorption or scattering) coefficients of phytoplankton Efficiency factors for absorption and scattering (subscripts a, b, respectively), defined as the ratios of energy absorbed within the particle, or scattered out from the particle, to the energy impinging onto its geometrical cross-section Relative size of a spherical particle, defined as a ¼ pDnw ðl0 Þ1 D diameter, nw refractive index of water, and l0 , wavelength in vacuo Relative (complex) refractive index of the particle, defined as the ratio of the index of the substance forming the particle to the refractive index of water (n, real part, n0 imaginary part) Van de Hulst parameter, defined as r ¼ 2aðn 1Þ Depth of the euphotic layer where PAR is reduced to 1% of its surface value Photosynthetic available radiation (within the 400–700 nm range) Attenuation coeffcient for downward irradiance (also K)

m1 m1 m1sr1 m1 – m1 m2 (mg Chl)1

a b b(y) bb b˜b c af , bf Qa , Qb

–

a

–

m ¼ n in 0

– m photons s1m2 m1

r Zeu PAR Kd

Optical Models for Individual Particle or Population of Particles In the open ocean Case 1 waters, phytoplankton with their accompanying retinue of living and detrital particles, are the principal agents responsible for the determination of the optical properties. The size of these particles extends from less than 0.1 mm (virus, colloids, debris), to less than 1 mm for heterotrophic bacteria and picoplanktonic algal species, and from 1 to tens or even several hundreds of micrometers for phytoplankton, protists, and large heterotrophic organisms (and actually up to tens of meters for whales). It is well known that the size distribution function of marine particles is rather monotonic, with numbers continuously increasing toward smaller sizes (Figure 1). A simple function, often sufficient to approximately describe the size distribution of oceanic particulate matter, is a power law (known as Junge distribution), dnðxÞ=dx ¼ NðxÞ ¼ kxj

½1

where x is the size (e.g., the diameter, if the particles can be considered as spherical), NðxÞ is the distribution function, i.e., the number of particles per unit

of volume, having a given size x, and within a dx interval (around x), k is a scaling factor, and j is an exponent, with typical values around 4 for oceanic particles. Such a distribution means that an increase by a factor of 10 in size corresponds to a reduction in number (in frequency of occurrence) by a factor of 10 000. As a consequence of this abruptly decreasing number of particles with size (combined with optical theories, see below), the particles which are the most optically significant are in the size range of 1–10 mm, and thus include most of the small heterotrophic organisms, phytoplanktonic cells, and various small debris. This is true for the scattering properties, but not for the back-scattering coefficient predominantly due to smaller particles (Figure 1). It is also true for absorption, even if in this process heterotrophs play a minor role (they are rather colorless); in contrast, algal cells containing a variety of pigments (chlorophylls, carotenoids, and occasionally phycobilins) are strongly absorbing bodies. The interaction between radiation and an optical object like a particle is conveniently described by dimensionless numbers, called efficiency factors for absorption and for scattering, and denoted Qa and Qb , respectively (Table 1). One advantage of these


387

BIO-OPTICAL MODELS DC?

0

14

10

V

HB

3

N (x) (m

_3

1 nP

0.5

2

Qb

6

bb

b

μP

3

2

10

_2

10

_1

0.2

0

3.0

10 Size x (μm)

Qa

0.0 10

1

2

1 MP

10

20

Qb

pP

10

10

10

15

10

_1

μm )

5

1.0

10

2

0 0

10

20

30

D (μm) Figure 1 Schematic plot (shaded stripe, left-hand side ordinate log-scale) of the approximate numerical concentration of particles according to their size in oceanic waters (eqn [1] with j ¼ 4). Approximate abundance versus mean size of major groups of microorganisms or particles are also indicated, with notations as follows: DC, debris and colloids; V, viruses; HB, heterotrophic bacteria; pP, picophytoplankton; nP nanoplankton; mP microplankton; MP, large macroplankton. (Adapted from Stramski and Kiefer (1991) Light scattering by microorganisms in the open ocean. Progress in Oceanography 28: 343–383.) Linear plot (right-hand ordinate scale, from 0 to 1) of the progressive value of the scattering coefficient, b, when the upper limit of the integral (eqn [3]) is increasing; the progressive value is relative and normalized by its final value (unity). A similar curve is drawn for the backscattering coefficient, bb . For these computations, the relative refractive index of the spherical particles is 1.05, and the wavelength is 550 nm. (Adapted from Morel A and Ahn Y-H (1991) Optics of heterotrophic nanoflagellates and ciliates: atentative assessment of their scattering role in oceanic waters compared to those of bacterial and algal cells. Journal of Marine Research 49: 177–202.)

factors lies in the fact that theories are available by which their values can be predicted as a function of the relative size (a, defined in Table 1), and the relative complex index of refraction of the particle (m, see Table 1). The Mie-Lorenz theory provide for spherical particles accurate Q values, and the angular values of the volume scattering function (see Radiative Transfer in the Ocean). When the refractive index of the particle is close to that of the surrounding medium (as is the case for most of the watery oceanic particles in suspension in water), the so-called van de Hulst approximation can apply and provide Q-factors more rapidly than via Mie computations. If the particles are made of the same substance (same refractive index, m), and are assumed to be spherical, with the same diameter, D, the absorption and scattering coefficients, a and b, of the medium

Figure 2 Efficiency factor for scattering, Qb , as a function of the parameter r (defined in Table 1), or of the diameter, D, when the relative index of the particle is set equal to 1.05, and the wavelength l is 675 nm. Curve represents Qb for a nonabsorbing particle. Curve 2 represents Qb for an absorbing particle, with n 0 ¼ 0:0075, n ¼ 1:05, and l ¼ 675nm. Curve 3 represents the efficiency factor for absorption Qa for the same n, n 0 , and l values as for curve 2. Note that when the size increases, Qb oscillates around, and tends asymptotically, toward 2, if the particle is not absorbing, or toward 1 when it is absorbing; Qa tends also toward 1.

which contains these particles, are simply expressed as a or b ¼ NpðD2 =4Þ Qa

or b

½2

where N is the number of particles per unit volume, pðD2 =4Þ represents the geometrical cross-section of a single (spherical) particle. The Q factors are simultaneously functions of D and m, through the parameter r (see Table 1). For perfectly transparent particles ðn0 ¼ 0Þ, a and Qa are obviously 0. In this case, Qb after oscillations tends asymptotically toward 2 for increasing size (Figure 2); such a particle is able to remove from the radiative field by scattering twice the amount of radiation intercepted by its geometrical cross-section (this is often called the ‘extinction paradox’). For absorbing particles, Qa increases with increasing size, and Qb , after some oscillations tends toward 1, like Qa (Figure 2). If the particles are not uniform in size, and the population is characterized by a size distribution function, NðDÞ the eqn [2] must be integrated over the appropriate size interval, according to a; or b ¼ ðp=4Þ


Z

N ðDÞD2 Qa or b ðD; mÞdðDÞ

½3

388

BIO-OPTICAL MODELS

The same parameters (size, complex refractive index) are used in the frame of the rigorous Mie theory, to compute the volume-scattering function (VSF, see Table 1) of any individual particle. The VSF for the entire particle population is simply obtained by adding the individual VSF with the appropriate weight, as derived from the distribution function NðDÞ. All the bio-optical models of the scattering properties of marine particles rest on such an approach. Their limitations do not originate from theory, but from the present lack of accurate information about the sizes and composition of the suspended material (in Case 1, and even more in Case 2 waters). If the actual pattern of the particle VSF is globally understood and can be reconstructed, the predictive skill of the models (actually the available information used as inputs) are still insufficient to allow the evolution of the back-scattering coefficient to be safely parameterized as a function of the bio-optical state (see reflectance modeling). What can be predicted, however, is the extremely low back-scattering efficiency exhibited by most of the unicellular organisms with low refractive index (such as algal cells, for instance). Also that this efficiency increases with the decreasing particle size is a theoretical evidence; as a consequence, the particles responsible for the formation of the scattering and back-scattering coefficients do not belong to the same size range (Figure 1). In summary, theoretical models are available which account very well for most of the observed optical properties. They have been validated in particular through in vitro experiments and by using various cells grown in culture (heterotrophic or photo-autotrophic organisms). Several phenomena, predicted through theories and subsequent bio-optical models, are worth mentioning.

•

•

•

Scattering by small (0.4–2 mm) organisms depends on the wavelength according to a l2 law. This spectral dependency is perfectly verified for heterotrophic bacteria (almost nonabsorbing bodies); for small phytoplanktonic cells (such as Prochlorococcus and Synechococcus), the presence of pigments results in localized features (minima) superimposed onto the general l2 spectral pattern. For larger organisms, the scattering spectrum may exhibit various shapes, including a ‘flat’ (l0) shape when the size exceeds approximately 10 mm. For algal cells, the various absorbing pigments always influence the scattering spectrum, by introducing minima and maxima in scattering throughout the absorption bands of these pigments. Absorption by pigmented cells in suspension differs from that of a ‘solution’ of the same material,

if the pigments were homogeneously distributed. This ‘packaging’ effect, and its corollary, the ‘flattening’ of the absorption spectrum, both originate from the behavior of the Qa factors with varying size and n0 . These effects are well understood, described by simple equations and accurately modeled. The above models, which address the optics of individual cells, or ultimately deal with populations, allow the properties observed in oceanic waters containing assemblages of these particles to be interpreted. In this sense, they are able to support the second category of bio-optical models, which are examined below.

Modeling the Optical Properties of Ocean Waters in Relation to Their Biological State In oceanic Case 1 waters, far from significant terrigenous influences, the origin of all materials present is necessarily to be found in the first link of the food chain, namely in photosynthesizing phytoplanktonic organisms. Heterotrophic organisms, as well as inanimate detritus or dissolved organic matter are related to algal biomass, and to the initial creation of organic matter (and particles) through photosynthesis. Therefore, the water optical properties are logically studied as a function of this vegetal biomass. Because chlorophyll a is the single ultimate photosynthetic pigment, the most abundant in all living plants, and because it is easily determined, its concentration in the water is a convenient, albeit imperfect, index of the bio-optical state. It is worth recalling that the chlorophyll concentration in oceanic waters, used as descriptor of the bio-optical state, and to which the optical properties are to be related, varies within about 3 orders of magnitude (say 0.02 and 20 mg m3, between oligotrophic zones and eutrophic conditions in upwelling areas). In practice, [Chl] and the optical properties within the upper layers must be measured simultaneously at sea to examine if statistically significant correlations can be found between some IOPs or AOPs and [Chl]. When such correlations are expressed mathematically (in general through nonlinear laws, and with a certain confidence interval), the corresponding expression can be used as a model or an algorithm. In contrast to what occurred for the previous category of bio-optical models (which are based on exact physical laws), the models examined below are in essence ‘empirical’. Varying uncertainties are therefore attached to each model; depending on inclusion


BIO-OPTICAL MODELS

of new data, the numerical formulation or the input parameters of these models are still liable to further evolution.

389

0.08

_1

The absorption coefficient of oceanic water Beside the fixed contribution of water itself ðaw Þ to this coefficient, the varying biological contribution (sometimes denoted abio ) can itself be partitioned into a component due to particulate material, ap , and another due to colored dissolved organic matter, acdom (all these coefficients are spectral quantities, even if the symbol l is omitted, when not necessary)

0.04

aφ 0.02

anp 0.00 400

450

500

550

600

650

700

λ (nm)

(A)

½4

ap, aφ

In turn, ap can be divided according to

0.2

40)

a φ (4

_1

where af and anap are the partial absorption coefficients by phytoplanktonic cells, and ‘nonalgal’ particles, respectively. This nonalgal compartment includes colored debris and all kinds of heterotrophic organisms. Techniques are available to discriminate between ap and anap and to determine their spectra; such measurements are performed with particles retained on a filter, before and after methanol extraction. The algal absorption af spectrum is indirectly obtained by difference. Typical shapes of these spectra are shown in Figure 3(A) when [Ch1] is 1 mg m3. Statistical analyses of absorption measurements performed systematically in Case 1 waters with increasing [Chl] have demonstrated that the partial coefficients (af and anap ) increase with [Chl] in a nonlinear manner (Table 2, examples in Figure 3B). As a consequence, the chlorophyll-specific absorption coefficient of phytoplankton (af , Table 1) is not a constant and decreases when [Chl] increases. Such a trend is, at least partly, due to the above-mentioned packaging effect. In addition, a regular change in the pigment composition is also at the origin of this decrease; recall that algal absorption originates from all accessory pigments, while normalization is made with respect to the sole chlorophyll a. Most of the recent studies in oceanic waters have shown that algal cells are the dominant term in forming ap. On average, af would represent about 70% of ap within the absorption blue maximum of algae (around 440 nm), and even more in the red peak (around 675 nm). The spectral af and ap patterns in Figure 3(A) are slightly changing with [Chl], as a consequence of the differences between the spectral

)

440

a p(

½5 (m )

ap ¼ af þ anap

ap

2

Inherent Optical Properties and [Chl]

a ¼ aw þ acdom þ ap

(m (mg chla) )

0.06

0.1

0)

a p (56

aφ (560) 0.0 0 (B)

5

10 _3

Chl (mg m )

Figure 3 (A) Mean spectral absorption values, as they result from statistical analyses, when the chlorophyll concentration within the water body is 1 mg m3. The curves represents ap ðlÞ, plotted inside a shaded area which represents 7 1SD, the and nonalgal particle absorption spectrum, anap ðlÞ, phytoplankton absorption spectrum, aj ðlÞ (see eqn [4]). (B) Nonlinear evolution of the mean absorption coefficients ap ðlÞ and aj ðlÞ, with increasing chlorophyll concentration (see also eqns [2.1] and [2.2] in Table 2); the two selected wavelengths correspond to the maximum (440 nm), and the minimum (560 nm) of algal absorption. (Adapted from Bricaud A, Morel A, Babin M et al. (1998) Variation of light absorption by suspended particles with chlorophyll a concentration in oceanic (case 1) waters: analysis and implications for bio-optical models. Journal of Geophysical Research 103: 31033–31044.)

values of the exponents (in eqns [2.1] and [2.2], Table 2). In comparison, nonalgal particles are less absorbing, and their absorption, regularly increasing toward the short wavelengths, is well modeled with an exponential function (Table 2, eqn [2.3]). The measurements of acdom (performed on filtered water) are rather scarce in oceanic waters; they have not yet provided a clear relationship (if any) between this term and [Chl], but have shown that the spectral shape, namely a monotonous (exponential) increase of acdom toward the short wavelengths, is rather


390

BIO-OPTICAL MODELS

Table 2 Bio-optical models for Case 1 waters, and corresponding equations by which the main inherent optical properties (IOP) of a water body can be related to its chlorophyll concentrationa ap ðlÞ ¼ Ap ðlÞ½ChlEpðlÞ

½2:1

aj ðlÞ ¼ Aj ðlÞ½ChlEjðlÞ

½2:2

Note that the terms Ap and Aj are displayed in Figure 3(A), as ap ðlÞ and aj ðlÞ are shown when [Chl] is set equal to 1 mgm3; the exponents EpðlÞ, and EjðlÞ are similar in magnitude but not equal, and they vary within the range 0.6–0.9, approximately. anap ðlÞ ¼ anap ðl0 Þexp Snap ðl l0 Þ

½2:3

acdom ðlÞ ¼ acdom ðl0 Þexp½Scdom ðl l0

½2:4

Note that in these two last expressions, the reference wavelength ðl0 Þ is arbitrary (often 440 nm is adopted), and the slopes (S) of the exponential decrease are approximately Snap ¼ 0.012 nm1 and Scdom ¼ 0.015 nm1. b ðChl;l0 Þ ¼ bw ðl0 Þ þ bp ðChl;l0 Þ

and

bp ðChl;l0 Þ ¼ Bp ðl0 Þ½Chlx

½2:5

with the exponent x in the range 0.6–0.7, approximately; with Bðl0 Þ, at l0 ¼ 550 nm, statistically found to vary between 0.15 and 0.45 m2 (mgChl)1 (see also, Figure 4). ½2:6 bp ðChl;lÞ ¼ bp ðChl;l0 Þðl=l0 Þy

with the exponent y between 0.5 and 2 (the value 1 is commonly adopted). bp ðC; l0 Þ ¼ Bp0 ðl0 Þ½Cx

0

½2:7

with the exponent x0 close also to 1 (see also Figure 4). a

Subscripts: w, water; p, particles; f, phytoplankton; nap, nonalgal particle; cdom, colored dissolved organic matter.

stable (eqn [2.4]). It is worth noting that this spectral dependency is close to that typical of anap, apart from a small difference in slope. All models presently proposed in the literature have the same exponential structure. The scattering coefficient In open ocean waters, the scattering coefficient, b ¼ bw þ bp , is the sum of the constant molecular scattering (bw), and of a varying contribution, bp, resulting from the presence of all kinds of particles, living organisms, and detritus. Unlike ap, bp cannot be split into algal and nonalgal contributions, as there is no experimental technique allowing such a discrimination. Bio-optical models aim at relating this bulk coefficient bp to [Chl]. In situ measurements of bp ðlÞ, at fixed wavelength, l0 (or indirect determination via the particle attenuation coefficient cp ðlÞ, from which ap ðlÞ is subtracted), have led, through least-squares analyses, to a nonlinear dependence upon [Chl] of the form bp ðl0 Þ ¼ Bp ðl0 Þ½Chlx

½6

Bp ðl0 Þ (which represents the value of bp when Chl ¼ 1 mg m3) has been found to vary within a factor 3 for Case 1 waters (see also Figure 4; and eqn [2.5]). The nonlinearity in eqn [6] (a power law with o1) is such that when the algal biomass increases, bp increases more slowly. This effect is generally attributed to a change (i.e., a decrease) in the relative contribution of detritus and of heterotrophic organisms to scattering when [Chl] increases (in eutrophic waters). This explanation is corroborated by the following observation: when bp is studied as a function of the particulate organic carbon concentration, [C], instead of [Chl], it varies linearly with [C] (see Table 2, eqn [2.7]). Spectral measurements of bp in the open ocean are rather scarce. If the size distribution function obeys a Junge distribution (eqn [1]), the spectral dependency of the scattering coefficient can be theoretically predicted as being a power law, with an exponent (y, in Table 2, eqn [2.6]) which is related to j (eqn [1]), simply by: y ¼ 3 j. With typical values for j around 4, y would be around 1, as roughly observed, and


BIO-OPTICAL MODELS _

[C] (mg m 3 ) 0

10

1

10

10

3

2

10

1

10

0

_1

bp (550) [m ]

10

10

_1

bp = f (Chl) 10

_2

bp = f (C) 10

_3 _3

10

_2

10

_1

10

10

0

10

1

_3

[Chl] (mg m ) Figure 4 Particle-scattering coefficient as a function of the chlorophyll concentration in Case 1 waters, and at l ¼ 550 nm (eqn [2.5] in Table 2); the natural variability in this relationship for Case 1 waters is represented by the shaded band. The dashed line represents the empirical relationship statistically obtained between bp and the organic carbon concentration (upper abscissa scale).

generally adopted in bio-optical modeling. Such a monotonic decrease of bp throughout the spectrum is probably oversimplifying, particularly at high algal concentration; indeed, phytoplankton scattering spectra are, as mentioned before, featured in reponse to the pigment absorption. Apparent Optical Properties and [Chl]

Downwelling irradiance The downwelling and upwelling irradiances (radiant flux per unit of area, see Radiative Transfer in the Ocean) are convenient measurements to make at sea to characterize the penetration of daylight into the water column. The depth variations of these quantities are quantified by diffuse attenuation coefficients (Radiative Transfer in the Ocean). When dealing with downward irradiance Ed, the corresponding coefficient is Kd (simply written K). It depends on the IOPs of the water and on the geometrical structure of the light field. Only by approximation, it can be seen as the sum of a term due to the water itself and a varying contribution of all materials (particulate and dissolved) originating from biological activity, so that KðlÞDKw ðlÞ þ Kbio ðlÞ

½7

391

The first term can be (again approximately) expressed as a function of the IOPs of optically pure seawater (aw and bw), whereas Kbio ðlÞ, resulting from the presence of all kinds of biological materials, can be related to [Chl]. On the basis of many field measurements in oceanic waters, it has been shown that the spectral Kbio ðlÞ values do not vary at random but are interrelated. The proposed optical classification is based on the realization that such a rather regular change affects simultaneously all the wavelengths and progressively modifies the entire spectrum. Bio-optical algorithms derived from statistical analyses of these field data allow the entire KðlÞ spectrum to be specified, as soon as Kðl0 Þ, at a reference wavelength l0 is known (Table 3, eqn [3.1]). A second way of analyzing the field data consists of relating the Kbio ðlÞ values to [Chl]. In Case 1 waters, the diffuse attenuation coefficients appear to be highly correlated to [Chl], and the statistical relationships (linear regression on log-transformed quantities) are expressed as Kbio ðlÞ ¼ wðlÞ½ChleðlÞ

½8

with exponents eðlÞ are always o1, whatever the wavelength. The corresponding KðlÞ bio-optical model consists of a set of such nonlinear expressions based on eqns [7] and [8] (eqn [3.2], Table 3; Figure 5). More complex models have also been used (eqn [3.4]). To the extent that K is largely determined by absorption, the nonlinear character of the correlation between K and [Chl] is not surprising and resembles that observed for a (Figure 3B). By integrating over the whole visible domain (the photosynthetic available radiation (PAR) domain, see Radiative Transfer in the Ocean), a relationship between Zeu, the depth of the euphotic zone (cf. Table 1), and [Chl] can be obtained. This nonlinear relationship can also be derived through a direct analysis of the column-integrated chlorophyll content and of Zeu, observed at sea by using a photometer able to determine the vertical PAR profile. This bio-optical algorithm is useful to predict, in Case 1 waters, the depth of the euphotic zone when the vertical chlorophyll profile has been determined (eqn [3.3], Table 3). Irradiance reflectance This apparent optical property, RðlÞ, is crucial in the interpretation of the remotely sensed ocean color. It is defined as the ratio of upwelling irradiance, Eu, to downwelling irradiance at the same depth (actually just beneath


392

BIO-OPTICAL MODELS

Table 3 Bio-optical models, and corresponding equations by which the main apparent optical properties (AOP) in Case 1 waters can be related to the chlorophyll concentration [Chl] K ðlÞ ¼ Kw ðlÞ þ M ðlÞ½K ðl0 Þ Kw ðl0

½3:1

where Kw ðlÞ is the attenuation coefficient for downwelling irradiance in pure water, MðlÞ are statistically derived coefficients and l0 a reference wavelength. ½3:2 K ðlÞ ¼ Kw ðlÞ þ wðlÞ½ChleðlÞ

wðlÞ and eðlÞ are empirical wavelength-dependent factors and exponents, derived from statistical analysis. Zeu ¼ z ½Chlz

½3:3

z is about 38 m, and the exponent z is close to 1/2; this expression is computed by using eqn [3.2], combined with a standard spectrum of solar radiation at sea level. n o K ðl½ChlÞ ¼ k ðlÞexp k 0 ðlÞlog10 ½Chl=½Chl0 2 þ 0:001½Chl2

½3:4

[Chl0] is a reference concentration (0.5 mg m3), and k ðlÞ and k 0 ðlÞ are empirical spectral parameters dervied from statistical analysis. log10 ð½ChlÞ ¼ a0 þ a1 r þ a2 r 2 þ a3 r 3

½3:5

where r ¼ log10 ½Rðl1 Þ=Rðl2 Þ is the decimal logarithm of a ratio of reflectances at two wavelengths; the cubic polynomial is an example (first order polynomials have also been proposed). The inverse relationships, which express various r ratios as a function of log10 ([Chl]), also have the same polynominal structure.

the surface in remote sensing applications) RðlÞ ¼ Eu ðlÞ=Ed ðlÞ

½9

Other expressions involving upward radiance (instead of Eu) are also in use in ocean color remote sensing; they are geometrically related to RðlÞ as defined, and physically modeled in a similar way. Historically, purely empirical models were the first to have been developed. As for Kd, quasi-simultaneous field measurements of RðlÞ and of [Chl] were carried out, and then statistically analyzed. The relationships obtained through such regression analyses were used as algorithms when processing ocean color data. Most of these algorithms consider the ratios of reflectance at two wavelengths, l1 and l2 . Statistical analyses have demonstrated that these ratios vary regularly with [Chl]. Therefore, they can be operated as predictor of [Chl], through a certain function F: ½Chl ¼ F½Rðl1 Þ=Rðl2 Þ

½10

When such ratios of reflectances (called band ratio in ocean color science) and [Chl] are plotted in log-log space, the scatterplot is sigmoid. In the so-called empirical ‘chlorophyll-algorithms’ in use in ocean color remote sensing, the functional dependency is

expressed through a polynomial with respect to the log-transformed quantities (Table 3; linear approximations have also been employed). Thanks to radiative transfer studies, it has been shown that (to a first approximation) RðlÞ can be expressed as a function of two inherent properties, a and bb, through RðlÞ ¼ f bb =ða þ bb Þ

½11

where f is not a constant factor (because R is not an inherent property), but depends in a complex manner on the structure of the light field, and on the VSF of the particles. Its value (between about 0.3 and 0.5 for common cases in oceanic waters) can only be computed by solving the RTE with the appropriate boundary conditions. As a consequence of eqn [11], a bio-optical model for reflectance reduces to a combination of models for a and bb, separately considered. Because a and bb are inherent properties, the additivity principle applies (eqns [4] and [5], for a). Therefore a purely analytical model would be built if a and bb (actually each term forming a and bb) could be completely parameterized in terms of [Chl]. Such a model is not yet available. Modeling the back-scattering coefficient is presently an unresolved problem. This coefficient is generally expressed as


BIO-OPTICAL MODELS

393

1 10

K (600) R (λ1) / R (λ2)

)

K (560 0) (44 K

0.1

λ1 = 443 nm λ2 = 555 nm

5

_

K = K w + K bio (m 1)

K (675)

λ1 = 490 nm λ2 = 555 nm

2 1 0.5

0.01 0.1

1

10 _3

0.01

[Chl] (mg m )

0.1

1

10 _3

Figure 5 Diffuse attenuation coefficient for downward irradiance, at selected wavelengths, as a function of the chlorophyll concentration in Case 1 waters. The initial values (the ordinates when [Chl] ¼ 0.02 mg m3) are almost the pure sea water values (Kw in eqn [7], see also eqn [3.1], in Table 3, for the term Kbio).

0.03

0.1

0.1 0.3

1 3

R (λ)

0.01 10

_

[Chl] (mg m 3)

0.001

300

400

500

600

700

λ (nm)

Figure 6 Example of a spectral model for reflectance as a function of [Chl] and for Case 1 waters; the various bio-optical models presently in use are not identical but very similar.

[Chl] (mg m ) Figure 7 Example of outputs of models for Case 1 waters providing the evolution with [Chl] of ratios of reflectances at two wavelengths (as indicated) (see eqn [3.2], Table 3).

value) have been proposed to express its dependency on [Chl] and on l. When dealing with absorption, the role of acdom in open ocean is largely unknown, or more precisely its magnitude has not yet been successfully related to [Chl]. For this reason, in particular, a can advantageously be replaced by its proxy, Kd, which cumulates the influence of all absorbing substances, and is easily modeled (see above). Some manipulations of the RTE, however, are necessary to derive a from Kd. Such models are called ‘semianalytical’ (see Figure 6). They are also used to predict the ratios of reflectances as a function of [Chl], and reciprocally (as through eqn [10]); they result in algorithms having the same structure as those derived from the empirical approach (eqn [3.4], Table 3, and Figure 7). Other bio-optical models, involving reflectances at more than two wavelengths, have also been proposed. Reasonable agreements have been reached between such semi-analytical approaches, empirical approaches, and independent data sets (not used when developing the models). The sigmoidal shape is well understood if not analytically reproduced with accuracy.

Conclusions being the product b b˜ b , of the scattering coefficient (which can be related to [Chl] through eqn [6]), and the dimensionless back-scattering efficiency b˜ b. The latter quantity is not presently well documented, and rather divergent hypotheses (including a constant

In terms of accuracy or predictive capabilities, the first category of bio-optical models, dealing with the properties of individual organisms, are robust particularly because they rest on firm theoretical bases. The second category, dealing with bio-optical properties of oceanic water bodies, in essence empirical,


394

BIO-OPTICAL MODELS

cannot be as accurate, especially because they also reflect the ‘natural’ variability around mean laws. This variability is expected to the extent that various substances, dissolved and particulate, living or inanimate, which are optically significant, are not strictly covarying with [Chl]. Therefore, the links between bulk optical properties and [Chl] cannot be tight. Actually, most of these properties may vary within a factor 2 or 3 (or more) around the mean value provided by the (nonlinear) models. This remark holds true for coefficients like ap, aj , bp, and K. The situation is somewhat better when ratios of optical coefficients are involved, and when covariations of these coefficients reduce the amplitude of fluctuations around the mean (as for reflectance and band ratios, for instance). In summary, the predictive skill of such models is inevitably limited by the tightness of the regressions on which they are based. It is necessary to bear in mind that further studies at sea will likely change the result of the regression analyses, and thus the numerical values of the tabulated parameters entering into the models. For this reason, only equations are given in Tables 2 and 3, the provisional numerical information to be found in the literature, is likely amenable to modification in the future). Most of the bio-optical models, by which optical properties and [Chl], are related, rely on nonlinear relationships, and often on power laws of [Chl], with exponents below 1. As a result, the optical properties vary within a narrower interval than does [Chl] in Case 1 waters, and do not span more than 2 orders of magnitude (instead of 3), which is still considerable. It can be envisaged that for Case 1 waters ‘regional’ bio-optical models, encompassing a less wide chlorophyll concentration range, and accounting for

local biological activity with its specific assemblages, will be more efficient in the future. For Case 2 waters, bio-optical models based on [Chl], are insufficient and must be supplemented by accounting for the presence of mineral and organic sediments, and dissolved colored substance, not covarying with the algal biomass.

See also Optical Particle Characterization. Radiative Transfer in the Ocean.

Further Reading Demers S (1991) Particle Analysis in Oceanography. NATO-ASI Series, vol. G 27. Berlin Heidelberg: Springer-Verlag. Gordon HR, Brown OB, Evans RH, et al. (1988) A semianalytical radiance model for ocean color. Journal of Geophysical Research 93: 10909--10924. Gordon HR and Morel A (1983) Remote sensing of ocean color for interpretation of satellite visible imagery: a review. Lecture Notes on Coastal and Estuarine Studies. New York, Berlin, Heidelberg, Tokyo: Springer-Verlag. Jerlov NG and Nielsen ES (1974) Optical Aspects of Oceanography. London, New York: Academic Press. Morel A (1988) Optical modeling of the upper ocean in relation to its biogenous matter content (Case 1 waters). Journal of Geophysical Research 93: 10749--10768. Morel A and Smith RC (1982) Terminology and units in optical oceanography. Marine Geodesy 5: 335--349. Smith RC and Baker KS (1978) The bio-optical state of ocean waters and remote sensing. Limnology and Oceanography 23: 247--259. Van de Hulst HC (1957) Light Scattering by Small Particles. New York: Wiley.


BIOTURBATION D. H. Shull, Western Washington University, Bellingham, WA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Activities of organisms inhabiting seafloor sediments (termed benthic infauna) are concealed from visual observation but their effects on sediment chemical and physical properties are nevertheless apparent. Sediment ingestion, the construction of pits, mounds, fecal pellets, and burrows, and the ventilation of subsurface burrows with overlying water significantly alter rates of chemical reactions and sediment–water exchange, destroy signals of stratigraphic tracers, bury reactive organic matter, exhume buried chemical contaminants, and change sediment physical properties such as grain size, porosity, and permeability. Biogenic sediment reworking resulting in a detectable change in sediment physical and chemical properties is termed bioturbation. It is critical to account for bioturbation when calculating chemical fluxes at the sediment– water interface and when interpreting chemical profiles in sediments. In the narrowest sense, bioturbation refers to the biogenic transport of particles that destroys stratigraphic signals. In the broader sense it can refer to biogenic transport of pore water and changes in sediment physical properties due to organism activities as well. Bioturbation and its effects on sediment chemistry and stratigraphy is a natural consequence of adaptation by organisms to living and foraging in sediments.

Particle Bioturbation Deposit feeding, the ingestion of particles comprising sedimentary deposits, is the dominant feeding strategy in muddy sediments. In fact, since the vast majority of the ocean is underlain by muddy sediments, deposit feeding is the dominant feeding strategy on the majority of the Earth’s surface. Because digestible organic matter typically comprises less than 1% of sediments by mass, to meet their metabolic needs deposit feeders exhibit rapid sediment ingestion rates, averaging roughly three body weights per day. Deposit feeders adapted to living in sediments with relatively low organic matter concentrations tend to exhibit elevated ingestion rates; there is no free lunch even for deposit feeders. Rates of deposit feeding of

individual organisms increase with increasing body size so that bioturbation rates in some sedimentary deposits may be controlled by a handful of larger species. Deposit feeders employ a wide variety of strategies to collect particles for food, but reworking modes due to deposit feeding can be broken down into the following categories: conveyor-belt feeding where particles are collected at depth and deposited at the sediment surface; subductive feeding, where particles are collected at or near the sediment surface and deposited at depth; and interior feeding where particles are collected and deposited within the sediment column. These feeding modes transport particles the length of the organism’s body or the length of its burrow. Some species of deposit feeders also ingest and egest sediment near the sediment surface, resulting in horizontal movement of particles but limited vertical displacement. Due to rapid particle ingestion rates and relatively large horizontal and vertical transport distances, deposit feeding is considered to be the primary agent of bioturbation. Benthic organisms also rework sediments through burrow formation. Muddy sediments behave more like elastic solids than granular material. A benthic burrower in muddy sediments uses its burrowing apparatus (bivalve foot, polychaete proboscis, amphipod carapace, or other burrowing tool) like a wedge to create and propagate cracks in sediments. After an organism passes through a crack, sediments tend to rebound viscoelastically resulting in relatively little net movement of particles compared to deposit feeding. An exception to this generality is burrowing by large epibenthic predators including skates, rays, and benthic-feeding marine mammals such as gray whales and walrus, which can cause extensive reworking in sediment patches where they are feeding. From a particle’s perspective, bioturbation consists of relatively short-lived intervals of particle movement due to deposit feeding or burrowing interspersed by relatively long periods during which the particles remain at rest. When particles pass through animal guts, in addition to changing location, the particles may be aggregated into fecal pellets (particles surrounded by or embedded in mucopolymers). When constructing burrows, some infauna produce mucopolymer glue to form sturdy burrow walls, locking particles into place for an extended period of time. Transport of subsurface particles to the sediment surface by conveyor-belt feeding results in downward gravitative movement of particles within the sediment column as subsurface feeding voids are


395

396

BIOTURBATION

filled with sediment from above. Within a particular sedimentary habitat many particle reworking mechanisms occur simultaneously. There are many ways to quantify mathematically the ensemble of particle motions that results in bioturbation. Traditionally, bioturbation has been modeled as though it were analogous to diffusion. This means that the collection of particle motions resembles a large number of small random displacements. Under these assumptions, bioturbation is included in the general diagenetic equation as a biodiffusion coefficient, DB. Ignoring vertical gradients in porosity and sediment compaction, the rate of change of a chemical tracer, C, in the vertical spatial dimension, x, can be represented as follows: @C @2C @C X ¼ DB 2 u þ R; @t @x @x

xoL

½1

where DB is the biodiffusion coefficient (cm2 yr 1), 1 u Pis the sediment accumulation rate (cm yr ), and R represents the sum of chemical reactions. In the absence of specific information on bioturbation mechanisms, it is often assumed that DB is constant throughout the reworked layer to the depth L. Below the depth L, DB is zero. The advantage of this formulation is that all the various particle reworking processes are quantified by one parameter, DB. The nondimensional Peclet number, Pe ¼ uLDB 1 , quantifies the relative importance of bioturbation and sediment accumulation in determining the profile of C within the reworked layer. Values of Pe less than 1 indicate a strong influence of bioturbation. Table 1 summarizes the general pattern of variation in DB, u, L, and Pe among benthic provinces at different water depths. The depth of the reworked layer, L, shows little systematic variation among habitats, averaging 10 cm. Although we would expect considerable variation in Pe, the low values in each province indicate that bioturbation is generally important throughout the ocean. An easy-to-

Table 1 Variation in the biodiffusion coefficient, DB, sedimentation rate, u, and the Peclet number, Pe, characteristic of different benthic environments. A Peclet number greater than 1 indicates sediment accumulation is more important than bioturbation DB Shallow water Cont. Shelf Slope Deep sea

10 0.1 0.05 0.01

u 100 10 1 0.5

0.1 0.01 0.001 0.0001

L 1 0.5 0.05 0.01

10 10 10 10

Pe 0.01 0.01 0.01 0.002

1 50 10 10

A Peclet number less than 1 indicates that bioturbation is more important for transport relative to sedimentation.

remember rule of thumb regarding bioturbation rates is that DB varies from c. 0.01 to 100 cm2 yr 1 from deep-sea to shallow-water depths. This variation is correlated with increased abundance of infauna, greater rates of food supply, and (with the exception of polar regions) elevated average bottom-water temperatures with decreasing water depth. Because bioturbation mechanisms can transport particles relatively large distances, roughly the length of the reworked zone, L, and because particle trajectories are often nonrandom, the biodiffusion coefficient is not appropriate for modeling the effects of bioturbation on transport of some tracers. A more general model of particle mixing that includes longer-distance, nonrandom particle trajectories is the nonlocal bioturbation model. Again neglecting variation in porosity: @C ¼ @t

ZL

Kðx0 ; x; tÞCðx0 Þdx

0

CðxÞ

ZL

@C X þ R ½2 Kðx; x0 ; tÞdx0 u @x

0

where K is the exchange function (dimensions: time 1) that quantifies the rate of particle movement from one depth, x, to any other depth, x0 . The first term on the right-hand side gives the concentration change at depth x due to the delivery of a particle tracer from other depths, x0 . The second term gives the concentration change at depth x due to transport of a tracer from depth x to other depths x0 . The other terms are defined as in eqn [1]. The exchange function, K, can potentially quantify a complex ensemble of bioturbation mechanisms. Analogs of eqn [2] that rely on discrete mathematics exist. In one dimension, nonlocal transport can be modeled as a transition matrix in which the rows of the transition matrix correspond to depths in the sediment and the matrix elements quantify the probability of transport of a tracer among depths. Multiple-dimensional automaton models can simulate complex modes of particle transport in both vertical and horizontal dimensions. These more complex models can better capture the complexities of bioturbation but sacrifice the one-parameter simplicity of eqn [1]. There are two common approaches for determining the values of the bioturbation parameters in these models. Mixing parameters can be estimated from direct measurements of deposit-feeder ingestion rates and organism burrowing rates. More often, these parameters are estimated by measuring sedimentbound tracers with known inputs to the sediment and


BIOTURBATION

P known reaction rates ( R). Mixing parameters are then calculated by fitting measured tracer profiles to profiles calculated by use of the appropriate mathematical model. Useful bioturbation tracers include excess activities of naturally occurring particlereactive radionuclides such as 234Th, 210Pb, or 7Be. These radionuclides have a relatively continuous source either from the atmosphere or from the overlying water column, are rapidly scavenged onto particles, and sink to the seafloor (see Sediment Chronologies). In addition, chlorophyll a, artificial tracers added to the sediment surface as an impulse such as glass beads or fluorescent luminophores, or other exotic identifiable material with a known rate of input are used as tracers of bioturbation. The profile of excess 210Pb in Figure 1 illustrates several effects of bioturbation on a tracer profile. The rate of bioturbation in the top 6 cm is rapid enough to completely mix excess 210Pb within this layer. The subsurface peak at 15–16 cm indicates subsurface deposition of surficial material. Below 16 cm, the slope of the profile is set by the rate of sediment accumulation and radioactive decay of 210Pb. Bioturbation has important consequences for sediment stratigraphy, chemistry, and biology. Bioturbation can homogenize tracers within the reworked layer (Figure 1). Bioturbation acts as a low-pass filter, destroying information deposited on short timescales, but preserving longer-term trends.

Bioturbation makes it generally difficult or impossible to resolve timescales of less than 103 years stratigraphically in deep-sea sediment cores. If the bioturbation mechanism is not known, it is difficult to separate changes in the input signal from changes due to mixing (Figure 2). If mixing is not complete, and the bioturbation mechanism is known, it may be possible to deconvolve the input signal to the stratigraphic record, although detailed information will be lost. If bioturbation in the surface reworked zone completely homogenizes a tracer, then knowing the mixing mechanism is unimportant. Once pancake batter is thoroughly mixed, for example, it no longer contains information on how the mixing was performed. Bioturbation has important consequences for sediment geochemistry. Bioturbation buries reactive organic matter. Subductive deposit feeders selectively feed on organic-rich particles near the sediment surface and deposit them at depth, perhaps as food caches. In the presence of horizontal transport of organic matter, or suspension-feeding benthos that locally enhance organic matter deposition through biodeposit formation, bioturbation can greatly enhance the organic matter inventory in sediments. Burial of organic matter exposes it to different oxidants, changing the degradation pathway. In particular, reworking of Mn and Fe oxides cycles them between oxidative and reducing environments,

Excess 210 Pb (dpm g−1) 0

5

10

397

Tracer concentration 15

0

0

0.2

0.4

0.6

0.8

1

0 Well-mixed surface layer

5

5

15

10 Subduction of surficial 210 Pb

20 Sediment accumulation below reworked layer

Depth (cm)

Depth (cm)

10

15

20

25 25 30 30 Figure 1 Excess 210Pb activity vs. depth in a sediment core from Narragansett Bay, Rhode Island. Data with permission from Shull DH (2001) Transition-matrix model of bioturbation and radionuclide diagenesis. Limnology and Oceanography 46(4): 905–916. Copyright (2001) by the American Society of Limnology and Oceanography, Inc.

Figure 2 Changes in the profile of a hypothetical conservative tracer present initially as two narrow subsurface peaks, as predicted from eqn [1]. DB ¼ 0.1 cm2 yr 1, u ¼ 0.1 cm yr 1, L ¼ 10 cm. Solid line: tracer profile at t ¼ 0. Dotted line: t ¼ 25 years. Dashed line: t ¼ 150 years.


398

BIOTURBATION

resulting in enhanced anaerobic degradation of organic matter. Bioturbation changes sediment properties as well. Pelletization changes the sediment grain size distribution. Furthermore, bioturbation rates are particlesize-dependent. Size-selective feeding by deposit feeders results in biogenic graded bedding with lag layers of large sediment particles either at the sediment surface or at depth, depending upon the bioturbation mechanism. Formation of pellets and burrows increases sediment porosity, counteracting the effects of sediment compaction. Sediment surface manifestations of bioturbation such as pits, mounds, and tubes alter seafloor roughness and flow characteristics of the benthic boundary layer, roughly doubling the drag compared to a hydrodynamically smooth bed.

Pore-Water Bioirrigation Most benthic infauna maintain a burrow that connects to the sediment–water interface to facilitate respiration, feeding, defecation, and other metabolic processes. These burrows exist in a range of geometries including vertical cylinders, U- or J-shaped tubes, or branching networks. In deep-sea sediments, dissolved oxygen can penetrate 30 m into the sediment. Near the shore, however, oxygen penetration is quite variable and in muddy sediments it often penetrates no farther than a few millimeters. To meet their metabolic requirements for oxygen, infauna ventilate their burrows by thrashing their bodies, flapping their appendages, by peristalsis, by beating cilia, or by oscillating like pistons in their tubes. These ventilation mechanisms result in intermittent burrow flushing, which exchanges a portion of the fluid inside the burrow with overlying water. In this way, organisms in the sediment can flush out metabolic wastes and toxins such as hydrogen sulfide that have accumulated in their burrows and they can restock the burrow water with dissolved oxygen. Observations of organisms in artificial tubes maintained in the laboratory indicate that burrow ventilation is episodic, with ventilation frequencies ranging from once per hour to 10 or more ventilation events per hour. Deposit-feeding infauna generally ventilate less frequently than suspension-feeding infauna, which pump water through their burrows for both respiration and food capture. The sediments into which these burrows are built are mixtures of particles and interconnected pore water. Surficial sediments may possess porosities (defined as the volume of interconnected pore water per unit volume of sediment) in excess of 90%. Thus,

surface sediments generally contain more pore water than particles. The rate of molecular diffusion of solutes through pore water is reduced relative to diffusion in a free solution because the solutes must follow a winding path through the particles, called sediment tortuousity. Particle bioturbation mechanisms redistribute this pore water along with the particles, but since rates of pore-water transport, inferred from dissolved tracer distributions, are typically an order of magnitude higher than rates of particle bioturbation, particle reworking is a relatively unimportant mechanism for transporting pore water. Rather, burrow ventilation seems to be the most important biogenic mechanism of pore-water transport. The consequences of burrow ventilation on pore-water transport in the surrounding sediments (termed bioirrigation) depend upon sediment permeability. Sandy sediment typically possesses high enough permeability to allow advective flow of pore water through the interconnected pore space surrounding sediment particles. Under these conditions, the pressure head within a burrow created by burrow ventilation activities can drive pore-water flow from the burrow into surrounding sediments. The velocity of this flow can be calculated using Darcy’s law: k ud ¼ ðrP rgrxÞ m

½3

where ud is the Darcy velocity, k is sediment permeability, m is the dynamic viscosity of pore water, P is pressure, r is the pore-water density, g is gravity, and r is a gradient operator (e.g., @/@x, @/@y). The velocity of pore water, u, is related to the Darcy velocity, ud ¼ uj 1, where j ¼ porosity. Substituting eqn [3] into the general diagenetic equation gives the expected change in concentration of a pore-water tracer subjected to an advection velocity driven by burrow ventilation, molecular diffusion, and chemical reactions: @C @2C @C X ¼ D0M 2 u þ R; @t @x @x

xoL

½4

where D0 M is the molecular diffusion coefficient, corrected for tortuousity. In contrast to sandy sediments, permeability of mud is generally too low to permit significant porewater advection so that u ¼ 0. Thus, pore-water transport in muddy sediments is dominated by molecular diffusion. Burrow ventilation in muddy sediments enhances pore-water transport by changing the diffusive geometry. Figure 3 shows the geometry of idealized equally spaced vertical


BIOTURBATION

(a)

(b)

399

(c) r

x

Figure 3 Idealized burrow geometry underlying eqn [5]. (a) Burrows as close-packed cylinders. (b) Rectangular plane intersecting an average burrow microenvironment. (c) The x r domain of eqn [5]. The shaded rectangle represents the burrow, while the unshaded rectangle represents the surrounding sediment. Reproduced from Aller RC (1980) Quantifying solute distributions in the bioturbated zone of marine sediments by defining an average microenvironment. Geochimica et Cosmochimica Acta 44(12): 1955–1965, with permission from Elsevier.

burrows embedded into sediment. If these burrows were rapidly flushed so that the solute concentration within the burrows were equal to the solute concentration in the overlying water, then the corresponding diagenetic equation governing pore-water transport in the vertical dimension, x, and in the radial dimension, r, would be given by ½5

The diffusion operator within the parentheses is similar to the diffusion operator in eqn [4], but quantifies molecular diffusion in both the x and r dimensions. A one-dimensional diagenetic model that incorporates the effects of bioirrigation on porewater transport can be derived from eqn [5]: X @C @2C R ¼ D0M 2 aðC C0 Þ þ @t @x 1

0 0

0.05

activity (dpm ml−1)

0.1

0.15

0.2

0.25

5

Depth (cm)

2 X @C @ C 1 @ @C 0 þ ¼ DM r þ R @t @x2 r @r @r

222 Rn

10

15 Measured 222 Rn activity 20

Supported 222 Rn activity Nonlocal model solution

½6

where a(day ) is the coefficient of nonlocal bioirrigation, and C0 is the concentration of the solute tracer in overlying water. The nonlocal bioirrigation coefficient, a, in eqn [6] treats bioirrigation as both a source and a sink for solutes at depth. The value of the bioirrigation exchange rate, a, is usually determined by measuring dissolved porewater tracers with known inputs and reaction kinetics. The most commonly used radionuclide tracer of bioirrigation is 222Rn. Produced within sediments from the decay of its parent 226Ra, 222Rn is a soluble noble gas. Pore-water exchange with overlying water results in lower 222Rn activity in sediment pore waters than would be expected, compared to the activity of its parent 226Ra. The shape of the 222Rn profile and the magnitude of the 222Rn deficit relative to 226Ra are used to determine rates of bioirrigation

25 Figure 4 Measured 222Rn activity and supported 222Rn activity (produced from the decay of 226Ra within sediment particles) vs. depth in a sediment core from Boston Harbor, Massachusetts. Horizontal error bars represent standard error from three replicate cores. The bioirrigation rate, a (day 1), was modeled as the exponential function. a ¼ 3.8e x, and the modeled profile was calculated from eqn [6]. Data from Shull DH, previously unpublished.

(Figure 4). Other tracers of pore-water exchange include inert solutes such as bromide or dissolved nutrients such as ammonium, nitrate, or silicate, if reaction kinetics can be estimated. Bioirrigation has important implications for sediment geochemistry. Bioirrigation accelerates sediment– water fluxes, changes rates of elemental cycling, catalyzes oxidation reactions in the sediment, changes vertical and horizontal gradients of pore-water solutes, elevates levels of dissolved oxygen, and reduces


400

BIOTURBATION

exposure of organisms to metabolic wastes. By changing the redox geometry of sediments, bioirrigation can significantly alter rates of redox-sensitive reactions that occur in sediments such as nitrification, denitrification, sulfate reduction, and mercury methylation.

influenced subsequent development of animal body plans during the Cambrian. Bioturbation also made a new food resource, buried organic matter, accessible to deposit feeders while radically changing the geochemistry of the seafloor.

Bioturbation and the Ecology and Evolution of Benthic Communities

See also

Bioturbation has numerous effects on benthic community structure. In muddy sediments, bioturbation by deposit feeders appears to reduce densities of suspension feeders. Conveyor-belt bioturbation can displace surface-dwelling benthos. Bioturbation changes the depth distribution of organic matter and can increase the inventory and quality of food for deposit feeders in sediments. It can increase nutrient fluxes leading to elevated rates of benthic primary production and increased microbial productivity as well. Furthermore, elevated nutrient recycling between sediments and overlying water helps to maintain water-column productivity in estuaries and other shallow-water marine environments. Marine benthic habitats of the late Neoproterozoic and early Phanerozoic (600–500 Ma) were very different from benthic habitats that existed later. Seafloors at this time were characterized by welldeveloped microbial mats, as suggested by studies of sedimentary fabric preserved in the geologic record. These extensive microbial mats and associated seafloor fauna, such as immobile suspension-feeding helicopacoid echinoderms, became scarce or extinct in the Cambrian. The substantial change that occurred in seafloor communities around this time, termed the ‘Cambrian substrate revolution’, is thought to be caused by the development of bioturbation. It is hypothesized that the emergence of both bioturbation and predation around this time led to the extinction of nonburrowing taxa and

Macrobenthos. Ocean Margin Sediments. Sediment Chronologies. Sedimentary Record, Reconstruction of Productivity from the. Uranium-Thorium Decay Series in the Oceans Overview. Uranium-Thorium Series Isotopes in Ocean Profiles.

Further Reading Aller RC (1980) Quantifying solute distributions in the bioturbated zone of marine sediments by defining an average microenvironment. Geochimica et Cosmochimica Acta 44(12): 1955--1965. Aller RC (1982) The effects of macrobenthos on chemical properties of marine sediments and overlying waters. In: McCall PL and Tevesz MJS (eds.) Animal–Sediment Relations, pp. 53--102. New York: Plenum. Boudreau BP and Jorgensen BB (2001) The Benthic Boundary Layer: Transport Processes and Biogeochemistry. Oxford, UK: Oxford University Press. Dorgan KM, Jumars PA, Johnson BD, Boudreau BP, and Landis E (2005) Burrowing by crack propagation: Efficient locomotion through muddy sediments. Nature 433: 475. Lohrer AM, Thrush SF, and Gibbs MM (2004) Bioturbators enhance ecosystem function through complex biogeochemical interactions. Nature 431: 1092--1095. Meysman F, Boudreau BP, and Middelburg JJ (2003) Relations between local, non-local, discrete and continuous models of bioturbation. Journal of Marine Research 61: 391--410. Shull DH (2001) Transition-matrix model of bioturbation and radionuclide diagenesis. Limnology and Oceanography 46(4): 905--916.


BLACK SEA CIRCULATION regular and has a length of 4445 km. The Black Sea contains only a few small islands, the largest being Zmeyiny in the west and Beresan in the northwest. The shores of the Black Sea are bounded by Bulgaria (whose length of shoreline is 300 km) and Romania (225 km) on the west, Ukraine (1540 km) and Russia (475 km) on the north, Georgia (310 km) on the east, and Turkey (1595 km) on the south (Figure 1). These data should be taken with caution as the length of coastline is a fractal measurement; if a more detailed map is used, the coastline will be longer. The seabed is divided into the shelf, the continental slope, and the deep-sea depression. The shelf occupies a large area (c. 25% of the total) in the northwestern part of the Black Sea where it is over 200 km wide, has a depth ranging from 0 to 160 m, and also receives major freshwater discharges from Europe’s large rivers. Near the Caucasian and Anatolian coasts, the shelf rarely exceeds 15 km in width and has numerous submarine canyons and channel extensions. In the geological past, the connection of the Black Sea to the oceans opened and closed a few times. Around 5–7 Ma the ancient Tethys Ocean disintegrated, giving birth to the enclosed Sarmatian Sealake. Consequently the modern Black, Caspian, and Aral Seas originated from its parts. During the Ice Age of the Pleistocene epoch, the level of the Black Sea rose, and the sea was connected to the Mediterranean and Caspian Seas several times. In the postglacial period the Black Sea contracted, became

G. I. Shapiro, University of Plymouth, Plymouth, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction The Black Sea is a unique marine environment, representing the largest land-locked basin in the world. It is situated at the southeastern edge of Europe and is connected to the remote waters of the Atlantic Ocean via the Marmara, Aegean, and Mediterranean seas through a chain of narrow straits: Bosporus, Dardanelles, and Gibraltar. The Black Sea extends for 1167 km from the most westerly point in the Burgas Bay, Bulgaria (271 270 E), to the most easterly point near the town of Kobuleti, Georgia (411 420 ) and for 624 km from the most northerly point in Beresan Liman, Ukraine (461 320 ), to Giresun, Turkey (401 550 ), in the south. The longest extents of water are 1150 and 610 km in the zonal and meridional directions, respectively. The sea has a surface area of 413 490 km2, maximum depth of 2245 m, a volume of c. 529 955 km3, and a mean depth of 1197 m. The apple-shaped Crimean Peninsula penetrates into the Black Sea from the north and separates it from its largest arm, the shallow Sea of Azov, which is linked to the Black Sea through the narrow Kerch Strait. Otherwise, the Black Sea coastline is fairly

47

Ukraine

Odesa

46

Kerch

45 Romania

Constanta N 44

3000 2500 2000 1500 1000 750 500 400 300 200 100 0 100 200 500 1000 1500 2000 2500

Rostov-na-Donu

Yalta

Russia Novorossiisk

Sevastopol

Tuapse Sochi

Varna 43 Bulgaria

Georgia

Burgas

Sinop

Batumi

42

Eregli

Samsun Giresun

41

Turkey 26

27

28

29

30

31

32

33

34

Height/depth in m

35

36

37

38

39

40

41

42

43

E Figure 1 Bathymetric map of the Black Sea based on ETOPO2 data.


401

402

BLACK SEA CIRCULATION

a brackish lake, and its water level fell below that of the ocean. Evidence suggests that c. 6000–8000 years ago the connection to the Mediterranean Basin was restored and the freshwater fauna of the Black Sea were replaced by the saltwater fauna of the Mediterranean Sea. The Black Sea has been studied since the time of Phoenician sailors. The Greeks, who colonized the shores of the sea in sixth to eighth century BC, called it Pontos Euxeinos (meaning hospitable sea), and had a fairly good estimate of its width (c. 640 km in modern units); however, they highly exaggerated its length, as gathered from writings by the famous Greek historian Herodotus (fifth century BC). The Romans colonized its shores in the third to first century BC. In the tenth century AD, the sea was often called the Russian Sea, as it was used by the Russians as a trade route ‘from Varyags to the Greeks’ connecting the Baltic Sea and Byzantium, and from the fifteenth to the eighteenth century it was a Turkish ‘lake’. A comprehensive atlas of the Black Sea was published in 1842, which was based on the 11-year-long observational campaign led by Captain E.P. Manganary and showed details of bottom topography of the coastal waters and general currents. Oceanographic measurements (temperature and water density) started in 1868 on board the Russian corvette Lvitsa (‘Lioness’). A breakthrough in the multidisciplinary study of the Black Sea was made in the 1920s: more than 1000 deep-water stations were occupied in 1924–28, measuring an extensive set of physical and chemical parameters. Incidentally, the first general scheme of basin-scale surface currents was presented by biologist N.M. Knipovich in 1932, and then improved by oceanographer G. Neumann in 1942. A database of hydrographic measurements in the Black Sea from the beginning of the century until the 1980s exists through the efforts of the erstwhile USSR and other countries. After 1990, coordinated multinational surveys were carried out with much increased coverage, and improved resolution and quality of data.

Water Budget Thermohaline properties of the Black Sea and hence the geostrophic circulation depend to a large extent upon the balance of inflowing and outflowing waters. There is no consensus about the specific figures, partly due to interannual and interdecadal variability of various components – river discharges, precipitation/evaporation, inflow of low-salinity (11 psu) water from the Sea of Azov and high-salinity (34.9 psu) Marmara Sea water with the Lower

Table 1

Freshwater balance for the Black Sea (km3 per year)

Freshwater supply by rivers Precipitation Inflow from Marmara Sea Inflow from the Sea of Azov Evaporation Outflow into Marmara Sea Outflow into the Sea of Azov

338 238 176 50 396 371 33

Adapted from Simonov AI and Altman EN (eds.) (1991) Hydrometeorology and Hydrochemistry of the Seas of the USSR: The Black Sea, issue 1, 449pp. St. Petersburg: Gidrometizdat.

Bosporus Current, outflow of low-salinity (17 psu) Black Sea water with the Upper Bosporus Current into the Marmara Sea and through the Kerch Strait into the Sea of Azov. Current estimates of average values for the water inflow/outflow for the period 1923–40 and 1945–85 based on a large number of individual studies are reported in Table 1. The Black Sea has a positive water balance, in which the inputs from freshwater sources (rivers and precipitation) exceed losses by evaporation by 180 km3 per year; this causes freshening of the upper layer, which is balanced by mixing with lower, moresaline waters, and explains why the salinity in the upper layer of the Black Sea (17–18 psu) is only half that of the World Ocean (35 psu). The Black Sea drainage area is c. 2 000 000 km2, which is nearly 5 times larger than its surface area, and covers almost a third of Europe (Figure 2). This results in a disproportionally large freshwater input. Unlike the Mediterranean, the Black Sea is an estuarine-type basin (or dilution basin). The principal source of freshwater is the Danube, which supplies on average 209 km3 of freshwater per year, although the instantaneous fluxes range from 4 103 to 9 103 m3 s 1 owing to seasonal and interannual fluctuations. Other large feeders are the Dnieper, Southern Bug, Dniester, and Rioni, and also the Don and Kuban via the Sea of Azov. The rivers flowing into the northern and western parts of the Black Sea carry much silt and form deltas, sandbars, and lagoons. An interesting two-layer current system (to and from the Black Sea) exists in the Bosporus Strait. These currents were first studied in situ and by laboratory experiments by Admiral S. Makarov in the nineteenth century. They are highly variable and dependent upon meteorological and hydrological forcing on both ends. The upper current (out of the Black sea) has a typical speed of 0.9 m s 1 accelerating in places to 2 m s 1 or more. The average speed of the lower current (below 20–40 m) exceeds 0.9 m s 1.



403

BELARUS RUSSIA

Kiev Munich Innsbruck AUSTRIA

Brno Vienna SLOVAKIA Bratislava

SLOVENIA

Dniepropetrovsk Rostov

Budapest

Graz

HUNGARY

Zagreb CROATIA

Kharkiv

UKRAINE

MOLDOVA

Cluj

Mariupol

Chisinau Odessa

ROMANIA

Crimea

Belgrade BOSNIAHERZEGOVINA

Azov sea Novorossiysk

Sevastopol Bucharest

SERBIA AND MONTENEGRO

Constanta Black Sea

Sukhumi

GEORGIA

Poti Burgas Samsun

Trabzon

Istanbul Ankara TURKEY

Figure 2 Drainage area of the Black Sea (shown in strong color). Graphic by Philippe Rekacewicz, UNIP/GRID-Arendal, http:// maps.grida.no/go/graphic/drainage_in_the_black_sea_area.

Meteorological Forcing Circulation in the Black Sea is a product of complex interaction of meteorological and thermohaline factors, influenced by detail of the coastline and bottom topography. The climate of the Black Sea is predominantly continental. Despite its modest size, the Black Sea experiences different climatic forcing in its different parts. In winter, severe cold and dry northeastern winds prevail over the northern and northwestern parts of the sea, lowering the sea temperature and bringing frequent storms. Cold and violent bora winds reach up to 30 m s–1 in the coastal region of Novorossiysk, just to the east of the Kerch Strait, making the harbor unusable for up to 70 days in a year. The eastern and southeastern parts of the sea are protected by the Caucasus mountains from the cold winds and hence retain their relatively warm temperature. This causes

basin-scale air temperature and air pressure gradients, which lead to the formation of cyclonic circulation patterns in the atmosphere above the sea. The western and southern areas are influenced by warm and humid Atlantic air, which enters the Black Sea with the Mediterranean cyclones. The monthly mean air pressure at sea level for the month of January clearly shows its cyclonic nature (Figure 3(a)). The lowest monthly average air temperature in the northern part of the Black Sea is about –2.5 1C in January (near Ochakiv) and occasionally falls as low as –30 1C. The coastal waters freeze up to a month or more, while the shallow bays, river mouths, and limans freeze up to 2–3 months. During this season the southern coast of the Crimea and the sheltered eastern and southeastern coasts enjoy a temperature of 6–8 1C, well above the freezing point. In the summer, the Azores high-pressure center extends far into Europe and influences much of the


404


(a)

(b)

1021

1012 1011 1014 1017 1020 1018 1016

1017

1015

1018 1013 1020 1021

1012 1010

1011

Figure 3 Mean atmospheric pressure at sea level in January (left panel) and July (right panel). Adapted from Leonov AK (1960) Regional Oceanography, Part 1: The Black Sea, pp. 623–765. Leningrad: Gidrometizdat.

Black Sea, causing anticyclonic air flow above the sea, and stable weather with clear skies from May to end of September; see Figure 3(b). The average July temperature is more uniform across the sea: in the north it is 22–23 1C and in the south, 24 1C. The northwestern coast has the lowest annual precipitation (300 mm), and the Caucasian coast has the highest (up to 1800 mm). The air pressure field and associated winds in July lose their cyclonic character and in many areas the vorticity of the wind becomes negative (anticyclonic); see the map of mean air pressure at sea level in Figure 3(b).

Water Mass Structure A combination of salt and heat budgets results in the unique thermohaline structure of water masses in the Black Sea. The most important feature of the Black Sea is that oxygen is dissolved and a rich sea life is made possible only in the upper water levels – the Black Sea is the world’s largest water body containing hydrogen sulfide. This is because it has a strong pycnocline, located at around 100–150-m depth, which separates the top layer (salinity 17.5–18.5 psu) influenced by fluvial inputs and the bottom layer (21–22 psu) fed by warm salty waters from the Mediterranean Sea through the Bosporus – so the sea is stratified and cannot be mixed even by strong winds or winter convection. This leads to oxygen depletion in layers below 80–150 m, which occupy up to 90% of the volume of the sea. In the anoxic deep waters, organic matter degradation uses oxygen bound in nitrates, and especially in sulfates; this generates hydrogen sulfide and leads to a gloomy, ‘dead’ zone populated only by adapted bacteria (Figure 4). Due to the low salinity, the fauna and flora of the Black Sea are qualitatively poor as compared to the

Mediterranean Sea. The Mediterranean has about 7000 plants and animal species, while the Black Sea has only about 1200. On the other hand, primary productivity based on nutrients coming from large rivers is relatively high. A cold intermediate layer (CIL; usually defined by 8 1C temperature contour) sits at the upper boundary of the main pycnocline. In the winter this layer is fed by even colder surface waters at the centers of cyclonic gyres and through cascading from the northwest shelf; in summer the surface waters get warmer and CIW becomes the coldest water in the Black Sea. An additional seasonal summer pycnocline usually develops between 10- and 40-m depth (Figure 5).

Basin-Scale Circulation Basin-scale circulation consists of a coherent, cyclonic boundary current often called the Rim Current or the Main Black Sea Current, and two sub-basin cyclonic gyres which are particularly prominent in winter and form a circulation pattern known as ‘Knipovich spectacles’; see Figure 6. These gyres are related to the shape of the coastline, which divides the sea into two sub-basins. Occasionally this basic circulation encompasses a smaller anticyclonic gyre south of the Crimean peninsula, a number of cyclonic and anticyclonic vortices and a quasipermanent anticyclonic gyre in the southeast (Batumi Eddy). Due to conservation of potential vorticity, the Rim Current tends to flow along the depth contours. Observations show that the location of the Rim Current generally coincides well with that of the continental slope: the current flows typically at a distance of 15–25 km from the shore; however, it can be as little as 3–5 km near the Georgian coast, or as great as 200 km in the northwestern part due to the



Depth (m)

0

200

180

160

140

120

100

80

60

40

20

0

200

180

160

140

120

100

80

60

40

20

9

18

7

17

T

15

21

T (C) 11 13

S (ppt) 19 20

S

Xmiss 3.04 3.08 3.12 3.16

3

21

18

S (ppt) 19 20

17

15

9

T (C) 11 13

S

7

T

Xmiss

0

22

2

4

0

4

0

17 0

3.2 0

22

17 0

2

16

6

O2

8

16

Mn (µM) 4 6 H2S (µM) 8 12

O2 (µM) 100 200

H2S

H2S (µM) 8 12

Mn (µM) 4

O2 (µM) 100 200

H2S

O2

0

20 0

10

300 0

Mn

0

20 0

8

300 0

Mn

NO3 (µM) 2 3

NO3 (µM) 2 3

NH4 (µM) 4 8

0.2

4

12

12

0.2

4

NH4

NO3

NH4 (µM) 8

NO2 (µM) 0.05 0.1 0.15

1

4

NO2 (µM) 0.05 0.1 0.15

1

NH4

NO2

16

0.25

5

16

0.25

5

NO3

0

0

PO4

0

0

20

2

20

2

PO4

Si (µM) 40 60

PO4 (µM) 4

Si

Si (µM) 40 60

PO4 (µM) 4

Si

80

6

80

6

0

100 3.1

3.2

8

0.2

3.2

8

0.4

CH4

8 7.8

0

100 3.1

8 7.8

CH4 (µM) 0.8 1.2

Alk 3.3

pH 8.2

Alk

1.6

3.4

8.4

pH

2

0

3.5 0

8.6 0

Urea

0.8 0

8.6 0

CH4 (µM) 0.4 0.6

8.4

3.5 0

Alk 3.3

pH 8.2

CH4

Urea

3.4

Alk

pH

0.2

4

0.2

5

16

6

0.8

Norg (µM) 8 12 Urea (µM) 2 4 Porg (µM) 0.4 0.6

Porg

Norg

0.8

6

Urea (µM) 2 4 Porg (µM) 0.4 0.6

20

Norg (µM) 10 15

Porg

1

20

1

25

Figure 4 Vertical distribution of temperature (T ), salinity (S), density (sy), transmission (Trans), oxygen (O2), hydrogen sulfide (H2S), total manganese (Mn2 þ ), silicates (Si), nitrates (NO3), nitrites (NO2), ammonia (NH4), urea (Urea), phosphates (PO4), and organic phosphorus (Porg) at a summer station near Gelendzhik. Concentrations of chemical parameters are in mM. Upper panel shows summer profiles measured on 2 July 2002 at 441 516N; 371 872E; lower panel shows winter profiles measured on 26 January 2004 at 441 415N; 371 317E. From Yakushev EV, Podymov OI, and Chasovnikov VK (2005) The Black Sea seasonal changes in the hydrochemical structure of the redox zone.Oceanography 18(2): 48–55.

Depth (m)


350

300

250

200

150

100

50

0

50

7.00

100

7.75

10.00

41 23 N 29 5 E

150

7.00

350

20.00

300

7.50

Distance (km)

250

8.50

7.25

200

8.75

8.00

16.00

400

450

7.00

8.00

44 35 N 33 15 E 24 22 20 18 16 14 12 10 9.5 9 8.75 8.5 8 7.75 7.5 7.25 7 6.75 6.5 6.25 6

Temperature (C)

350

300

250

200

150

100

50

0

0

50

100

41 23 N 295 E

150

21.6

250

20.4

300

Distance (km)

200

21.0

21.4

19.8

18.6

350

19.2

18.0

400

450

22.2 22 21.8 21.6 21.4 21.2 21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19 18.8 18.6 18.4 18.2 18 17.8 17.6 17.4

44 35 N 33 15 E Salinity (psu)

Figure 5 Composite temperature and salinity distributions across the western Black Sea gyre on a transect from Istanbul to Sevastopol. Based on data from 522 stations taken in the month of July over the period 1956–96 and collated by Suvorov AM, Eremeev VN, Belokopytov VN, et al. (2004) Digital Atlas: Physical Oceanography of the Black Sea (CD-ROM). Kiev: Environmental Services Data and Information Management Program, Marine Hydrophysical Institute of the National Academy of Sciences of Ukraine. Note the dome-shaped structure of the CIL, main thermocline, and the halocline due to cyclonic general circulation in the Black Sea.

Depth (m)

0

Depth (m)


30

15

407

40

45

45

Figure 6 Mean basin-scale circulation – ‘Knipovich spectacles’. Reproduced from Nauka i Zhizn, 2006, No. 2, http://www.nkj.ru/ archive/articles/4019/.

extensive shelf. The Rim Current has a width of 40–60 km and a typical speed of 0.3–0.5 m s–1, sometimes accelerating to 1 ms–1. The Rim Current is subject to meandering both on- and offshore. Specific physical processes that form the observed circulation pattern are complicated. Classically, the cyclonic wind pattern (positive curl of wind stress) has been recognized as the main forcing for the cyclonic surface circulation. On the other hand, numerical studies indicate a seasonal thermohaline circulation driven by nonuniform surface fluxes, and density distribution due to river runoff complementary to the wind-driven circulation. Thermohaline sources in the shelf regions generate density fronts and doming of density surfaces around the entire sea. This results in surface currents of comparable magnitude. The low-salinity surface waters, which are formed over the northwest shelf due to intense river discharge, travel with the Rim Current and reach the Anatolian coast in a modified form due to mixing; the travel time from the Danube River mouth to the Bosporus Strait is 1–2 months. Currents in the coastal zone and on the shelf are highly variable both in time and space but rarely exceed 0.3 m s–1. The Rim Current is clearly seen up to a depth of 300– 400 m. There were early hypotheses that at depth there is a permanent counter-current. This was not confirmed later, however; occasionally weak countercurrents were recorded above the continental slope, probably induced by mesoscale eddies. Currents in the central parts of the sea are weak, typically 0.05–0.15 m s–1, and extend to greater depths.

The western and eastern cyclonic subgyres slightly change their location from year to year and are known to move north after cold winters. The thermohaline component of the Black Sea circulation (caused by differences in density) can already be seen from an early calculation of geostrophic currents based on temperature and salinity observations of the 1920s and 1930s, as shown in Figure 7. Diurnal and semi-diurnal tides are nearly nonexistent in the Black Sea – the tidal amplitude at Constant¸a is 7 cm, and at Sevastopol only 1–3 cm. However, the water level of the Black Sea is subject to seasonal fluctuations averaging about 20 cm. The two-layer density structure of the Black Sea is reflected in the circulation pattern – the most energetic currents are concentrated in the upper 200 m, above the permanent pycnocline.

Mesoscale Dynamics In the 1980s a strange behavior of the Rim Current was recorded near the town of Gelendzhik located on the northeastern coast. The along-shore current sometimes flowed in the opposite direction for a few days, producing a bimodal distribution in the direction probability diagram at the coastal station (3 miles offshore), while 20 nautical miles offshore the prevailing direction of currents was toward northwest as usual; see Figure 8. This was later found to be caused by mesoscale (30–50 km in size) coastal anticyclonic eddies, which


?


Jstanbul

>7.5?

>7.5

30

.0

10

Kefken

5

7.

5.0 7.5

2.5

0.0

2.5

Eregli

0.0

2.5

5.0

10.0

>10.0

< 2.5

2.5

< 0.0

a r ytsc h

Sewastopol Jalta 15.0

.S

35

pe o Sin

12.5

Samsan

Befra

0.0 0.0

5

2.

Anapa Noworossijsk

Vona burnu

< 2.5 5.0

< 5.0?

>10.0

10.0

Kertsch

0.0 2.5 5.0 7.5 10.0

Feodosia M.Meganom

Kerempe Jneboll Amasra

>15.0

?

M.Tarchankutskij Jewpatoria

Perekop

Cherson

M

Midye

Sosopol

Emine Burgas

C.Sabla

Konstantza

>20.0

5.0–10.0

?

15.0–20.0

0.0–5.0

Odessa

Giresun

5.0 7.5 .0 10 12.5 75.0

< 2.5

.0

Jeros

< 5.0

40

Trapezunt

< 7.5

0 15. .0 10

Rise

> 22.5

Poti

17.5 15.0

45

Batum

> 20.0

M.Kodor

M.Pizunda Suchum

Sotschi

Tuapse

Abb. 5. Absolute Topographie des Physikalischen Meeresniveaus. Die Zehlen geben die Abweichung von einer willkürlich gewählten Niveaufläche in dyn. cm an.

40

Figure 7 Dynamic topography and geostrophic currents in the Black Sea. From Neumann G (1942), adapted from Neumann G and Pierson WJ, Jr. (1966) Principles of Physical Oceanography. New York: Prentice-Hall.

45

10.0–15.0

< 0.0

dyn. cm.

5

35

.0 15 .0 10 7.5 5.0

2.

30

20


are typically elongated, anticyclonic gyres wedged between the Rim Current and the coast. The coastal anticyclones are formed principally by fluid elements shedding from elongated recirculating wakes structures formed behind coastal headlands. They are often related to anticyclonic meanders of the Rim Current, which typically have length scales of 100– 200 km. Charts of dynamic topography (in dyn mm)

P% 5

Gelendzhik 5%

5 CT6

10

10

P% 15 10 5 20%

15

10

5 N1

Figure 8 Probability distribution of current direction at two locations in the northeastern part of the Black Sea. Coastal station shows a strong bimodal pattern. Adapted from Titov VB (1992) About the vortex role in the formation of current regime on the Black. Sea shelf and in the coastal zone ecology. Oceanology 32(1): 39–48.

409

clearly show the horizontal structure of the coastal anticyclones. In contrast to mesoscale rings generated by meandering of jet currents in the ocean (i.e., Gulf Stream rings), coastal anticyclones in the Black Sea are located to the right of the main jet, and hence they are not formed by circling of an extended anticyclonic meander on the left-hand side of the jet, but rather by the instability of the Rim Current due to the high horizontal shear and horizontal friction. The Rim Current can be identified on satellite images as it transports colored waters rich in phytoplankton, yellow substance, and suspended particles originating from the shelf areas, as in Figure 9. The upper ocean circulation pattern also reveals a series of recurrent anticyclonic eddies (sometimes individually named Sevastopol, Kali-akra, Bosporus, Sakarya, Sinop, Kizilirmak, Caucasus, and Crimea) on the periphery between the Rim Current and the undulations of the coast. Even in the presence of a strong stratification, the continental slope can affect the dynamics, structure, and stability of the near-shore current as well as the processes of generation, movement, and transformation of eddies, particularly over the shelf edge and the top part of the continental slope, where the sea depth is comparatively low. Since eddies are one of

Figure 9 Satellite image of the Black Sea based on MODIS data from NASA, 22 May 2004. Colors are exaggerated to show the contrasting coastal and open sea water masses. Adapted from http://visibleearth.nasa.gov.


410


the main mechanisms of the interaction of the nearshore zone with the deep sea, the character and intensity of the water exchange over the slope must be affected by the topography of the continental slope. In the central part of the Black Sea, larger anticyclonic eddies of 80–100 km in diameter have been observed both from in situ measurements and with satellite imagery; these are quite deep with currents penetrating the main pycnocline. Their life span ranges from 1 to 8 months and the orbital velocities can reach 0.5 m s–1 at the surface. The open-sea eddies can interact with the Rim Currrent, deflect it away from the coast, and cause intense horizontal exchanges and mixing between the shelf and deep sea. Numerical models and satellite altimetry have shown that Black Sea eddies tend to form in the eastern basin and propagate westward as Rossby waves with a speed of c. 0.03 m s–1. The narrow Black Sea section south of the Crimean peninsula strongly affects that eddy propagation. Dissipation increases in the western basin where eddies slow down and their scales become smaller. The deep-sea anticyclones originate from those coastal anticyclonic eddies which manage to escape from their near-shore locations when the Rim Current weakens, becomes intermittent, and its main stream moves further offshore. Strong mesoscale activity takes place over the wide northwest shelf where both cyclones and anticyclones are generated. Shelf-born eddies are sometimes transported by the

Rim Current counterclockwise, and disintegrate due to friction or collision with other current systems such as upwellings.

Temporal Variability Due to the temporal variability of currents it is not easy to obtain a snapshot of the circulation system. In recent years, numerical modeling has become an important tool to study Black Sea circulation. An example of model simulation of sea surface elevation and related geostrophic currents based on climatic mean external forcing is given in Figure 10. A second example is shown in Figure 11: when the models assimilate real-time observational data (mostly from satellite-borne altimeters) they are able to provide a snapshot of Black Sea circulation on a specific date. Figure 11 shows enhanced mesoscale activity – eddies, filaments, and meanders are formed in all areas of the sea. Their size range is a few tens of kilometres, consistent with the fact that the strong thermocline makes the internal Rossby radius equal to 15–20 km. A set of marginal anticyclones is clearly defined both in climate-driven and operational velocity charts (Figures 10 and 11). Model simulations have shown that, regardless of the forcing mechanism, the Rim Current is absent if the topography is not included. Changes in bottom slope and coastline orientation along the coasts generate meanders and instability features in the Rim

46° N

45° N

44° N

43° N

42° N

41° N 28° E

30° E

10

8

32° E

6

4

34° E

2

0

38 °E

36° E

2

4

6

8

40 °E

10

12

Figure 10 Snapshot of sea level elevation (cm) and surface currents from a numerical model. From Staneva JV, Dietrich DE, Stanev EV, and Bowman MJ (2001) Rim Current and coastal eddy mechanisms in an eddy-resolving Black Sea general circulation model. Journal of Marine Systems 31: 137–157.



411

Current velocity (cm s−1). Depth 2.5 m. Date 2006.08.01. Time 23(h):56(m) 100 46 80 Latitude (north)

45 60 44 40

43

20

42

41 28

30

32

34 36 Longitude (east)

38

40

Figure 11 Current velocities obtained with a numerical model, which assimilates satellite altimetry data. From Korotaev G, et al. (2006) http://dvs.net.ua.

N

46.00

44.00

42.00

28.00

30.00

32.00

34.00

36.00

38.00

40.00

E Figure 12 Tracks of 54 surface drifters in the Black Sea. Courtesy of S.V. Motyzhev, 2006.

Current on a wide range of space scales and timescales. The Rim Current is stronger, more coherent, and attached to the continental slope in the winter when the wind is strong; the circulation tends to be intermittent and dominated by mesoscale eddies in the summer when the wind is much weaker. Mesoscale eddies are often very energetic and play a much greater role in forming the circulation pattern in the Black Sea than was earlier thought; in particular, anticyclonic meanders and eddies lying between the Rim Current and the coast provide an efficient exchange mechanism between the areas of

different depth (i.e., shelf and deep sea). Mesoscale eddies and localized jets (filaments) of various origins are capable of transporting the coastal waters to the open sea over distances up to B200 km, which is only slightly smaller than the half-width of the sea itself. As mesoscale structures are formed over the entire perimeter of the sea, they tend to homogenize the chemical and biological parameters over its area. This kind of horizontal mixing cannot be achieved by a large-scale current alone. New information on basin- and mesoscale circulation was obtained from a recent international


412


44 33 N; 30 32 E

43 50 N; 32 29 E

0 12.0

12.0 10.0 8.0 7.5

10.0 7.5

7.0

8.0

50

7.5 13.0

100

7.5

12.0

8.0

11.0 10.0

47

9.0 Latitude, N

150

0

46

8.0

T (C) 7.5 7.0

0

Depth (m)

14.0

0

0

45

0

10

00 5 00 0 150

44 500

6.5 43

6.0

28

29

30 31 32 Longitude, E

33

34

200 0

50

150

100

200

Distance (km) Figure 13 Temperature section through Sevastopol Eddy, May 2004. The bar at the top shows the horizontal extent of the eddy. The eddy deepens temperature contours below 15-m depth and lifts them up in the top layer. Cascading of cold winter water from the shelf is clearly seen. Reproduced from Shapiro GI et al. (2005) http://www.research.plymouth.ac.uk/shelf/projects/Black-sea/ Black-sea.html.

0 20 40 60 80 Depth (m)

drifter experiment (1999–2003), where 54 satellitetracked drifters were deployed in the open part of the Black Sea (Figure 12). However, half of the drifters terminated on coastal shoals, driven there by mesoscale currents from the deep sea. Of the 54 drifters, only 11 executed at least one full circuit around the entire basin. Although the drifters showed the general cyclonic circulation in the Black Sea, they did not reveal the eastern and western subbasin cyclonic gyres which are represented in traditional schemes of the Black Sea circulation.

100 120 140 160 180

Coherent Structures and Overturning Circulation One interesting mesoscale feature is the Sevastopol Eddy located southwest of Crimean peninsula. Fine-scale anticyclonic eddies and vortex filaments that peel off the Crimean headlands merge into a particularly strong and coherent recirculation eddy during the period of weak stratification. Mesoscale processes near the Crimean peninsula and their

200 220 0

100

200

300

400

Oxygen (µM) Figure 14 Anticyclonic eddy facilitates downward penetration of oxygen. The open sea stations outside the eddy (red dots) show a sharp decrease of oxygen concentrations below the upper mixed layer. Greatest deepening of oxygenated waters takes place in the region where the eddy is in contact with the continental slope (blue dots). From Romanov AS et al. (2006).



36 E

37 E

38 E

39 E

413

C

N0AA–17 28 Apr 2003 8:13 GMT

45 N

Anapa

Sea surface temperature

Novorossiysk Gelendzhik

Tuapse

10

5

44 N

Sochi

0

Figure 15 Sea surface temperature map showing various types of mesoscale activity in the northeastern Black Sea. Courtesy of S.V. Stanichny, 2003.

interaction with the shelf can substantially affect the vertical mixing of chemicals and biological organisms. The distortion of the thermocline and the deepening of oxygenated waters by the Sevastopol Eddy are shown in Figures 13 and 14, respectively. The periphery of the Sevastopol Eddy entrains water originating in the basin interior and on the shelf and wraps it around its core water. The process of leakage of coastal waters from the eddy due to its breakdown contributes to long-distance horizontal mixing of water, and exchanges of dissolved matter and vorticity. Busy mesoscale activity in the northeastern part of the sea, characterized by a narrow shelf, and formation of a few types of coherent structures (jets, eddies, and filaments) are shown in Figure 15. Some of the localized jets give birth to cyclonic–anticyclonic pairs (mushroom currents); an offshore example is seen off Tuapse, the vortex pair with a near-shore component eddy is located next to Novorossiysk. Overturning circulation in the Black Sea is driven by winter convection. Cold dense water is produced more efficiently over the shelves and in the centers of sub-basin cyclonic eddies than elsewhere as the water layer involved in cooling is thinner there. This is due to limited water depth over the shelves and upward doming of the pycnocline in cyclones. Winter convection feeds the intermediate cold layer, which then

spreads over the entire sea due to horizontal mixing enhanced by mesoscale eddies and basin-scale oscillations.

See also Coastal Circulation Models. Deep Convection. Double-Diffusive Convection. Estuarine Circulation. General Circulation Models. Meddies and Sub-Surface Eddies. Mediterranean Sea Circulation. Mesoscale Eddies. Ocean Circulation. Ocean Circulation: Meridional Overturning Circulation. Open Ocean Convection. Wind Driven Circulation.

Further Reading Izdar E and Murray JW (eds.) (1991) NATO Advanced Science Institute Series, Series C, Vol. 351: Black Sea Oceanography. Dordrecht: Kluwer. Leonov AK (1960) Regional Oceanography, Part 1: The Black Sea, pp. 623–765. Leningrad: Gidrometizdat. Murray JW (2005) Special Issue: The Black Sea. Oceanography 18(2). Neumann G and Pierson WJ, Jr. (1966) Principles of Physical Oceanography. New York: Prentice-Hall.


414


Oguz T, Latun VS, Latif MA, et al. (1993) Circulation in the surface and intermediate layers of the Black Sea. Deep Sea Research 40: 1597--1612. ¨ zsoy E and U ¨ nluäta U¨ (1997) Oceanography of the Black O Sea: A review of some recent results. Earth Science Review 42: 231--272. Simonov AI and Altman EN (eds.) (1991) Hydrometeorology and Hydrochemistry of the Seas of the USSR: The Black Sea, issue 1, 449pp. St. Petersburg: Gidrometizdat. Sorokin Yu I (2002) The Black Sea Ecology and Oceanography, 875pp. Leiden: Backhuys. Staneva JV, Dietrich DE, Stanev EV, and Bowman MJ (2001) Rim Current and coastal eddy mechanisms in an eddy-resolving Black Sea general circulation model. Journal of Marine Systems 31: 137--157. Suvorov AM, Eremeev VN, Belokopytov VN, et al. (2004) Digital Atlas: Physical Oceanography of the Black Sea (CD-ROM). Kiev: Environmental Services Data and Information Management Program, Marine Hydrophysical Institute of the National Academy of Sciences of Ukraine. Titov VB (1992) About the vortex role in the formation of current regime on the Black. Sea shelf and in the coastal zone ecology. Oceanology 32(1): 39--48.

UNEP/GRID-Arendal (2001) Drainage in the Black Sea Area. In: UNEP/GRID-Arendal Maps and Graphics Library. http://maps.grida.no/go/graphic/drainage_in_ the_black_sea_area (accessed Feb. 2008). Yakushev EV, Podymov OI, and Chasovnikov VK (2005) The Black Sea seasonal changes in the hydrochemical structure of the redox zone. Oceanography 18(2): 48--55. Zatsepin AG and Flint MV (eds.) (2002) Multidisciplinary Investigations of the Northeast Part of the Black Sea, 475pp. Moscow: Nauka.

Relevant Websites http://dvs.net.ua – Marine Hydrophysical Institute National Academy of Sciences of Ukraine, Remote Sensing Department. http://research.plymouth.ac.uk – Shelf Sea Oceanography and Meteorology research group. http://blacksea.orlyonok.ru – The Living Black Sea Marine Environmental Education Program. http://visibleearth.nasa.gov – Visible Earth.


BOTTOM WATER FORMATION A. L. Gordon, Columbia University, Palisades, NY, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 334–340, & 2001, Elsevier Ltd.

Introduction Meridional sections of temperature and salinity through the Pacific and Atlantic Oceans (Figure 1) reveal that in the Pacific below 2000 m, more than half of the ocean depth, the water is colder than 21C. The Atlantic is somewhat warmer, but there too the lower 1000 m of the ocean is well below 21C. Only

within the surface layer, generally less than the upper 500 m of the ocean is the water warmer than 101C, amounting to only 10% of the total ocean volume. The coldness of the deep ocean is due to interaction of the ocean with the polar atmosphere. There, surface water reaches the freezing point of sea water. Streams of very cold water can be traced spreading primarily from the Antarctic along the sea floor, warming en route by mixing with overlying water, into the world’s oceans (Figure 2). The coldest bottom water, Antarctic Bottom Water (AABW), is derived from the shores of Antarctica. There, freezing point, high oxygen concentration, water is produced during the winter over the continental shelf. At a few sites the shelf water salinity is

Figure 1 Meridional sections of temperature and salinity through the Pacific and Atlantic Oceans. Antarctica is at the center of the figure.


415

416

BOTTOM WATER FORMATION 90° W

0°

90° E

180°

90° W

60°N

60° N

30° N

30° N

4 2

0°

0°

0 30° S

30° S

60° S

⊗ ●

90° W

⊗ ●

⊗ ● 0°

90° E

_2

60° S

⊗ ●

180°

90° W

Figure 2 Bottom potential temperature along the seafloor, for oceanic areas deeper than 4000 m. The four symbols along the coast of Antarctica mark the places where Antarctic Bottom water forms.

sufficiently high, greater than 34.61%, that, on cooling to the freezing point, the surface water density is sufficiently high to allow it to sink to great depths of the ocean. As the shelf water descends over the continental slope into the deep ocean it mixes with adjacent deep water, but this water is also quite cold so the final product arriving at the seafloor at the foot of Antarctica is about 1.01C. Definitions used by different authors vary, but generally AABW is defined as having a potential temperature (the temperature corrected for adiabatic heating due to hydrostatic pressure) less than 01C. AABW spreads into the lower 1000 m of the world ocean, where it cools and renews oxygen concentrations drawn down by oxidation of organic material within the deep ocean. AABW is said to ventilate the deep ocean. In the Atlantic Ocean the 21C isotherm marks the base of a wedge of relatively salty water, associated with high dissolved oxygen and low silicate concentrations (see Figure 1). This water mass is called North Atlantic Deep Water (NADW). The densest component of NADW is formed as cold surface waters during the winter in the Greenland and Norwegian Seas. This water sinks to fill the basin north of a ridge spanning the distance from Greenland to Scotland. Excess cold water overflows the ridge crest, mixing on descent with warmer more saline water, producing a bottom water product of about þ 1.01C. The overflow water stays in contact with the sea floor to near 401N in the Atlantic Ocean, where on spreading southward it is lifted over the remnants of denser AABW.

Export of Greenland and Norwegian Sea bottom water has been estimated from a series of current measurements. Transports of about 2 106 m3 s1 of near 0.41C water occur between the Faroe Bank and Scotland, 1 106 m3 s1 of similar water passes through notches between Iceland and Faroe Bank, and 3 106 m3 s1 of near 01C is exported through the Denmark Strait, between Greenland and Iceland. The overflow plumes rapidly entrain warmer waters, producing bottom water of near þ 1.01C. With entrainment of other deep water, a production rate of about 8 106 m3 s1 of overflow water is likely. Less dense components of NADW, that do not contact the seafloor are formed in the Labrador Sea and Mediterranean Sea. The total production of NADW is estimated as 15 106 m3 s1. As the Antarctic is the primary source of the cold bottom waters of the world ocean, Antarctic Bottom Water is discussed in this article. See North Atlantic Deep Water for further information on that Northern Hemisphere deep water mass.

Formation of Antarctic Bottom Water Antarctic Bottom Water is formed at a few sites along the continental margin of Antarctica during the winter months (Figure 2). The shelf dense water slips across the shelf break to descend to the deep ocean, perhaps within the confines of incised canyons on the continental slope. The shelf water is made dense (Figure 3) by the very cold winds from Antarctica, which spur the formation of sea ice. Normally sea ice acts to insulate the ocean from further


BOTTOM WATER FORMATION

417

Cold win

d

Ice shelf

Coastal polynya

Sea surface

Sea ice Front

Meeting

Ice shelf wat

er

_2°C

Relatively warm deep water (~ +1°C)

_1.9°C Cold salty shelf water

se Den

Shelf

mes

r plu

wate Slope _1°C

Antartic Bottom Water 0°C

Figure 3 Schematic of circulation and water masses in the Antarctic continental shelf.

heat loss and thus attenuates the continued formation of sea ice. But along the shores of Antarctica sea ice is blown northward by the strong winds descending over the cold glacial ice sheet of Antarctica. The removal of the sea ice exposes the ocean water to the full blast of the cold air, forming coastal polynyas (persistent bands of ice-free ocean adjacent to Antarctica; Figure 4). Production and removal of yet more sea ice continues within the coastal polynyas, which act as ‘sea ice factories’. As sea ice has a lower salinity than the sea water from which it formed, approximately 5% versus 34.5% of the sea water, salt rejected during ice formation, concentrates in the remaining freezing point sea water making it saltier, and hence denser. The exposure of the ocean to the atmosphere also raises the dissolved oxygen concentration within the shelf water. As the ability of sea water to hold oxygen increases with lowering temperature, the oxygen concentration of the shelf water is very high, about 8 ml ll.

The shelf water at some sites, such as the Weddell and Ross seas, is made even colder on contact with floating ice shelves that are composed of glacial (freshwater) ice. Ocean contact with glacial ice occurs not only at the northern face of the ice sheet, but also at hundreds of meters depth along the bases of floating ice shelves. As the freezing point of sea water is lowered with increasing pressure ( 0.071C per 100 m of depth), the shelf water in contact with the base of the ice shelves often at a depth of many hundreds of meters, attains temperatures well below 2.01C. The cooling of shelf water in contact with the glacial ice is linked to melting of the glacial ice, hence this very cold water is slightly less saline than the remaining shelf water. Ocean–glacial ice interaction is believed to be a major factor in controlling Antarctica’s glacial ice mass balance and stability. The resultant water, called ice shelf water, can be identified as streams within the main mass of shelf water by its low potential temperatures. Ice shelf


418


Figure 4 Satellite image of a coastal polynya shown as the dark region extending from the ice-covered shore and sea ice on the left. (See also Polynyas.)

water with potential temperatures as low as 2.21C have been measured. This very cold water may act to encourage formation of AABW, as the seawater compressibility increases with lowered temperature. Thus as the shelf water begins its descent to the deep ocean, compressibility of the ice shelf water induces water of greater density, which accelerates the descent, limiting mixing with adjacent water. As shelf water escapes the shelf environs, offshore water must flow onto the shelf to compensate for the shelf water loss. The offshore water is warmer than the dense shelf water, with temperatures closer to þ 1.01C. Once on the shelf this water cools to renew the reservoir of freezing point shelf water. The onshore flow is drawn from deep water of the world ocean that slowly flows southward and upward, crossing the Antarctic Circumpolar Current. It may be viewed as ‘old’ AABW returning from its northern sojourn. In this way the overturning thermohaline circulation cell forced by AABW formation is closed; the escape of very cold shelf water, spreading northward, mixing en route with warmer overlying water, eventually results in upwelling to return to the Antarctic. The whole process takes some hundreds of years. Only now can the chlorofluorocarbons (CFC) added to the ocean surface layer in the last 70 or so

years, be detected reaching along the seafloor into the midlatitudes of the Southern Hemisphere.

Formation Rate of Antarctic Bottom Water Measurement of the formation and escape of Antarctic shelf water and the subsequent formation of Antarctic Bottom Water is difficult because the formation regions are geographically remote, covered by sea ice year-round, and distributed along a shelf break frontal region that extends more than 18 000 km around Antarctica. In addition, the thermohaline properties of source waters vary spatially, which can lead to quite different ideas regarding the mixing ‘recipes’ and specific processes leading to the ultimate cold products. It can be argued that for every 1 106 m3 of shelf water escaping 2 106 m3 of AABW is formed. The best observed bottom water formation is in the Weddell Sea (the extreme southern Atlantic Ocean). There, deep-reaching convective plumes of very cold surface water descend over the continental slope into the deep ocean (Figure 5), producing Weddell Sea Bottom Water. This is a particularly cold form of AABW, having initial potential temperatures



50

28 18 20 21 22 23 29 25 19 26

27 17

0

0

100

100

200

200

300

300

Pressure (dbar)

Pressure (dbar)

28 18 20 21 22 23 29 25 19 26

400 500 600

500 600 700

800

800 (°C) HELO 3

900

1000 _ 56

_ 55

_ 54

Longitude 28 18 20 21 22 23 29 25 19 26

_ 53

_ 57

50

_ 55 _ 54 Longitude

28 18 20 21 22 23 29 25 19 26

27 17

_ 53

50

27 17

0

500

500

1000

1000

Pressure (dbar)

Pressure (dbar)

_ 56

(B)

0

1500

2000

2500

1500

2000

2500

(°C) HELO 3

3000

3000

S HELO 3

3500

3500 _ 57 (C)

S HELO 3

1000 _ 57

(A)

27 17

400

700

900

50

419

_ 56

_ 55

_ 54

Longitude

_ 57

_ 53

_ 56

(D)

_ 55 _ 54 Longitude

_ 53

Figure 5 Potential temperature (A,C) and salinity (B,D) along a section at 671400 S in the western Weddell Sea, showing the varied stratification over the continental shelf and slope. The sea floor is stippled. The values inserted along the seafloor are the bottom temperature and salinity within the descending plume of dense water. The sharp change in water properties on the upper 500 m at 561W marks the shelf/slope front. (Reproduced from Gordon AL (1998) Western Weddell Sea thermohaline stratification. In: Jacobs SS and Weiss R (eds) Ocean, Ice and Atmosphere, Antarctic Research Series, vol. 75, pp. 215–240. Washington, DC: American Geophysical Union.)


420


at the seafloor of 1.51C. The transport of bottom water of less than 0.71C emanating from the Weddell Sea is estimated to be 2–5 106 m3 s1, presumably drawing from a shelf water flux of 1–3 106 m3 s1. Because of its coldness Weddell Bottom Water has a major effect on the bottom water properties of the World Ocean, particularly within the Atlantic Ocean which has the coldest bottom water. Circumpolar estimates for the formation rate of AABW, generally defined as having a potential temperature of less than 01C, are in the range of 10–15 106 m3 s1, but such values are not well constrained by the sparse observations. Based on CFC measurements a firm estimate of 9.4 106 m3 s1 has been made for the circumpolar production of AABW colder than 1.01C descending to depths greater than 2500 m. This value is similar to estimates of descending NADW from the Greenland and Norwegian Seas overflow. It is not clear what controls the rate of shelf water export from the continental shelf. At the edge of the shelf is a strong ocean front (Figure 5). Movement of this front may be associated with escape of dense shelf water. The presence of canyons incised into the continental slope may also act as paths for descent to the deep ocean.

Deep Convection Within the Southern Ocean In addition to descending dense water plumes over the continental slope of Antarctica, deep reaching convection over the deep ocean may occur. In winter, a thin veneer of sea ice stretches from Antarctica northward, reaching half the distance to the Antarctic Circumpolar Current. A delicate balance is achieved between the cold atmosphere and upward flux of oceanic heat into the atmosphere, resulting in the formation of approximately 0.6 m of sea ice. In this balance cold, low-salinity water sits stably over a warmer, more saline deep water layer. During the austral winters of 1974 to 1976, near the Greenwich Meridian and 661S in the vicinity of a seamount called Maud Rise, the ice displayed strange behavior, which has not been repeated since. A large region normally covered by sea ice in winter remained ice free throughout the winter, though it was surrounded by sea ice. This remarkable anomaly, is referred to as the Weddell Polynya. Though a full Weddell Polynya has not been observed since the mid-1970s, short lived polynyas (lasting roughly one week) are frequently observed by satellite imaging in the Maud Rise region. During the Weddell Polynya

episode the normal stratification was disturbed as cold surface water convected to depths of 3000 m. It is estimated that during the three winters of the persistent Weddell Polynya, 1–3 106 m3 s1 of surface water entered the deep ocean, as a form of AABW. The Weddell Polynya convection may represent another mode of operation of Southern Ocean processes, one in which surface water sinks into the deep ocean at sites away from the continental margins, in what oceanographers refer to as open ocean convection, in contrast to the continental slope plume convection discussed above. It is not clear what triggered the Weddell Polynya of the mid-1970s. It is clear that if enough deep water can be brought to the sea surface to melt all of the ice cover, then further upwelled deep water, not to be diluted by ice melt, would produce cold saline water that can then sink back into the deep ocean. What would cause enhanced upwelling of deep water? This is not clear, though interaction of ocean circulation with the Maud rise is suspected to play a key role.

Conclusion Bottom water of the western Weddell Sea in the early 1990s is colder and fresher than observed in previous decades. The same is true for the Southern Ocean south of Australia. These observations suggest an increased role of the low salinity, ice shelf water in recent decades. However, the period over which observations have been taken within the hostile Southern Ocean is not long enough to place much importance on this trend. The presence of sea ice and the rather small spatial and temporal scales associated with the convective plumes, makes AABW formation processes very difficult to observe and model. Advancement of AABW research represents a significant technological challenge to field and computer oceanographers.

Summary The ocean is cold. Its average temperature of 3.51C is far colder than the warm veneer capping much of the ocean. Waters warmer than 101C amount to only 10% of the total ocean volume; about 75% of the ocean is colder than 41C. Along the seafloor the ocean temperature is near 01C. The cold bottom water is derived from Southern Ocean, the ocean belt surrounding Antarctica. Sea water at its freezing point of 1.91C is formed in winter over the continental shelf of Antarctica. Where the salt content of shelf water is high enough, roughly 34.61% (parts



per thousand) the water is sufficiently dense to descend as convective plumes over the continental slope into the adjacent deep ocean. In so doing Antarctic Bottom Water is formed. It is estimated that on average between 10 and 15 106 m3 of Antarctic bottom water forms every second! Antarctic Bottom Water spreads away from Antarctica into the world oceans, chilling the deep ocean to temperatures near 01C. Bottom water warms en route on mixing with warmer overlying waters. Cold winter water also forms in the Greenland and Norwegian Seas of the northern North Atlantic. This water ponds up behind a submarine ridge spanning the distance from Greenland to Scotland. This water overflows the ridge crest into the ocean to the south. As the overflow water mixes with warmer saltier water during descent into the deep ocean, it results in a warmer, more saline water mass than Antarctic Bottom water. The Greenland and Norwegian Sea overflow water forms the densest component of the water mass called North Atlantic Deep Water and is estimated to form at a rate of around 8 106 m3 s1. The overflow water stays in contact with the seafloor until the northern fringes of Antarctic Bottom Water encountered in the North Atlantic near 401N, lifts it to shallower levels.

See also

421

Further Reading Fahrbach E, Rohardt G, Scheele N, et al. (1995) Formation and discharge of deep and bottom water in the northwestern Weddell Sea. Journal of Marine Research 53(4): 515--538. Foster TD and Carmack EC (1976) Frontal zone mixing and Antarctic Bottom Water formation in the southern Weddell Sea. Deep-Sea Research 23: 301--317. Gordon AL and Tchernia P (1972) Waters off Adelie Coast. Antarctic Research Series, vol. 19, pp. 59--69. Washington, DC: American Geophysical Union. Jacobs SS, Fairbanks R, and Horibe Y (1985) Origin and evolution of water masses near the Antarctic continental marginE: vidence from H2 18 O=H2 16 O ratio in seawater. In: Jacobs SS (ed.) Oceanography of Antarctic Continental Margin, Antarctic Research Series, vol. 43, pp. 59--85. Washington, DC: American Geophysical Union. Jacobs SS and Weiss R (eds.) (1998) Ocean. Ice and Atmosphere: Interactions at the Antarctic Continental Margin, Antarctic Research Series, vol. 75. Washington DC: American Geophysical Union. Nunes RA and Lennon GW (1996) Physical oceanography of the Prydz Bay region of Antarctic waters. Deep-Sea Research 43(5): 603--641. Orsi AH, Johnson GC, and Bullister JL (1999) Circulation, mixing, and production of Antarctic Bottom Water. Progress in Oceanography 43: 55--109. Tomczak M and Godfrey JS (1994) Regional Oceanography: An Introduction. London: Pergamon Press.

Antarctic Circumpolar Current. Polynyas. Rotating Gravity Currents. Sub Ice-Shelf Circulation and Processes. Weddell Sea Circulation.


BRAZIL AND FALKLANDS (MALVINAS) CURRENTS A. R. Piola, Universidad de Buenos Aires, Buenos Aires, Argentina R. P. Matano, Oregon State University, Corvallis, OR, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 340–349, & 2001, Elsevier Ltd.

Introduction The zonal component of the mean prevailing winds, low latitude easterlies and mid-latitude westerlies induce anticyclonic1 upper ocean circulation patterns referred to as subtropical gyres. The latitudinal rate of change of the Earth’s rotation induces a zonal asymmetry in these gyres, and intensifies the flow near the western boundaries. The Brazil Current is the western limb of the subtropical gyre that carries warm and salty waters poleward along the continental slope of South America (Figure 1). Near 391S the Brazil Current collides with a northward branch of the Antarctic Circumpolar Current (ACC), the Malvinas (Falkland) Current, which transports cold and relatively fresh subAntarctic waters equatorward. The collision between these distinct water masses generates one of the most energetic regions of the world ocean: the Brazil/Malvinas Confluence (BMC). This article reviews in situ and remote observations and the results of numerical simulations that describe the mean structure and time variability of the Brazil and Malvinas Currents and the frontal region that they generate.

from the North Atlantic, South Pacific, and Antarctic regions. To illustrate the water mass structure of the upper layer of the western South Atlantic, Figure 2 shows a diagram of potential temperature2 versus salinity (y– S) from summer stations collected within the cores of the Brazil and Malvinas Currents (1000–2000 m depth range). From 201S to 351S, Figure 2 illustrates the y/S characteristics associated with the water masses advected by the Brazil Current and, from 551S to 401S, with those advected by the Malvinas Current. In addition, Figure 2 also shows the y–S diagram of a hydrographic station collected downstream from the separation of both boundary currents from the continental margin (e.g. within the core of the BMC). Upper Ocean

The western South Atlantic has been referred to as the ‘cross-roads of the world ocean circulation’, because it hosts water formed in remote areas of the world, and brought into this region by the large-scale ocean circulation. This meeting of water masses generates a highly complex vertical stratification structure. In the upper ocean, this structure is dominated by the confluence of subtropical and subAntarctic waters associated with the opposing flows of the Brazil and Malvinas Currents. In the deep ocean, the vertical stratification structure is dominated by contributions from deep and bottom waters

The upper portion of the water mass carried poleward by the Brazil Current is referred to as Tropical Waters (TW), and is characterized by high potential temperature (y4201C) and salinity (S436 PSU, Figure 2). The high temperatures of the TW are due to heat gained from the atmosphere at low latitudes, while the high salinities are due to freshwater losses at mid-latitudes. The upper portion of the Brazil Current is also characterized by the presence of relatively thin low salinity layers capping the TW structure (e.g. the 351S curve in Figure 2). These low salinity layers are thought to be caused by mixing between TW and shelf and river waters. Below the TW, but still within the Brazil Current, there is a sharp thermocline and halocline (see the quasilinear y–S relation in the 20–101C temperature range) that is referred to as South Atlantic Central Water (SACW). The SACW shows a very stable y–S pattern with only minor variations induced by winter sea–air interactions near the southern limit of the Brazil Current. The upper layer of the Malvinas Current (i.e. the curves corresponding to 401 and 501S in Figure 2) is substantially colder (yo151C) and fresher (So34.2 PSU) than the corresponding layer of the Brazil Current. These properties reflect the subAntarctic origin of the Malvinas waters. In the northern portion of the Drake Passage, the source for the Malvinas transport, the surface temperature is close to

1 Clockwise in the northern hemisphere and counter-clockwise in the southern hemisphere.

2 y is the temperature of a water parcel raised adiabatically to the sea surface, thus removing the effect of pressure.

Water Masses

422


BRAZIL AND FALKLANDS (MALVINAS) CURRENTS

423

10

>37

15 °S 20

Br az

25

C. il >36.2

30

S

35

outh Atl

an

tic

C.

Malvinas C

40

45

.

34

.2

po

l ar C .

50

Burdwood B.

55

Drake

Antarctic

Circ

um

Ewing Bank

N. Scotia Ridg e

Scotia Sea

Passage 60 70

60

50

40

°W

30

Figure 1 Schematic diagram of the upper layer circulation of the South Atlantic western boundary currents. Black lines are used for the Antarctic and subAntarctic water flows, associated with the Antarctic Circumpolar Current and the Malvinas Current. Red lines are used for the flow of the subtropical waters carried by the Brazil Current. Over the Patagonian continental shelf the arrows represent the mean surface currents. The thin contour lines show the salinity field at 200 m depth that was used to infer part of the circulation scheme. The salinity at 200 m ranges from 34.2 south of the Confluence to 37 near 151S. A sharp salinity front develops at the Brazil/ Malvinas Confluence and extends in a meandering fashion towards the ocean interior where it marks the South Atlantic Current. The background shading represents the bottom topography with darker shading corresponding to deeper waters. The deepest area in the southern Argentine Basin is the Argentine Abyssal Plain where depth is greater than 6000 m. Major topographic features and currents cited in the text are labeled.

41C and increases northward up to 161C at the latitude where the Malvinas separates from the continental boundary (B401S). Although the subAntarctic waters of the Malvinas Current and the SACW of the Brazil Current thermocline occupy the same density range (syB25.5–27.0 kg m3) they

have very different thermohaline characteristics and the convergence of these water masses, in the BMC, leads to the formation of alternate layers of subAntarctic and subtropical water. These intrusions are referred to as interleaving or fine-structure (see station at BMC, Figure 2).


424


30

24 20°S

25

35°S

25 Brazil / Malvinas Confluence

TW

26


30°S 20

27 SACW

15

28 40°S 10

50°S 5 55°S NADW AAIW

UCDW

0 34

35

36

37

Salinity (PSU) Figure 2 Potential temperature–salinity diagrams from hydrographic stations collected during austral summer along the paths of the Brazil Current (from 201S in the Brazil Basin to 351S, solid lines) and along the Malvinas Current (from 551S in the northern Drake Passage and 401S, dashed lines). These stations are located between the 1000 m and 2000 m isobaths near the cores of the western boundary currents. Also included is a station from the Brazil/Malvinas Confluence after separation from the western boundary (dashed-dotted line). Constant density anomaly (sy in units of km m3) lines are included. See Figure 3 for abbreviations.

Antarctic Intermediate Water

The water mass structure of the Brazil Current at intermediate depths (700–1000 m) is dominated by the presence of Antarctic Intermediate Water (AAIW). The AAIW, characterized by a salinity minimum (So34.3 PSU), has contributions from the coldest and densest (syB27.3 kg m3) member of the southern hemisphere Subpolar Mode Water or SubAntarctic Mode Water (SAMW), which originates from deep winter convection along the SubAntarctic Zone. The Malvinas Current carries newly formed AAIW and SAMW into the Argentine Basin. Data collected during the austral winter show that, as the AAIW/SAMW enter into the Argentine Basin from the south, they are exposed to the atmosphere and are subject to further modification by local air–sea interactions. South of the BMC, the AAIW/SAMW

are less salty (So34.1 PSU) than within the Brazil Current (Figure 2) and these lateral property gradients across the BMC induce interleaving. Similarly to the upper layer flow, the temperature of the AAIW core increases from 31C, at the Drake Passage, to 3.51C at 401S. It is interesting to note that although on average the AAIW must spread northward (away from the region of formation), direct current observations at 281S, and close to the continental margin indicate that, in the subtropical basin, the AAIW follows the upper ocean anticyclonic gyre, and flows southward below the Brazil Current. After leaving the continental boundary the AAIW turns into the subtropical gyre, where vertical and lateral mixing increase its salinity and decrease its dissolved oxygen concentration. The water mass resulting from this recirculation process is known as recirculated AAIW.



Deep and Abyssal Water

Circulation

The deep layers of the western South Atlantic show a variety of water masses which are depicted by their properties in Figure 3. Below the Brazil Current, there is the poleward flow of North Atlantic Deep Water (NADW), which is the primary source of ventilation underneath the main thermocline. The NADW originates at the high latitudes of the North Atlantic Ocean, from where it spreads southward along the continental slope of the American continent. At 301S the NADW is characterized by relatively high potential temperature (yB31C), salinity (SB34.8 PSU), and dissolved oxygen (O2B250 mmol kg1). Below 800– 1000 m Circumpolar Deep Water (CDW) flows northward within the Malvinas Current. Although the nutrient-rich CDW originates from NADW, mixing along its path around the Antarctic Continent leads to decreased concentrations of oxygen and salinity. Consequently, although CDW is still identified by a relative salinity maximum, its salinity at the core still is lower than that of NADW. In the western Argentine Basin the NADW splits the CDW into two layers: the upper CDW (UCDW) and the lower CDW (LCDW). The latter are identified by two minima in dissolved oxygen above and below the high salinity, oxygen-rich NADW. The existence of two separate oxygen minimum layers is apparent north of 501S (Figure 3). From the Drake Passage the UCDW flows into the Argentine Basin closely following the 1000–1500 m isobaths. At 401S the UCDW is characterized by deep (B1400 m) temperature (yo2.91C) and dissolved oxygen minima (O2o200 mmol kg1). The LCDW is the densest water flowing eastward through Drake Passage. It enters the Argentine Basin over the Falkland Plateau and primarily east of Ewing Bank, flows westward along the escarpment located at 491S and continues northward along the continental slope of the Argentine Basin at 3000–3500 m depth (Figure 3). The abyssal waters of the southern hemisphere oceans are derived from southern high latitudes and are generally referred to as Antarctic Bottom Water. In the western South Atlantic the bottom waters are cold (yo01C), oxygen-rich (O2B225 mmol kg1), and nutrient-rich. These abyssal waters are denser and colder than the densest water found in the Drake Passage and must derive from the Weddell Sea. Underneath the continental ice shelves of the southern Weddell Sea the densest water mass of the world ocean is formed, but it is the Weddell Sea Deep Water (WSDW), a product of mixing between the CDW and the Weddell Sea Bottom Water, which flows northward around the Scotia Trench and enters into the Argentine Basin as an abyssal western boundary current.

Brazil Current

425

The Brazil Current originates along the continental slope of South America, between 101 and 151S, through a branching of the westward-flowing South Equatorial Current. The northern branch of the South Equatorial Current forms the North Brazil Current, and represents a loss of upper layer mass from the South Atlantic to the North Atlantic. The southern branch forms the Brazil Current, the western boundary current of the subtropical South Atlantic Ocean. A substantial amount of the southward upper ocean flow occurs on the outer continental shelf of Brazil. Although the term Brazil Current usually refers to the flow within the upper 1500 m, there is evidence that the current may extend well beyond that depth. In fact, hydrographic observations suggest that the AAIW layer is also part of the southward-flowing western boundary current. Direct current measurements off southern Brazil also reveal that although the upper layer flow of the South Equatorial Current reaches South America near 151S, at intermediate depths the bifurcation shifts south of 241S. The addition of re-circulated AAIW to the southward flow would contribute to the increase of volume transport of the Brazil Current observed south of approximately 281S. Geostrophic calculations, and a few direct current measurements of the Brazil Current transport, yield a value of only about 4–6 Sv, between 10 and 201S, and this increases to about 20 Sv at 381S, near the BMC. The rate of transport increase for the Brazil Current is comparable to that of the Gulf Stream. The increase of the Brazil Current’s transport is partially associated with a tight recirculation cell near the western boundary and, perhaps, the addition of intermediate waters near 251S. In situ observations, between 201 and 281S, have shown that the poleward increase of volume transport of the Brazil Current to 16 Sv is associated with a deepening of the current from 100 m to 600 m. While the discussion on the Brazil Current’s transport has focused on the upper 1000 m, there are also important poleward, western boundary undercurrents below the thermocline. At 271S, for example, the core of the southward-flowing NADW (S434.94 PSU) is found at approximately 2000 m and east of the upper ocean jet. If this undercurrent is included in the transport calculation then the southward volume transport relative to a deep reference level is close to 11 Sv at 271S, and increases southward to 70–80 Sv at 361S. Although this estimate may include some southward recirculation of subAntarctic water and CDW from the Malvinas Current it is, nevertheless, much larger than previous values.



4000

3500

3000

2500

2000

1500

1000

500

0

16 13. 10. 0 0 8.0

3.2 3.0

4.0

200

0 .2

5

1.0

2.0

2.5

2.8

< _ 0.1

WSDW 0

50

Salinity

.20

100

>34.9

Distance (km)

0

34.2

150

200

0

34.7

34.80

5

34.8

NADW

34. 80 34.85

0

34.7

0

.5

34

20

34.

AAIW

.20

34

>35.6

0

_1

50

300

150

220

200

22

ag, where a ¼ 0.5 for the limiting Stokes wave but ao0.39 for the so-called almost steepest waves, and Z(t) is the surface elevation record at a fixed point. This threshold mechanism has been the basis for several theoretical studies of wave breaking; however, it is not conclusively supported by observations, where wave breaking has been observed at accelerations below the threshold value. Similarly, the theoretical maximum steepness of the limiting Stokes wave, ak ¼ p/7, where a and k are wave amplitude and wave number, respectively, is hardly ever observed in the field, and extensive observations reveal that breaking and nonbreaking waves cannot be separated on the basis of local steepness alone. Wave breaking may occur at all wave scales. However, the spectral distribution of breakers depends on the wave development and the most common breakers are associated with waves of higher frequencies than op, the frequency at the peak of the energy spectrum. In fact, only in young wave fields,

t=T No wave breaking

t = 2T Wave breaking at x =

Figure 1 Idealized wave breaking periodicity. Waves tend to break in the center of wave groups. Thus, breakers are separated by one wave length l, and the period between successive breakers equals two wave periods T.


BREAKING WAVES AND NEAR-SURFACE TURBULENCE

where the wave saturation at the peak is sufficiently large, are breaking waves observed at all scales, including the dominant waves. As the wave field develops the distribution of breaking scales narrows and its peak shifts further away from the dominant wave frequency toward higher frequencies. The total breaking rate, defined as the number of breaking waves of all frequencies passing a fixed location, depends on the distribution of breaking scales and thus on the shape of the wave saturation s(o). Limiting the breaking spectrum to breaking waves that result in visible whitecaps, one typically finds about 50–100 breakers per hour for open-ocean conditions and 12-m s 1 wind speed.

Turbulence beneath Breaking Waves Direct measurement of the fine-scale velocity field in the ocean and especially in the near-surface layer is extremely challenging. Surface waves are a source of enhanced turbulence. However, they also create the major difficulties in near-surface velocity measurements. Typical turbulent velocity fluctuations are O(10 2–10 1 m s 1) and thus are 10–100 times smaller than the wave-related velocities. In terms of kinetic energy, the wave motion contains 2–4 orders of magnitude higher energy levels than the turbulent motion. Furthermore, the nonlinear advection associated with the wave orbital motion modulates the turbulent flow observed at a fixed mooring and affects the dissipation estimates obtained from single point velocity records that rely on Taylor’s hypothesis of frozen turbulence. Nevertheless, significant progress in measuring wave-induced turbulence under natural conditions has been made, starting around the mid-1980s. However, most detailed information stems from controlled laboratory experiments, although the breaking characteristics in these studies are often very different to those of natural breaking waves. Focused superposition of dispersive mechanically generated waves leads to well-defined wave breaking even in the absence of wind forcing. This setup allows repeatable turbulence measurements. Turbulence beneath these breaking waves is seen to spread downward; within the first two wave periods this spreading is approximately a linear function of time and occurs more slowly thereafter. The final spreading depth of roughly twice the wave height is reached after four wave periods and by then about 90% of the energy lost by the breaking wave has been dissipated. Thereafter, the remaining decaying turbulence spreads only slightly further and may persist for tens of wave periods but can only be detected in an otherwise rather calm environment.

433

Behind the breaking crest, vortices of size comparable to the wave height are generated. These rotors may play an important role in mixing gases and pollutants such as small oil droplets. Nearly half of the energy lost from the breaking wave is associated with the entrainment of air bubbles, although part of it will be converted into turbulence kinetic energy (TKE) as larger bubbles rise through the water column. Under natural conditions, turbulence is commonly characterized by the dissipation rate of TKE e, which may be inferred from the turbulence velocity shear @u/@z or rate of strain @u/@x, or from wavenumber velocity spectra S(k). Thus, two fundamentally different approaches exist in oceanic turbulence measurements: (1) observation of the velocity shear or the rate of velocity strain, and (2) observation of the velocity field in space or time. In isotropic turbulence, the rate of dissipation is related to the rate of strain or the turbulence shear by

@u 2 15 @u 2 ¼ n e ¼ 15n @x 2 @z where n is the kinematic viscosity of the fluid, u the horizontal velocity component, and x and z are the horizontal and vertical coordinates, respectively. These relations are the basis for the pioneering studies of near-surface turbulence measurements made with towed hot-film anemometers, electromagnetic current meters, and the common microstructure profilers utilizing airfoil shear probes. The second class of turbulence measurements makes use of Kolmogorov’s inertial subrange hypothesis; within a subrange of the wavenumber band the onedimensional wavenumber spectrum S(k) ¼ Ae2/3k 5/3 has a universal form which depends only on the energy dissipation, where k is the wavenumber and A is a universal constant. This simple relationship between energy dissipation and wavenumber spectra allows the estimation of e from velocity measurements. Due to recent advances in sonar technology it is now possible to resolve instantaneous velocity profiles at spatial scales of a few millimeters and temporal resolution of a tenth of a second. These scales are suitable for turbulence measurements in the upper ocean. A common reference level for turbulence studies in boundary layers is the flow along a rigid wall. In this classic reference case, often labeled wall layer or constant stress layer, the velocity profile is logarithmic and the turbulent stress in the inner boundary layer t ¼ ru2 is nearly constant, where r is the fluid density and u the friction velocity. The TKE


434


dissipation per unit mass is given by e ¼ u3 ðkzÞ1, with k ¼ 0.4 being the von Ka´rmań constant. Many studies show that turbulence in the ocean surface layer is enhanced compared to turbulence in a constant stress layer and there is strong evidence that the turbulence enhancement is due to breaking waves. The magnitude as well as the depth dependence of the time-averaged TKE dissipation in the near-surface layer of a wind-driven ocean departs significantly from the classic constant-stress-layer form. Observations indicate that the surface layer may be divided into three regimes. In the top layer wave breaking directly injects TKE down to a depth zb. In this injection layer, dissipation is highest and most likely depth-independent. Below this layer, the waveinduced turbulence diffuses downward and dissipates, as has been also demonstrated in the laboratory experiments. In this diffusive region, the decay of turbulence with depth is stronger than the wall-layer dependence epz 1. However, the exact depth dependence of the wave-induced turbulence is not well established and profiles consistent with epzn, with n in the range 4 to 2, as well as exponential profiles, epe z, have been observed. Some open ocean observations under strong wind forcing and significant swell revealed enhanced dissipation values but depth dependence consistent with wall-layer scaling. Further down in the water column at a depth zt, sufficiently far from the air–sea interface, the contribution of waves becomes small compared to local shear production, and turbulence properties are well described by the constant stress layer scaling. There is, as yet, no conclusive observational evidence for the vertical extension of the different regimes. In particular, the depth of direct TKE injection dominates the total dissipation. Thus, zb is a crucial parameter in turbulence closure models, where it is implemented as the surface mixing length, which will also affect the mean profiles of tracers such as salt or heat. Mean dissipation profiles are commonly referenced to the mean water surface, and the oscillation of the sea surface poses a challenge for observation and interpretation of near-surface turbulence. Mooring or tower-based observations are limited to observations below the troughs, and depth is referenced to the mean still water line. Surface-following measurements from floats or ships are, in principle, also suitable to monitor the region above the troughs and depth reference is made with respect to the instantaneous surface. For example, in waves of 0.5-m wave height, a nominal 1-m depth observations from a tower is equivalent to a surface-referenced depth varying from 0.75 m at the location of the wave trough to 1.25 m in the crest region. The same

observation from a float would be converted to depth values ranging from 1.25 m at the trough to 0.75 m at the crest, if referenced to the still water line (see Figure 2). Due to the strong depth dependence of dissipation, the choice of coordinate systems affects the resultant dissipation profile. Microstructure profilers operated in a rising mode are capable of observing turbulence up to the sea surface. However, enhanced dissipation levels associated with wave breaking are very intermittent, and the profiling frequency is too low to adequately resolve these events. Despite the observational challenges, quality turbulence data in the aquatic near-surface layer in the presence of breaking waves have been collected starting in the mid-1980s. (Pioneering studies started in the 1960s.) The general consensus is that the enhancement of average dissipation en ¼ e=ðu3 ðkzÞ1 Þ in the diffusive layer is of order 10–100, where e is taken as the mean over a few minutes, and the depth of the wave enhanced layer (zt) is confined to a depth corresponding to a few times the significant wave height, that is 2–6 m typically. However, under severe storm conditions, the extent of the wave enhanced layer could be more than 10 m. Bubble clouds generated by wave breaking have been observed to such depths, but it is not known to what extents wave turbulence or coherent structures such as Langmuir circulation are the responsible bubble transport mechanisms. The depth of the injection layer zb is not well established. Estimates range from zb ¼ O(0.1 Hs) to zbXHs, or about 0.2–1 m. Instantaneous dissipation levels can be much higher than seen in the mean profiles. Beneath an active breaking wave, turbulence enhancements en ¼ O(104) have been observed. These high values persist only for a few seconds. Turbulence beneath individual breaking waves decays as eptm, where observations indicate m E 4 in the diffusion layer and m E 7.6 in the injection layer. Approximately five wave periods after the onset of breaking, the turbulence levels have decayed to the background level of wall-layer flows. Thus, sufficiently fast sampled dissipation measurements reveal the coexistence of two distinct contributions, a wide distribution centered on constant-stress-layer turbulence levels (log(en)E0) and a smaller and narrower distribution representing breaking waves and centered on log(en)X2. The broader distribution of lower enhancement rates is associated with periods between breaking events and is broadly consistent with a wall-layer flow. The largest turbulence levels, occurring beneath the actively breaking crest, play an important role in the breakup of air cavities and thus determine the initial bubble size distribution.



Tower observation reference: mean water line

435

Float observation reference: mean water line

c

c

t t

Tower observation reference: free surface

Float observation reference: free surface

c

c

t

t

Figure 2 Perceived depth of a measurement, depending on observation platform and choice of surface reference.

The balance between surface tension gw and turbulent pressure forces leads to the so-called Hinze scale aH, which describes the resulting bubble radius: aH ¼ Aðgw =rÞ3=5 e2=5 , where A is a constant in the range 0.36–0.5. Field observations yielded aHC10 3m, but more observations are required to establish a possible range of these initial bubble sizes. In spilling breakers, wave breaking occurs on the wave crest and turbulence levels have decayed significantly by the time the succeeding trough is reached. Therefore, turbulence levels in the crest region are larger than in the trough region and more than half of the energy is dissipated above the mean water line.

Wave breaking is a very intermittent phenomenon and the resulting turbulence fields are very patchy. Therefore, long time series or a suite of several sensors are required to obtain reliable statistics and sound estimates of the contribution of wave breaking to upper ocean processes. Alternatively, it was suggested in 1985 that the scale and strength of breaking may be characterized by the length of the breaking crest and its propagation speed. This opens up the possibility of remote sensing of the integral contribution of processes associated with wave breaking. The central quantity in this concept is the spectral density function L(c). It is defined in a way that L(c)dc describes the average total length of breaking


436


wave crests (perpendicular to the wave propagation), per unit area, that have speeds in the range c to c þ dc. Within a given surface area there might be several breaking crests at any given time, many of them that are only breaking along a fraction of the total crest lengths. To determine L(c), breaking crests that propagate at a similar speed are combined, the total lengths of these breaking crests are added up, and then the sum is divided by the area of the observed surface patch. The passage rate of breaking crests propagating at speed c past a fixed point is cL(c). The Rfractional surface turnover rate per unit time is R ¼ cL(c)dc, which can also be interpreted as the breaking frequency at a fixed point, that is the number of breakers passing a fixed location per unit time. Furthermore, the fourth and fifth moment of this spectral density function may be related to the dynamics of wave breaking. These relationships are based on similarity scaling of breakers and were confirmed in wave tank experiments. Therefore a further challenge is the proper scaling of the laboratory experiments to wave scales observed under natural conditions. Quasi-steady breakers can be generated by towing a submerged hydrofoil along a test channel. These experiments established that the rate of energy loss per unit length of breaking crest is proportional to c5. Therefore, the wave energy dissipation due to the breaking of waves of scale corresponding to phase speed c is e(c)dc ¼ brg 1 c5L(c)dc, where b is an unknown, nondimensional proportionality factor, originally assumed to be constant. However, b might depend on nondimensional expressions of, for example, the wave-scale or the wave-field nonlinearity. The total energy dissiassociated with whitecaps is E ¼ brg 1 Rpation 5 c L(c)dc. Momentum and energy are related by M ¼ Ec 1 and the spectrally resolved momentum flux from breaking waves to currents is m(c)dc ¼ brg 1c4L(c)dc. The total momentum flux from the field to currents is given as M ¼ brg 1 Rwave 4 c L(c)dc. Evaluation of this integral over all scales (phase speeds c) of breaking waves, including microbreakers, and assuming no wave growth in space yields M ¼ tw , where tw is the atmospheric momentum flux supported by the form drag of the waves. This momentum flux balance might prove to be a key relation in estimating the proportionality factor b. However, so far observations of the breaking crest density L(c) in the ocean do not adequately resolve the small-scale waves and microbreakers of the breaking spectrum. In a wind-driven sea, breaking waves provide the strongest contribution to near-surface turbulence. Other effects of surface waves are increased

dissipation levels prior to the onset of air entrainment. This prebreaking turbulence is consistent with wave–turbulence interaction in a rotational wave field. Increased near-surface Reynolds stresses due to the Stokes drift of nonbreaking waves may increase horizontal transports. Interaction of the Stokes drift with the wind-driven current may trigger Langmuir circulation, conceptually described as counterrotating cells aligned in the wind direction. Langmuir cells are large eddies and may be an important part of the turbulence field as well as influencing the turbulence generated by other processes such as breaking waves. Little is known how small-scale wave-enhanced turbulence affects lateral dispersion; however, it is likely that this process is dominated by Langmuir turbulence.

Conclusion Especially in mid- to high latitudes breaking waves are a ubiquitous feature of the open ocean. The last two to three decades have seen increased study of breaking deep water waves and new insight has been gained from laboratory experiments and field observations as well as through theoretical wave modeling. Earlier attempts to relate breaking to geometrical or kinematic features of individual wave crests are slowly being replaced by the concept of nonlinear hydrodynamics of the wave field leading to wave breaking. Wave breaking plays an important role in many processes of air–sea interaction, and the waveinduced turbulence is a relevant quantity in assessing its contributions. Near-surface turbulence observations show a mean dissipation enhancement of 1–2 orders of magnitude due to the effect of wave breaking. However, the detailed structure of the turbulence field and the length scales involved are still not resolved conclusively. Instantaneous dissipation levels are up to 4 orders larger than dissipation levels in a wall-layer flow. These initial high turbulence levels decay rapidly, but are likely to play a defining role in the breakup of air bubbles and thus in the setup of the bubble size distribution. Turbulence levels are highest beneath the wave crest and there is a need for more observations with adequate sampling of the region above the trough line. The concept of relating breaking wave kinematics and dynamics to whitecap properties that can be observed remotely, say with aerial video imagery or subsurface acoustical tracking, is intriguing and holds the promise of new observational insight in wave breaking processes. However, any quantitative assessments of energy dissipation and momentum fluxes based on this concept directly depend on the



proportionality factor b. Currently, only very limited data exist and estimates of b are inconclusive; in fact, it is not even established that b is constant.

Nomenclature a aH c E g k m(c) M R

S(o) u u x z zb zt gw e en e(c) k L(c) n r s(o) t tw o op

wave amplitude Hinze scale wave phase speed total energy dissipation associated with wave breaking gravitational acceleration wave number spectrally resolved momentum flux from breaking waves to currents total momentum flux associated with wave breaking fractional surface turnover rate per unit time; equivalent to breaking frequency at a fixed location wave energy spectrum horizontal velocity component friction velocity horizontal coordinate vertical coordinate injection layer depth depth of enhanced wave-induced turbulence surface tension dissipation rate of turbulent kinetic energy enhancement of average dissipation spectrally resolved energy dissipation by breaking waves von Ka´rmań constant spectral density function of breaking crest lengths kinematic viscosity fluid density wave saturation turbulent stress atmospheric momentum flux supported by the form drag of the waves wave frequency frequency of the peak of the wave energy spectrum

See also Air–Sea Gas Exchange. Bubbles. Estimates of Mixing. Langmuir Circulation and Instability. Rogue Waves. Surface Gravity and Capillary Waves. Turbulence in the Benthic Boundary Layer. Wave Generation by Wind.

437

Further Reading Banner ML (2005) Rougue waves and wave breaking – how are these phenomena related? In: Mu¨ller P and Henderson D (eds.) Proceedings ‘Aha Huliko’ a Hawaiian Winter Workshop, Jan. 2005. http://www.soest.hawaii.edu/Pub Services/2005pdfs/Banner.pdf (accessed Feb. 2008). Banner ML and Peregrine DH (1993) Wave breaking in deep water. Annual Review of Fluid Mechanics 25: 373--397. Baschek B (2005) Wave-current action in tidal fronts. In: Mu¨ller P and Henderson D (eds.) Proceedings ‘Aha Huliko’ a Hawaiian Winter Workshop, Jan. 2005. http:// www.soest.hawaii.edu/PubServices/2005pdfs/Baschek.pdf (accessed Feb. 2008). Colbo K and Li M (1999) Parameterizing particle dispersion in Langmuir circulation. Journal of Geophysical Research 104: 26059--26068. Donelan MA and Magnusson AK (2005) The role of focusing in generating rogue wave conditions. In: Mu¨ller P and Henderson D (eds.) Proceedings ‘Aha Huliko’ a Hawaiian Winter Workshop, Jan. 2005. http:// www.soest.hawaii.edu/PubServices/2005pdfs/donelan.pdf (accessed Feb. 2008). Garrett C, Li M, and Farmer DM (2000) The connection between bubble size spectra and energy dissipation rates in the upper ocean. Journal of Physical Oceanography 30: 2163--2171. Gemmrich J (2005) A practical look at wave-breaking criteria. In: Mu¨ller P and Henderson D (eds.) Proceedings ‘Aha Huliko’ a Hawaiian Winter Workshop, Jan. 2005. http://www.soest.hawaii.edu/PubServices/2005pdfs/ Gemmrich.pdf (accessed Feb. 2008). Gemmrich JR and Farmer DM (1999) Observations of the scale and occurrence of breaking surface waves. Journal of Physical Oceanography 29: 2595--2606. Gemmrich JR and Farmer DM (2004) Near surface turbulence in the presence of breaking waves. Journal of Physical Oceanography 34: 1067--1086. Holthuijsen LH and Herbers THC (1986) Statistics of breaking waves observed as whitecaps in the open sea. Journal of Physical Oceanography 16: 290--297. Melville WK (1996) The role of surface-wave breaking in air–sea interaction. Annual Review of Fluid Mechanics 26: 279--321. Melville WK and Matusov P (2002) Distribution of breaking waves at the ocean surface. Nature 417: 58--62. Mu¨ller P and Henderson D (eds.) (2005) Proceedings ‘Aha Huliko’ a Hawaiian Winter Workshop, Jan. 2005. http:// www.soest.hawaii.edu/PubServices/2005pdfs/TOC2005. html (accessed Feb. 2008). Phillips OM (1985) Spectral and statistical properties of the equilibrium range in wind-generated gravity waves. Journal of Fluid Mechanics 156: 505--531. Rapp R and Melville WK (1990) Laboratory measurements of deep water breaking waves. Philosophical Transactions of the Royal Society of London A 331: 735--780. Song J-B and Banner ML (2002) On determining the onset and strength of breaking for deep water waves.


438


Part 1: Unforced irrotational wave groups. Journal of Physical Oceanography 32: 2541--2558. Sullivan PP, McWilliams JC and Melville WK (2005) Surface waves and ocean mixing: Insights from numerical simulations. In: Mu¨ller P and Henderson D (eds.) Proceedings ‘Aha Huliko’ a Hawaiian Winter Workshop, Jan. 2005. http://www.soest.hawaii.edu/PubServices/2005 pdfs/Sullivan.pdf (accessed Feb. 2008).

Terray EA, Donelan MA, Agrawal YC, et al. (1996) Estimates of kinetic energy dissipation under breaking waves. Journal of Physical Oceanography 26: 792--807. Thorpe SA (1995) Dynamical processes of transfer at the sea surface. Progress in Oceanography 35: 315--352. Thorpe SA (2005) The Turbulent Ocean. Cambridge, UK: Cambridge University Press.


BUBBLES D. K. Woolf, Southampton Oceanography Centre, Southampton, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 352–357, & 2001, Elsevier Ltd.

Introduction Air–sea interaction does not solely occur directly across the sea surface, but also occurs across the surface of bubbles suspended in the upper ocean, and across the surface of droplets in the lower atmosphere. This article describes the role of bubbles in air–sea interaction. There are three quite different types of bubbles in the oceans that can be distinguished by their sources (atmospheric, benthic, and cavitation). Benthic sources of bubbles include vents and seeps and consist of gases escaping from the seafloor. Common gases from benthic sources include methane and carbondioxide. Cavitation is largely an unintentional byproduct of man’s activities; typically occurring in the wake of ship propellors. It consists of the rapid growth and then collapse of small bubbles composed almost entirely of water vapor. Cavitation may be thought of as localized boiling, where the pressure of the water falls briefly below the local vapor pressure. Cavitation is important in ocean engineering due to the damage inflicted on man-made structures by collapsing bubbles. Both cavitation bubbles and bubbles rising from the seafloor are encountered in the upper ocean, but are peripheral to air–sea interaction. Atmospheric sources of bubbles are a product of air–sea interaction and, once generated, the bubbles are themselves a peculiar feature of air–sea interaction. The major sources of bubbles in the upper ocean are the entrapment of air within the flow associated with breaking waves and with rain impacting on the sea surface. Once air is entrapped at the sea surface, there is a rapid development stage resulting in a cloud of bubbles. Some bubbles will be several millimeters in diameter, but the majority will be o0.1 mm in size. Each bubble is buoyant and will tend to rise towards the sea surface, but the upper ocean is highly turbulent and bubbles may be dispersed to depths of several meters. Small particles and dissolved organic compounds very often collect on the surface of a bubble while it is submerged. Gas will also be slowly exchanged across the surface of bubbles, resulting in

a continual evolution of the size and composition of each bubble. The additional pressure at depth in the ocean will compress bubbles and will tend to force the enclosed gases into solution. Some bubbles will be forced entirely into solution, but generally the majority of the bubbles will eventually surface carrying their coating and altered contents. At the surface, a bubble will burst, generating droplets that form most of the sea salt aerosol suspended in the lower marine atmosphere. The measurement of bubbles in the upper ocean depends largely on their acoustical and optical properties. At the same time, the effect of bubbles on ocean acoustics has long been a major motivation for bubble studies. The generation of noise at bubble inception may be exploited. For example, acoustic measurements of rainfall depend on bubble phenomena. Fully formed bubble clouds attenuate and scatter both sound and light in the upper ocean. Climatologies of the distribution of bubbles in the upper ocean are based on both acoustical and optical measurements of bubbles. The global distribution of bubbles reflects the dominance of wave breaking as a source of bubbles, and the high sensitivity of wave breaking to wind speed. Bubbles are an important component of global geochemical cycling through their transport of material in the upper ocean and surface microlayer (see Surface Films), and especially their role in the air–sea exchange of gases and particles.

Sources of Bubbles As described already, bubbles may originate in a variety of ways, but this section will concentrate on the major natural processes of air bubble formation. The atmosphere is clearly a potential source of air bubbles, and generation involves the ‘pinching off’ of part of the atmosphere, or the ‘condensation’ of gases dissolved from the atmosphere within a body of water. Generation of bubbles within the body of water, when the surface water is sufficiently supersaturated with air, is similar to ‘vapor’ cavitation, but involves the major constituents of the atmosphere (nitrogen, oxygen, etc.) rather than water vapor alone. In the absence of hydrodynamic pressure effects associated with flow, the radial pressure into a cavity, Pb, is the sum of the atmospheric pressure, Pa, the hydrostatic pressure at a depth, z, and a component associated with the surface tension, g, and the curvature of the


439

440

BUBBLES

cavity (or radius ‘r’): Pb ¼ Pa þ rgz þ 2g=r For a bubble to grow, the pressure within a bubble (equal to the sum of partial pressures of the gases diffusing into the bubble), must exceed atmospheric pressure by a margin that increases both with water depth and the curvature of the cavity. A sufficiently large initial cavity is necessary for inception. The explosive dynamics of ‘true’ cavitation are associated with the rapid transport of water vapor across the surface of the cavity. However, the conditions for water vapor cavitation can only be achieved at normal temperatures where pressure is very low. A sufficient pressure anomaly may occur in an intense acoustic pulse, or in the wake of a fast moving solid object, but is not a common natural phenomenon. The conditions for growth of a bubble by diffusion of atmospheric gases are possible within the normal range of natural variability. For gases other than water vapor, molecular transport of the dissolved gases near the surface of the bubble is sufficiently slow that a virtual equilibrium between the internal and external pressures on the bubble must exist. We might observe bubble generation at home within a bucket of water, or in a soda bottle, where warming induces supersaturation (the solubility of most gases decreases with increasing temperature) and defects in the container walls provide the initial cavity. Warming, mixing, or bubble injection may occasionally force supersaturations of several percent at sea, in which case growth of bubbles on natural particles and microbubbles may release the excess pressure. Entrapment of air at the sea surface is more common than inception within the body of the water. Most of us are familiar with plumes of bubbles generated by paddling and by boats, but the entrapment of air in the absence of a solid boundary is less intuitive. In general, air is rarely entrapped by enclosure of a large air volume, but is usually drawn into the interior (‘entrained’) where there is intense and convergent flow of water at the sea surface. Sufficiently energetic convergence occurs where precipitation impacts on the sea surface, and where waves break. Bubble formation is associated with all common forms of precipitation (rain, hail, and snow), but the details of bubble formation are highly specific to the details of the precipitation. In particular, bubble formation by rain is known to be sensitive to the size, impact velocity, and incidence angle of the rain drops. Large drops, exceeding 2.2 mm in diameter, entrain most air. In heavy tropical downfalls, the

volume of air entrained can be fairly significant (B 106 m3 m2 s1), although much lower than rates associated with wave breaking in high winds. Bubbles up to 1.8 mm in radius are entrained by large rain drops, but smaller drops (0.8–1.1 mm in diameter) generate bubbles of only 0.2 mm in radius. When waves break at the seashore, the large ‘dominant’ waves dissipate their energy partly in entraining and submerging quite large volumes of air. On the open ocean, some of the largest and longest waves break, but wave breaking also occurs at much smaller scales. Some very small breaking events may be too weak to entrain air; however, small but numerous breaking events entraining small volumes of air occur on steep waves as short as 0.3 m in wavelength. The energy dissipated in wave breaking is derived from wind forcing of surface waves, and the amount of wave breaking and air entrainment is very sensitive to wind speed. The stage of development of the wave field also has some influence on air entrainment – the size of the largest breaking event is limited to the largest wave that has developed. An important feature of bubble generation at the sea surface is that a myriad of very small (o0.1 mm radius) bubbles is produced. Very large cavities several millimeters in diameter are likely to be torn apart by large shear forces at the sea surface, but it is difficult to explain how bubbles of o1 mm might be fragmented. Also, the same processes in fresh water (e.g. a lake or a waterfall) do not produce many small bubbles. The explanation can be found in the influence of dissolved salts on surface forces. In sea water, a surface deformation will tend to grow more and more contorted, so that when a large bubble is fragmented it will often shatter into numerous much smaller bubbles. The same factors will usually prevent the coalescence of bubbles in sea water.

Dispersion and Development Bubbles entrained by a breaking wave may be carried rapidly to a depth of the order of the height of the breaking wave by its energetic turbulent plume. For some wave breaking and other forms of bubble production the initial injection will be much shallower (B1–100 mm). Most of the bubbles are very small, but the majority of the volume of air is comprised of fairly large (B1 mm) bubbles entrained by breaking waves. Most of these larger bubbles will soon rise to the surface (typically in B1 s) in a highly dynamic plume close behind the breaking wave. The less buoyant, smaller bubbles are generally carried to a greater depth and are easily dispersed by mixing processes in the upper ocean.


BUBBLES

Bubbles are mixed into the ocean by small-scale turbulence associated with the ‘wind-driven upper ocean boundary layer’, but also by relatively large and coherent turbulent structures, especially Langmuir circulation (see Langmuir Circulation and Instability). Langmuir circulation comprises sets of paired vortices (cells) aligned to the wind. Bubbles will be drawn to the downwelling portions of the Langmuir cells, producing lines of enhanced bubble concentration, parallel to the wind. Langmuir cells can be up to tens of meters deep and wide, and downwelling speeds may exceed 0.1 m s1. In principle, even quite large bubbles may be forced downwards, but generally bubbles of only B20 mm

radius are most common at depths of Z1 m. Concentrations fall off rapidly with increasing radius, at radii exceeding the modal radius. The development of a bubble cloud does not solely concern the movement of bubbles, but also concerns the development of each and every bubble. Material will be transferred between the bubble and the surrounding water as a result of the flow of water around the bubble (largely induced by the buoyant rise of bubbles relative to their surroundings) and molecular diffusion close to the surface of the bubble. This transport plays a large part in the role of bubbles in geochemical cycling, which is illustrated schematically in Figure 1. The transport of both

Aerosol production Deposition

Jet drops

Film drops Atmosphere

Bursting bubble

Microlayer Ocean

Bubble scavenging

441

Gas exchange

Figure 1 A schematic illustration of the role of bubbles in geochemical cycling.


442

BUBBLES

volatile (i.e. gases) and nonvolatile substances is of interest. Nonvolatile substances will not penetrate the bubble itself, but may be transported between the surface of the bubble and the surrounding water. Many substances are ‘surface-active’, that is, they tend to stick to the surface and alter the dynamic properties of the surface. Some substances will already be adsorbed on the surface at the point of formation at the sea surface. During the lifetime of a bubble, further material (both dissolved and small particles) will accumulate on the surface of a bubble. One consequence of the ‘bubble scavenging’ process is the cycling of surface-active substances. Also, the surface-active material will alter the dynamic properties of the bubble, critically affecting the rise velocity of the bubble and transport across the surface of the bubble. A pure water surface is ‘mobile’, but it may be immobilized by surface-active material. The flow near a mobile (or free) surface and a rigid surface is quite different. Generally, a ‘dirty’ bubble with a contaminated, rigid surface will rise more slowly and will exchange gas at a much slower rate compared with a ‘clean’ bubble of the same size. The surface of small bubbles is immobilized by only a small amount of contamination, and bubbles o100 mm radius are likely to behave as dirty bubbles for most or all of their life. Larger bubbles will also be contaminated, but their dynamic behavior may remain close to that of a ‘clean’ bubble for several seconds (depending on bubble radius and the contamination level of the water). The transfer of gases across the surface of bubbles is important to the evolution of each bubble, and to the atmosphere–ocean transport of gases. Gases will diffuse across the surface of a bubble. The net transport of each gas across the surface of a single bubble depends on its concentration in the two media and the mechanics of transport: bubblewater flux ¼ j4pr2 ½Cw Spb As explained in the previous section, the gases within a bubble are compressed so that the pressure of gases in the bubble generally exceeds those in the atmosphere. This excess leads to a tendency for bubbles to force supersaturation of gases in the upper ocean. Many bubbles may be forced entirely into solution (possibly leaving a fragment enclosed in a shell of organics and small particles – a microbubble). The total (integral) effect of bubble clouds on air–sea gas exchange can be described by the following formula (see Air–Sea Gas Exchange): airsea flux ¼ KT ½Cw Spa ð1 þ DÞ

(per unit area of sea surface) ¼ Kb ½ð1 þ dÞCa =H Cw þ Ko ½Ca =H Cw The effect of bubbles on air–sea exchange is described by two coefficients: the contribution to the transfer coefficient, Kb, and a ‘saturation anomaly’, D. Both of these coefficients depend greatly on the solubility of the gas and the bubble statistics. For relatively soluble gases, such as carbon dioxide, the saturation anomaly due to bubble injection is generally negligible, but for less soluble gases, including oxygen the anomaly is usually significant, particularly at high wind speeds. The contribution to the transfer coefficient is again greater for less soluble gases, but is likely to be significant for most gases, at least for high wind speeds. We have focused on unstable bubbles that will either surface or dissolve within a few minutes of their creation. When a bubble totally dissolves it may leave a conglomeration of the particles and the organic material it accumulated. Some of the bubbles may not entirely dissolve, but may be stabilized at a radius of a few micrometers by their collapsed coating. (The mechanism of stabilization is rather mysterious, external pressures will be high and the coating can not entirely prevent the diffusion of gas, therefore total collapse must be resisted by the structural integrity of the coating – perhaps like a traditional stone wall.) Stable microbubbles might also be generated by a biological mechanism. Microbubble populations are denser in coastal waters where biological productivity and organic loading are generally higher. Microbubbles influence the acoustic properties of natural waters and are a common nucleus for cavitation.

Surfacing and Bursting Many small bubbles dissolve in the upper ocean, but generally the majority of the bubbles (and almost all the large bubbles) eventually surface. Phenomena that occur when a bubble surfaces are again significant to geochemical cycling (Figure 1). The release of gas from a bubble to the atmosphere completes the process of air–sea gas exchange mediated by the bubble. The approach of a bubble, or more especially a plume of bubbles, can disrupt the surface microlayer, enhancing turbulent transport directly across the sea surface. The bubble carries material to the sea surface accumulated by scavenging within the upper ocean. Most important are the energetic processes that occur when a bubble bursts on the sea surface. Bubble bursting is responsible for ejecting droplets into the atmosphere, creating the sea salt aerosol.


BUBBLES

Droplets can also be torn directly from wave crests, but bubbles generate almost all of the very small droplets that are easily suspended in the lower atmosphere and that will be dispersed over large distances. When a bubble surfaces its upper surface will project beyond the sea surface. This ‘film cap’ will drain and shatter. The shattering of the film cap produces ‘film drops’. In some cases, the film cap can shatter into many remarkably small (o1 mm) droplets, while in other cases a few large B10 mm radius droplets will be produced. The open cavity left after the film cap shatters will collapse inwards, leading to the upward ejection of a ‘Worthington jet’. This jet will pinch off into a few ‘jet drops’. The drop radii will typically be one-tenth of the radius of the parent bubble, producing drops from B2 mm to tenths of a millimeter in radius from a typical bubble population. The droplets generated by bursting bubbles will be enriched by material brought to the sea surface by the bubble and drawn from the sea surface. The sea salt aerosol will include organic material, metals, viruses, and bacteria.

Acoustical and Optical Properties Our knowledge of bubble distributions in the upper ocean is based on acoustical and optical measurements. Bubbles also have a significant impact on the acoustical and optical properties of the upper ocean. The acoustic properties of bubbles have attracted a great deal of attention. The generation of bubbles, both by breaking waves and rain, is an important source of noise in the upper ocean. Bubbles also absorb and scatter sound. The scattering of sound by an individual bubble is frequency-dependent with three primary regimes: close to, above, and below the ‘breathing frequency’ of the bubble. The breathing frequency of a bubble is the natural frequency at which a bubble will oscillate radially (‘breathe’) and is determined by its radius, surface tension, and the external pressure. The breathing frequency is inversely related to bubble radius, and in the upper ocean, bubbles of different radii will respond in resonance to acoustic frequencies from 10 kHz to a few hundred kHz. Scattering cross-sections close to resonance are very high. When the acoustic frequency is much higher than the breathing frequency of the bubble, the scattering by the bubble is related simply to its physical size (‘geometric scattering’). At low acoustic frequencies the acoustic crosssection of an individual bubble is much lower than its geometric cross-section (Rayleigh scattering). The scattering by a bubble is equal in every direction (isotropic) at most practical frequencies, but

443

becomes more anisotropic at very low frequencies. At low acoustic frequencies (o10 kHz), scattering is largely a communal response of clouds rather than of individual bubbles. Many measurements of bubbles have taken advantage of the resonant response of bubbles to sound. In particular, measurements at a number of acoustic frequencies can be inverted to calculate the size distribution of bubbles. A pair of transmitting and receiving ‘transducers’ can measure backscatter remotely along a profile. This technique has been used to infer the concentration and size of bubbles as a function of depth. The very high scattering by the concentrated plumes near breaking waves defy remote measurement. Instead bubbles near the surface may be studied by measuring absorption or scattering along a short path length. Other techniques include applying the influence of air void on the conductivity of the water, and optical measurements. Casual observation of the milky water marking a developing bubble cloud is enough to understand that bubbles in the upper ocean can alter the optical properties (e.g. color and brightness) of the sea, but among the numerous and complicated influences on ocean optics, bubbles have received relatively little attention. Bubble populations have been measured photographically, but for the sparse populations a meter or so beneath the sea surface this method is tedious if ultimately effective. Video footage of wave breaking and the early development of bubble plumes can be used to understand the many processes involved.

Summary of Bubble Distribution Measurements of bubbles in the ocean are still fairly sparse, and the relationship of wave breaking and bubble injection to environmental conditions is only partly understood, but we can at least summarize the general relationship of unstable bubble populations to wind forcing. Away from the immediate plume of a breaking wave, the mean concentration of bubbles of radius, r, at a depth z, typically follows a distribution of the form, Npr4 expðz=LÞ for radii as small as 30 mm, but there is a maximum in N typically at 25 mm radius. A typical attenuation depth, L, is 1 m. Some studies have suggested only a weak, approximately linear relationship between the attenuation depth and wind speed, but recent extensive studies imply attenuation depths proportional to the square of wind speed. There are fewer measurements of bubbles in the upper ocean


444

BUBBLES

within half a meter of the sea surface, but it is clear that concentrations are much higher near a breaking wave, and larger bubbles are far more common. The injection rate of bubbles is expected to increase with the third or fourth power of the wind speed. As vertical dispersion of the bubbles (and the attenuation depth) also increase with wind speed, the concentration of bubbles below the sea surface is extremely sensitive to wind speed. Air–sea gas exchange, scavenging, and other geochemical transport processes associated with bubbles will share this sensitivity to wind speed, suggesting that a large fraction of activity may occur in fairly rare storm conditions.

See also Acoustic Noise. Air–Sea Gas Exchange. Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, NonMethane Hydrocarbons, Organo-Halogens. Air–Sea Transfer: N2O, NO, CH4, CO. Breaking Waves and Near-Surface Turbulence. Carbon Dioxide (CO2) Cycle. Evaporation and Humidity. Heat and Momentum Fluxes at the Sea Surface. Langmuir

Circulation and Instability. Photochemical Processes. Sonar Systems. Surface Films. ThreeDimensional (3D) Turbulence. Upper Ocean Mixing Processes. Wave Generation by Wind. Whitecaps and Foam.

Further Reading Blanchard DC (1983) The production, distribution, and bacterial enrichment of the sea-salt aerosol. In: Liss PS and Slinn WGN (eds.) The Air–Sea Exchange of Gases and Particles, pp. 407--454. Dordrecht: Kluwer. Leighton TG (1994) The Acoustic Bubble. San Diego: Academic Press. Medwin H and Clay CS (1998) Fundamentals of Acoustical Oceanography. San Diego: Academic Press. Monahan EC (1986) The ocean as a source for atmospheric particles. In: Buat-Meńard P (ed.) The Role of Air–Sea Exchange in Geochemical Cycling, pp. 129--163. Dordrecht: Kluwer. Woolf DK (1997) Bubbles and their role in gas exchange. In: Liss PS and Duce RA (eds.) The Sea Surface and Global Change, pp. 173--205. Cambridge: Cambridge University Press.


CALCIUM CARBONATES L. C. Peterson, University of Miami, Miami, FL, USA

Carbonate Producers

Copyright & 2001 Elsevier Ltd.

The most important carbonate producers in the open ocean are planktonic coccolithophorids and foraminifera, unicellular phytoplankton and zooplankton respectively, which inhabit the upper few hundred meters of the water column (Figure 1). Coccolithophorids are the dominant carbonateprecipitating organisms on Earth. During part of their life cycle, they produce a skeletal structure (the coccosphere) consisting of loosely interlocking plates, often button-like in appearance, known as coccoliths. Deep-sea carbonates generally contain only the individual coccoliths, as the intact coccospheres are rarely preserved. Foraminifera produce a calcareous shell, or ‘test’, a few hundred microns in size that sinks after death or reproduction to the sea floor. Both coccolithophorids and the planktonic foraminifera construct their skeletal elements out of the mineral calcite, the more stable polymorph of CaCO3. Calcareous sediments dominated by one or the other component are termed coccolith oozes or foraminiferal oozes, although in reality most carbonate-rich sediments are a mixture of both. Coccolithophorids made their first appearance in the geological record in the earliest Jurassic, while planktonic foraminifers evolved somewhat later in the middle Jurassic. The appearance of these two dominant pelagic carbonate producers, and their rapid diversification in the Cretaceous, would have had major effects upon the carbonate geochemistry of the oceans. Before this, most carbonate was deposited in shallow seas, accounting for the high proportion of limestones among older rocks on the continents. Since the Mesozoic, deep-ocean basins have become enormous sinks for carbonate deposition. Smaller contributions to the deep-sea carbonate budget come from a variety of other sources. Pteropods, free-swimming pelagic gastropods, construct a relatively large (several millimeters) but delicate shell out of the metastable form of CaCO3 known as aragonite. However, while pteropods can be unusually abundant in certain environments, the increased solubility of aragonite leads to very restricted preservation of the shells and pteropod oozes are relatively rare in the ocean. In the vicinity of shallow, tropical carbonate platforms such as the Bahamas or Seychelles Bank, shedding of aragonitic bank-top sediments derived from algal and coral production can lead to aragonite-rich ‘periplatform oozes’ in deep waters around the perimeters of the platform.


Introduction The ocean receives a continual input of calcium from riverine and groundwater sources and from the hydrothermal alteration of oceanic crust at midocean ridge spreading centers. Balancing this input is the biological precipitation of calcium carbonate (CaCO3) by shell-and skeleton-building organisms in both shallow marine and open-ocean environments. In the deep sea, the primary contributors to the carbonate budget of open-ocean sediments are the skeletal remains of calcareous plankton that have settled down from the surface after death. Seafloor sediments consisting of more than 30% by weight calcium carbonate are traditionally referred to as calcareous or carbonate ooze; such oozes accumulate at the rate of 1–4 cm per 1000 years and cover roughly half of the ocean bottom. Carbonate oozes are the most widespread biogenous sediments in the ocean. While the biological production of calcium carbonate in oversaturated surface waters determines the input of carbonate to the deep sea, it is the dissolution of carbonate in undersaturated deep waters that has the dominant control on calcium carbonate accumulation in the open ocean. Since carbonate production rates in the surface ocean today greatly exceed the rate of supply of calcium, this ‘compensation’ through dissolution must occur in order to keep the system in steady-state. Increased dissolution at depth is largely a function of the effect of increasing hydrostatic pressure on the solubility of carbonate. However, superimposed on this bathymetric effect are regional preservation patterns related to differences in carbonate input and the carbonate chemistry of deep water masses. Carbonate oozes in the deep sea serve as a major reservoir of calcium and carbon dioxide on the Earth’s surface. Their spatial and temporal accumulation patterns in the marine stratigraphic record are thus a primary source of data about the carbonate chemistry and circulation of past oceans, as well as of the global geochemical cycle of CO2.


445

446

CALCIUM CARBONATES

500 μm

Magn 50x (A)

5 μm

Magn 6000x (B)

Figure 1 Carbonate oozes in the deep sea are dominated by the skeletal remains of (A) planktonic foraminifera ( 50 magnification) and (B) coccolithophorids ( 6000 magnification). Specimens shown here were isolated from a Caribbean sediment core.

In general, contributions from bottom-dwelling organisms (e.g. benthonic foraminifera, ostracods, micromollusks) are negligible in deep-sea sediments.

Carbonate Distribution and Dissolution The distribution of carbonate sediments in the ocean basins is far from uniform (Figure 2). If it were

possible to drain away all of the ocean’s water, carbonate oozes would be found draped like snow over the topographic highs of the seafloor and to be largely absent in the deep basins. The lack of carbonate-rich sediments in the deepest parts of the world’s oceans has been recognized since the earliest investigations. Although surface productivity and dilution by noncarbonate sediment sources can locally influence the concentration of carbonate in


CALCIUM CARBONATES

447

80 20 40 60 20

60

60

60 60

20

20

20

80

20

80

20 20

Figure 2 Global distribution of calcium carbonate (weight-% CaCO3) in surface sediments of the ocean. Data compilation from Archer (1996); reproduced with permission from the American Geophysical Union.

deep-sea sediments, the clear-cut relationship between calcium carbonate content and water depth indicates that carbonate dissolution plays the major role in governing carbonate distribution patterns. To a first approximation, the dissolution of carbonate on the seafloor is a function of the corrosiveness or saturation state of the overlying bottom waters. The amount of calcium carbonate that will dissolve in sea water if thermodynamic equilibrium is reached is governed by the following reaction: CaCO3 ðsÞ2Ca2þ ðaqÞ þ CO3 2 ðaqÞ At equilibrium, the rate of carbonate dissolution is equal to the rate of its precipitation and the sea water is said to be saturated with respect to the carbonate phase. In the deep sea, the degree of calcium carbonate saturation (D) can be expressed as: 2þ Ca seawater CO3 2 seawater D ¼ 2þ Ca saturation ½CO3 2 saturation where [Ca2þ]seawater and [CO3 2 ]seawater are the in situ concentrations in the water mass of interest and [Ca2þ]saturation and [CO3 2 ]saturation are the concentrations of these ions at equilibrium, or saturation, at the same conditions of pressure and temperature. Since shell formation and dissolution cause the concentration of [Ca2þ] to vary by less than 1% in the ocean, the degree of calcium carbonate saturation (D) can be simplified and expressed in terms of the concentration of the carbonate ions only: CO3 2 seawater D¼ ½CO3 2 saturation

D is thus a measure of the degree to which a seawater sample is saturated with respect to calcite or aragonite, and so provides a measure of the strength of the driving force for dissolution. Values of D41 indicate oversaturation while values of Do1 indicate undersaturation and a tendency for calcium carbonate to dissolve. Since the saturation carbonate ion concentration increases with increasing pressure and decreasing temperature, calcium carbonate is more soluble in the deep sea than at the surface. At the depth in the water column where D ¼ 1, the transition from oversaturated to undersaturated conditions is reached. This depth is known as the saturation horizon (Figure 3). Aragonite is always more soluble than calcite, and its respective saturation horizon is shallower, because the saturation carbonate ion concentration for aragonite is always higher for the same conditions of pressure and temperature. Observations from studies of surface sediments have allowed definition of regionally varying levels in the ocean at which pronounced changes in the presence or preservation of calcium carbonate result from the depth-dependent increase of dissolution on the seafloor. The first such level to be identified was simply the depth boundary in the ocean separating carbonate-rich sediments above from carbonate-free sediments below. This level is termed the calcite (or carbonate) compensation depth (CCD) and represents the depth at which the rate of carbonate dissolution on the seafloor exactly balances the rate of carbonate supply from the overlying surface waters. Because the supply and dissolution rates of carbonate differ from place to place in the ocean, the depth of the CCD is variable. In the Pacific, the CCD is typically found at depths between about 3500 and 4500 m. In the North Atlantic and parts of the South Atlantic, it is found


448

CALCIUM CARBONATES

2_

Δ CO 3 _ 30

_ 20

_ 10

0

(Calcite)

10

20

30

40

50

1000 GEOSECS data _1 (μ moles kg )

Water depth (m)

2000

3000

4000 Atlantic 5000

6000

Pacific Undersaturated

Oversaturated

Figure 3 Bathymetric profiles of calcium carbonate (calcite) saturation for hydrographic stations in the Atlantic and Pacific Oceans (data from Takahashi et al. 1980). Carbonate saturation here is expressed as DCO3 2 , defined as the difference between the in situ carbonate ion concentration and the saturation carbonate ion concentration at each depth DCO3 2 ¼ [CO3 2 ]seawater [CO3 2 ]saturation). The saturation horizon corresponds to the transition from waters oversaturated to waters undersaturated with respect to calcite (D CO3 2 ¼ 0). This level is deeper in the Atlantic than in the Pacific because Pacific waters are CO2-enriched and [CO3 2 ]-depleted as a result of thermohaline circulation patterns and their longer isolation from the surface. The Atlantic data are from GEOSECS Station 59 (301120 S, 391180 W); Pacific data come from GEOSECS Station 235 (161450 N,1611230 W).

closer to a depth of about 5000 m. Close to continental margins the CCD tends to shoal, although much of this apparent rise can be attributed to carbonate dilution by terrigenous input from the continents. Rarely does carbonate ooze accumulate on seafloor that is deeper than about 5 km. In practice, the CCD is identified by the depth transition from carbonate ooze to red clay or siliceous ooze that effectively defines the upper limit of the zone of no net CaCO3 accumulation on the seafloor. Given the practical difficulty (e.g. analytical precision, redeposition) of determining the depth level at which the carbonate content of sediment goes to zero, some investigators choose instead to recognize a carbonate critical depth (CCrD), defined as the depth level at which carbonate contents drop to o10% of the bulk sediment composition. The CCrD lies systematically and only slightly shallower than the CCD. A similar boundary to the CCD can be recognized marking the lower depth limit of aragonite-bearing sediment in the ocean, the aragonite compensation depth or ACD. Because of the greater solubility of aragonite as compared with calcite, the ACD is always much shallower than the CCD. Above the CCD, the level at which significant dissolution of carbonate first becomes apparent is called the lysocline. As originally defined, the term

lysocline was used to describe the depth level where a pronounced decrease in the preservation of foraminiferal assemblages is observed. It thus marks a facies boundary separating well-preserved from poorly preserved assemblages on the seafloor. This level is now more specifically referred to as the foraminiferal lysocline to differentiate it from the coccolith lysocline and pteropod lysocline, which may differ in depth because of varying resistance to dissolution or differences in solubility (in the case of the aragonitic pteropods). In addition, it is customary to recognize a sedimentary or carbonate lysocline as the depth at which a noticeable decrease in the carbonate content of the sediment begins to occur. In theory, the lysocline records the sedimentary expression of the saturation horizon, that is the depth-dependent transition from waters oversaturated to waters undersaturated with respect to carbonate solubility (Figure 4). The lysocline thus marks the top of a depth zone, bounded at the bottom by the CCD, over which the bulk of carbonate dissolution in the ocean is expected to occur in response to saturation state-driven chemistry. The thickness of this sublysocline zone, as indicated by the vertical separation between the lysocline and CCD, is variable and is governed by the rate of carbonate supply, the actual dissolution gradient, and


CALCIUM CARBONATES

449

2_

Δ CO3 (Calcite) _ 20

Water depth (m)

2000

_ 10

0

Percent 10

20

0

20

40

% Fragments

GEOSECS Stn. 441 Indian Ocean

60

80

100

% CaCO3

3000

Lysocline

Saturation horizon 4000

5000 CCD

Figure 4 Comparison of carbonate saturation profile for the eastern equatorial Indian Ocean with measurements of foraminiferal fragmentation and carbonate content (weight-%) from depth-distributed modern sediment samples in this region. The saturation horizon with respect to calcite (DCO3 2 ¼ 0) occurs locally in the water column at a depth of 3800 m. This level corresponds with both the foraminiferal lysocline and carbonate lysocline as recognized in the sediments. The carbonate compensation depth (CCD) in this region is found at a depth of approximately 5000 m. Increased foraminiferal fragmentation and decreases in sedimentary carbonate content are the result of dissolution and carbonate loss below the lysocline. Carbonate saturation data are from GEOSECS Station 441 (5120 S, 911470 E; Takahashi et al. 1980); modern sediment data are from Peterson and Prell (1985).

potentially by noncarbonate dilution in certain regions of the ocean. While the term lysocline was originally used to define a preservational boundary, it has also been used in a fundamentally different sense to denote the depth at which dissolution rates of carbonate on the seafloor greatly accelerate. Whether these levels may or may not coincide, and the nature of their relationship to the saturation horizon or ‘chemical lysocline’, has been the subject of much discussion and debate. One of the reasons for uncertainty in this regard is the fact that both the carbonate content (%) of a sediment sample and the preservation of the calcareous microfossil assemblages there in can be surprisingly poor indicators of the extent to which dissolution has occurred. For example, the loss of carbonate (L) from sediment, expressed as a weight percentage of the total sediment, is given by: L ¼ 100ð1 Ro =RÞ where Ro and R are the initial and final values of the noncarbonate (or residual) material. Thus, for a

sample initially containing 95% carbonate and a Ro value of 5%, 50% of the carbonate in the sample must be dissolved in order to double the noncarbonate fraction and reduce the carbonate content to 90%. Since the carbonate fraction of the pelagic rain in the open ocean often approaches 95%, this inherent insensitivity means that significant loss of carbonate can occur before detectable changes in the carbonate content are observed. As a consequence, the carbonate lysocline, traditionally defined as the level where the carbonate content of sediments begins to sharply decrease with water depth, may lie deeper than the depth at which significant loss of carbonate to dissolution actually begins to occur. Dissolution leads to an increase in surface area during the etching of carbonate skeletal material. Etching produces roughness and widens pores, leading to weakening and ultimately to breakage. Because of their larger size, planktonic foraminifera have usually been the subject of dissolution studies that focus on the preservation state of the microfossils themselves. Planktonic foraminifera have a wide range of morphologic characteristics that


450

CALCIUM CARBONATES

enhance their abilities to remain suspended in the upper water column while alive. These same characteristics largely dictate their resistance to dissolution after death. Taxa living in warm, tropical surface waters, where density is generally low, tend to be open-structured with thin shells and porous walls. Taxa that live deeper in cooler, denser subsurface waters, or in colder surface waters at high latitudes, tend to be more heavily calcified with thicker shells and small or closed up pores. On the seafloor, the thinshelled, more fragile species tend to dissolve more readily than the robust taxa. In effect, this means that individual species each have their own ‘lysocline’, which can be offset shallower or deeper from the foraminiferal lysocline determined from the total assemblage. There are additional consequences of this selective preservation of taxa that must be considered in paleo-oceanographic or paleoclimatic studies. For example, the selective preservation of more heavily calcified taxa tends to impart a generally ‘cooler’ appearance to the overall microfossil population and can bias attempts to derive paleotemperature information from seafloor assemblages, as well as other population properties such as diversity. For carbonate particles produced in the upper ocean, settling rates play an important role in their distribution and preservation. Smaller planktonic foraminifers settle at about 150–250 m d1, while larger (4250 mm) foraminifers may settle as much as 2000 m d1. These rates are rapid enough that little dissolution is thought to occur in the water column. Solitary coccoliths, on the other hand, sink at rates of 0.3 to B10 m d1, slow enough that dissolution within the water column should theoretically prevent their ever reaching the ocean bottom. However, sediment trap studies have shown that transport by fecal pellets is the dominant process by which small phytoplankton skeletons are transferred to the seafloor. Protection offered by the organic fecal pellet covering may also protect the coccoliths after deposition and account for the fact that the coccolith lysocline is generally observed to lie somewhat deeper than the foraminiferal lysocline. While the seafloor depths of the lysocline and CCD can be readily identified from sedimentary criteria, this information is of limited use without realistic knowledge of the rates at which calcium carbonate is lost from the sediments to dissolution. In practice, it is much easier to determine carbonate accumulation in the deep sea than it is to estimate carbonate loss. Yet the latter information is clearly needed in order to close sediment budgets and to reconstruct changes in the carbonate system. Carbonate-rich sediments deposited above the saturation horizon should experience little in the way

of saturation-driven dissolution because they lie in contact with waters oversaturated with respect to calcite. Nevertheless, evidence for significant supralysoclinal dissolution has been found in a number of studies. Much of this dissolution at shallower water depths is thought to be driven by chemical reactions associated with the degradation of organic carbon in the sediments. Organic carbon arriving at the seafloor is generally respired as CO2 or remineralized to other organic compounds by benthic organisms. The metabolic CO2 generated by organisms that live within the sediment can contribute to the dissolution of calcite even above the lysocline by increasing the chemical corrosivity of the pore waters. Studies of organic matter diagenesis in deep-sea sediments suggest that rates of supralysoclinal dissolution vary greatly with location, ranging from minimal loss to 440% calcite loss by weight. Temporal and spatial changes in the rain rate of organic carbon relative to carbonate can affect this process. Whether above or below the lysocline, carbonate dissolution is mostly confined to the bioturbated surface sediment layer (typically r10 cm in the deep sea). As carbonate is depleted from this bioturbated layer, older ‘relict’ carbonate is entrained from the sediments below. This results in ‘chemical erosion’ and can produce substantial hiatuses or gaps in the record. Dissolution, and hence erosion, eventually stops when nonreactive materials fill up the mixed layer and isolate the underlying sediment from the overlying water. Many clay layers interbedded within carbonate-rich sequences are likely produced by this mechanism; the resulting lithologic contrasts often show up as subsurface seismic horizons which can be traced for long distances and tell a story of changing dissolution gradients and carbonate chemistry in the past.

Basin-to-Basin Fractionation in the Modern Ocean Superimposed on the general depth-dependent decrease of carbonate accumulation observed everywhere in the deep sea are preservation patterns that differ between the major ocean basins. Today, carbonate-rich sediments tend to accumulate in the Atlantic Ocean, while more carbonate-poor sediments are generally found at comparable water depths in the Indian and Pacific Oceans. This modern pattern is largely the product of the ocean’s thermohaline circulation and has been termed ‘basin-to-basin fractionation’. In the Atlantic, deep and bottom waters tend to be produced at high latitudes because cold temperatures and high sea surface salinities lead to the formation of dense water


CALCIUM CARBONATES

masses that sink and spread at depth. These young, relatively well oxygenated and [CO32]-enriched waters tend to depress the depth of the saturation horizon and allow carbonate to accumulate over much of the Atlantic basin, as manifested by a deep lysocline and CCD. In contrast, neither the Indian nor Pacific Oceans today experience surface conditions that allow deep or bottom waters to form; water masses at depth in these basins largely originate in the Atlantic sector as part of what is sometimes described as the ocean’s conveyor belt circulation, with a general upwelling of waters from depth balancing the formation and sinking of deep waters in the Atlantic source areas. Since deep and bottom waters in the Indian and Pacific Oceans are further removed from their modern source areas in the Atlantic, they tend to be CO2-enriched and [CO32]depleted because of their greater age and the cumulative effects of organic matter remineralization along their flow path. In particular, the in situ decrease in [CO32] concentration leads to an increase in undersaturation of the water masses and a progressive shoaling of the saturation horizon (Figure 3). Thus, Indian and Pacific deep waters are generally more corrosive to the biogenic carbonate phases than Atlantic waters at comparable depth, the lysocline and CCD are shallower, and a smaller area of the seafloor experiences conditions suitable for carbonate preservation and accumulation. This pronounced modern pattern of basin-to-basin fractionation is illustrated by the fact that roughly 65% of the present Atlantic seafloor is covered by carbonate ooze, while only 54% of the Indian Ocean floor and 36% of the Pacific Ocean floor share that distinction. Naturally, if thermohaline circulation patterns have changed in the past, then carbonate preservation and accumulation patterns will change accordingly. The mapping and reconstruction of such trends has emerged as a powerful paleoceanographic tool.

Temporal Changes in Carbonate Accumulation and Preservation The patterns of carbonate accumulation and preservation in the deep sea contain important information about the chemistry and fertility of ancient oceans. Numerous studies have now shown that variations in the carbonate system have occurred on a variety of timescales, both within and between ocean basins. On a local or even regional scale, such variations can often be used as a correlation tool. This has come to be known as ‘preservation stratigraphy’. A number of criteria have commonly been used as indicators of the intensity of carbonate dissolution in

451

deep-sea sediments. Variations in the measured carbonate content of sediments are commonly used to correlate between cores in a region, but are difficult to interpret strictly in terms of dissolution and changing deep-water chemistry. This is because the weight percent carbonate content of a sample can also be affected by changing carbonate input (i.e. surface production) and by dilution from noncarbonate sources. More useful are indices based on some direct measure of preservation state, such as the percentage of foraminiferal fragments in a sample relative to whole shells (Figure 5). However, while clearly recording dissolution, preservationbased indices can also be affected by other factors, including ecologic changes that may introduce variable proportions of solution-susceptible species into a region over time. Because carbonate dissolution is a depth-dependent process, it is best studied where existing seafloor topography allows for sampling of sediments over a broad depth range. Given this sampling strategy, one way to circumvent the problems of using measured carbonate content and other relative dissolution indices (e.g. fragmentation) is to calculate carbonate accumulation histories for the individual sampling locations and examine depth-dependent differences in accumulation rates and patterns. To do so requires an accurate knowledge of sedimentation rates (e.g. cm per thousand years) and measurements of sediment bulk density (in g cm3), in addition to the data on carbonate content. The product of these three measures yields a mass accumulation rate for the carbonate component expressed in g per cm2 per thousand years. Differences in accumulation between depth-distributed sites can provide insights into dissolution gradients and carbonate loss. As the relative importance of calcium supply from weathering and carbonate production vary through time, the depth of the CCD must adjust to control dissolution and to keep calcium levels in balance. Studies of CCD behavior during the Cenozoic (Figure 6) have generally shown that CCD fluctuations were similar in the various ocean basins and were likely to have been driven by a global mechanism, such as a change in sea level and/or hypsometry of the ocean basins or a change in supply of calcium to the oceans. There are, however, clear ocean-to-ocean differences in this general pattern that are likely to have been the result of changes in regional productivity and the interbasinal exchange of deep and surface waters. By examining such differences, estimates of past circulation and of the relative differences in carbonate productivity in different regions can be determined from regional offsets in the depth of the CCD.


452

CALCIUM CARBONATES

1 _ 1.6

2

3

4

5

8

7

6

9

Core P6408_9

18

δ O (PDB)

_ 1.2 _ 0.8 _ 0.4

0

40 30 20 10

% Foraminiferal fragments

50

0.4

0

70

% CaCO3

60

50

40

0

50

100

200 150 Age (1000 y BP)

250

300

350

Figure 5 Measurements of foraminiferal fragmentation and calcium carbonate content (weight-%) spanning the last 350 000 years in Caribbean sediment core P6408-9. Stratigraphy and age control come from the oxygen isotope (d18O) record shown at the top of the figure; odd-numbered stages are warm, interglacial intervals and even-numbered stages indicate cold, glacial climates with greatly expanded Northern Hemisphere ice cover. Variations in the ratio of foraminiferal fragments to whole shells can be directly related to the intensity of carbonate dissolution on the seafloor. Greatly increased preservation (i.e. decreased numbers of fragments)during cold, glacial stages indicates reductions in the chemical corrosivity of deep Caribbean waters in response to climate and ocean circulation changes. Note that variations in carbonate content at this location are not as clearly linked to climate-induced changes in deep-water chemistry as the fragmentation record. This is because the carbonate content of the sediments can also be affected by carbonate productivity at the surface and by dilution on the seafloor by noncarbonate sediment types. (Unpublished data from L. Peterson.)

Seafloor Diagenesis With time and burial, carbonate oozes undergo a progressive sequence of diagenesis and are transformed first to chalk and then to limestone through a combination of gravitational compaction,

dissolution, reprecipitation, and recrystallization. Porosity is reduced from about 70% in typical unconsolidated carbonate oozes to roughly 10% in cemented limestones, while overall volume decreases by about one-third. Drilling results have shown that the transformation from ooze to chalk typically


CALCIUM CARBONATES

Q

Oligocene

Miocene

Plio.

453

Eocene

3000 South Atlantic (B 75)

Indian (P 92)

Water depth (m)

Atlantic (V 75) 4000

Nonequatorial Pacific (V 75)

5000

Equatorial Pacific (V 75)

Indian (S 77)

0

10

20

30

40

50

Age (Ma) Figure 6 Compilation of reconstructed variations in the depth of the CCD from selected studies covering the last 50 million years for different oceanic regions. The overall similarity of the CCD behavior between regions suggests a common forcing mechanism, such as global sea level or a long-term change in the supply of calcium to the ocean. Variations between the oceans are probably the result of differences in regional surface productivity and deep circulation patterns. Cited CCD studies include: V75, van Andel (1975); B75, Berger and Roth (1975); S77, Sclater et al. (1977); P92, Peterson et al. (1992).

occurs within a few hundred meters of burial, while limestones are produced by further cementation under about 1 km of burial. Although the transformation of ooze to chalk to limestone is the expected diagenetic sequence, smaller scale reversals in lithification are often observed. Such reversals in pattern have led to the concept of diagenetic potential, which simply states that different sediments will take different lengths of time to reach equal stages of lithification depending upon the original character of the deposited sediment. Such factors as the original proportions of coccoliths to foraminifera (affecting grain size), the amount of dissolution experienced before burial, sedimentation rates, and numerous other subtle factors can influence the diagenetic potential of a carbonate sediment. To the extent that these factors reflect original oceanographic conditions, the sub-bottom acoustic reflectors that result from changing lithification state and diagenetic potential preserve a history of paleooceanographic events that can often be traced across large regions within ocean basins.

See also Acoustics in Marine Sediments. Carbon Dioxide (CO2) Cycle. Cenozoic Climate – Oxygen Isotope Evidence. Cenozoic Oceans – Carbon Cycle Models. Cold-Water Coral Reefs. Ocean Carbon System, Modeling of. Plankton and Climate. Pore Water Chemistry. Protozoa, Planktonic Foraminifera. Sea Level Change. Sedimentary Record, Reconstruction of Productivity from the.

Further Reading Archer DE (1996) An atlas of the distribution of calcium carbonate in sediments of the deep sea. Global Biogeochemical Cycles 10: 159--174. Arrhenius G (1988) Rate of production, dissolution and accumulation of biogenic solids in the ocean. Palaeogeography, Palaeoclimatology and Palaeoecology 67: 1119--1146. Berger WH (1976) Biogenous deep sea sediments: production, preservation and interpretation. In: Riley JP


454

CALCIUM CARBONATES

and Chester R (eds.) Chemical Oceanography, vol. 5, pp. 266--388. London: Academic Press. Berger WH and Roth PH (1975) Oceanic micropaleontology: progress and prospect. Reviews of Geophysics and Space Physics 13: 561--585. Broecker WS and Peng T-H (1982) Tracers in the Sea. Palisades, NY: Lamont-Doherty Geological Observatory Press. Emerson S and Archer DE (1992) Glacial carbonate dissolution cycles and atmospheric pCO2: a view from the ocean bottom. Paleoceanography 7: 319--331. Jahnke RA, Craven DB, and Gaillard J-F (1994) The influence of organic matter diagenesis on CaCO3 dissolution at the deep-sea floor. Geochimica Cosmochimica Acta 58: 2799--2809. Milliman JD (1993) Production and accumulation of calcium carbonate in the ocean: budget of a nonsteady state. Global Biogeochemical Cycles 7: 927--957. Peterson LC and Prell WL (1985) Carbonate dissolution in recent sediments of the eastern equatorial Indian Ocean: Preservation patterns and carbonate loss above the lysocline. Marine Geology 64: 259--290. Peterson LC, Murray DW, Ehrmann WU, and Hempel P (1992) Cenozoic carbonate accumulation and compensation depth changes in the Indian Ocean.

In: Duncan RA, Rea DK, Kidd RB, von Rad U, and Weissel JK (eds.) Synthesis of Results from Scientific Drilling in the Indian Ocean, Geophysical Monograph 70, pp. 311--333. Washington, DC: American Geophysical Union. Schlanger SO and Douglas RG (1974) The pelagic oozechalk-limestone transition and its implications for marine stratigraphy. In: Hsu¨ KJ and Jenkyns HC (eds.) Pelagic Sediments on Land and Under the Sea, Special Publication of the International Association of Sedimentologists, 1, pp. 117--148. Oxford: Blackwell. Sclater JG, Abbott D, and Thiede J (1977) Paleobathymetry and sediments of the Indian Ocean. In: Heirtzler JR, Bolli HM, Davies TA, Saunders JB, and Sclater JG (eds.) Indian Ocean Geology and Biostratigraphy, pp. 25--60. Washington, DC: American Geophysical Union. Takahashi T, Broecker WS, Bainbridge AE, and Weiss RF (1980) Carbonate Chemistry of the Atlantic, Pacific and Indian Oceans: The Results of the GEOSECS Expeditions, 1972–1978, Lamont-Doherty Geological Observatory Technical Report 1, CU-1-80. van Andel TH (1975) Mesozoic-Cenozoic calcite compensation depth and the global distribution of calcareous sediments. Earth and Planetary Science Letters 26: 187--194.


CALIFORNIA AND ALASKA CURRENTS B. M. Hickey, University of Washington, Seattle, WA, USA T. C. Royer, Old Dominion University, Norfolk, VA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 368–379, & 2001, Elsevier Ltd.

Introduction The clockwise North Pacific Subtropical Gyre and the counterclockwise Subarctic or Alaska Gyre, the two principal current gyres of the North Pacific, have dimensions similar to those of the basins, i.e., several thousand kilometers. The West Wind Drift and Subarctic Current flow approximately zonally across the Pacific basin with origins in the Kuroshio and Oyashio Currents, respectively, in the western Pacific basin (Figure 1). As this broad eastward flow nears the west coast of North America, it bifurcates, splitting into the northward flowing Alaska Current, the eastern limb of the Alaska Gyre, and the southward flowing California Current, the eastern limb of the North Pacific Subtropical Gyre. This bifurcation takes place several hundred kilometers offshore and depends both on the ocean currents and the wind fields over the North Pacific. The water type is primarily Subarctic, relatively cool and fresh surface water. The volume transport within each of the currents is about 10–15 Sv. Both the California Current 140

160°E

180

and the Alaska Current are coupled to current systems and processes along the adjacent continental margins, where the majority of the seasonal variability occurs. On both seasonal and ENSO (El Ninõ) timescales, the strengths of the two current systems vary out of phase. For each boundary current system (the Alaska system and the California system) both the large-scale gyres and the coastal current systems, as well as the interactions between them, are described in the text below.

The Alaska Current System The Alaska Current, the eastern limb of the Subarctic Gyre, flows northward along the west coast of North America beginning at about 48–501N (Figure 2). A companion coastal current flows parallel to the Alaska Current but closer to the coast. This current has several local names – the Vancouver Island Coastal Current, the Haida Current, and the Alaska Coastal Current. These flows move in a counterclockwise sense around the Gulf of Alaska, bringing relatively warmer water northward along the eastern boundary of the Pacific Ocean. As they follow the topography in the northern gulf, they are diverted westward. Early explorations of the Subarctic North Pacific, initiated by the Russian czar, Peter the Great, illustrate this general circulation of the northern North Pacific region. In June 1741, two ships set out from the Kamchatka Peninsula to investigate the eastern 160°W

140

120

60° N

JAPAN

50

Figure 1 Schematic surface circulation of the North Pacific relative to 1000 db (i.e., assuming no flow near a depth of 1000 m) showing the Alaska and California Current systems. (Adapted with permission from Thomson, 1981.)


455

456

CALIFORNIA AND ALASKA CURRENTS

60°N

55°

50° 170°W

160°

140°

150°

Figure 2 Schematic of North Pacific Subarctic Gyre with Alaska Current, Alaskan Stream and Alaska and Haida Coastal Currents. (Adapted with permission from Reed and Schumacher, 1986.)

side of this unexplored basin. They sailed following the currents and winds south-eastward and then eastward across the North Pacific at about 48–501N. Near the date line, the two vessels were separated under foggy conditions, never to see one another again. Nevertheless, explorers on both ships eventually observed North America: from the St. Peter commanded by Vitus Bering, land was first seen off south-east Alaska on 15 July and from the St. Paul commanded by Aleksei Chirikov, land was seen near Kayak Island on 16 July. Both ships traveled back toward the west along the southern side of the Aleutian Islands, completing a counterclockwise path. This path is approximately the configuration of the mean ocean currents in the Subarctic Gyre. The St. Paul returned to Kamchatka in fall 1741, but the St. Peter ran aground on an island in the Bering Sea (now known as Bering Island) where many perished, including Bering himself. Atmosphere–Ocean Interactions

In winter, cold, dry Siberian air masses frequently sweep out over the northern North Pacific, rapidly extracting heat and moisture from the ocean. The introduction of this heat into the atmosphere intensifies the atmospheric circulation. These storms generally move from west to east across the Pacific basin. The path of an individual storm across the North Pacific depends on the global scale atmospheric circulation which changes both seasonally and interannually. In winter, storm tracks often cross the Gulf of Alaska, resulting in a large zone of low

atmospheric pressure known as the Aleutian Low. In summer, the North Pacific High strengthens and pushes northward into the gulf. Winter over the North Pacific Ocean is dominated by strong counterclockwise wind systems; in summer weak clockwise winds occur. Since the earth’s rotational force (‘Coriolis’) tends to move near surface water to the right of the wind direction in the northern hemisphere, the counterclockwise winter winds over the Gulf of Alaska transport upper layer water shoreward, away from the central deep ocean. The subsurface waters move upward to replace these surface waters. The rising of these subsurface waters together with wind mixing brings nutrient-rich waters into the upper layers of the ocean where they can be utilized by phytoplankton. Over the continental shelf and along the coastline, the surface waters are transported shoreward in winter, leading to convergence and downwelling. The pattern of upwelling and downwelling near the coast changes seasonally from intense downwelling in winter to weak upwelling or neutral conditions in summer (Figure 3, upper panel). Progressing southward along the west coast of North America, the seasonal cycle of upwelling and downwelling changes from a downwelling-dominated wind system (northward winds along the coast) in the north to a more upwellingdominated wind system (southward winds along the coast) farther south off Washington, Oregon, and California in the California Current system (see below). The seasonal progression of storm tracks across the Gulf of Alaska also affects precipitation rates,



457

50

25 0

40 −25 Upwelling index −50 (60°N, 149°W)

Discharge 3 −1 3 (10 m s )

30

−75 20 −100 −125

10 F

J

M

A

M

J

M

J

J

A

S

O

N

D

J

N

D

33.0

32.0 31.0 10 m

Salinity

20 m

30.0

30 m 50 m

29.0 75 m 100 m

28.0

150 m 250 m

27.0 J

F

M

A

J

A

S

O

J

Figure 3 Seasonal changes in strength of upwelling/downwelling in the northern Gulf of Alaska and coastal freshwater discharge (upper panel). Upwelling (positive) or downwelling (negative) is indicated via the ‘upwelling index’, a measure of cross-shore transport by alongshelf winds (units are m3 s1 per 100 m of coastline). Seasonal changes in salinity versus depth in the Alaska Coastal Current (601N, 1491W) (lower panel).

especially along the coast. As these storms encounter the coastal mountain ranges that rim the eastern boundary of the northern North Pacific, heavy precipitation occurs. This provides vast quantities of fresh water in the coastal region and adds heat to the atmosphere. On average, about 2.4 m of rain and snow fall in a relatively narrow (B100 km) coastal drainage area and more than 8 m of precipitation have been reported for a single year. The majority of this precipitation runs directly into the ocean via coastal rivers when the air temperatures are above freezing. Otherwise, it is stored as snow and ice with about 20% of the region being glacial. The average annual coastal runoff is estimated to be 24 000 m3 s1, about one third greater than the outflow of the Mississippi River. However, unlike the Mississippi, the runoff here is distributed along the coast in a number

of smaller rivers rather than through a single major river. Fresh water is continually added to the coastal currents as they progress around the Gulf of Alaska. The coastal discharge (Figure 3, upper panel) is least in winter when most of the moisture is contained in snow and ice. A small peak occurs in spring corresponding to seasonal heating at lower elevations. However, maximum discharge occurs in September– November prior to annual freeze-up, when both precipitation caused by storms and runoff from snowmelt contribute to the coastal runoff. The Alaska Current

The Alaska Current is affected by both winds and precipitation. Winds, which are usually downwelling-favorable along the coast, maintain the


458


density contrast between central Gulf of Alaska water and fresher, lower density water on the shelf. Precipitation and coastal runoff also diminish the water density on the shelf. In the eastern Gulf of Alaska, the Alaska Current is a relatively broad meandering flow, typically several hundred kilometers wide (Figure 2). Frequently it contains large mesoscale eddies that often move westward at several centimeters per second, taking years to complete their journey (Figure 4). These eddies serve as a mechanism for the transfer of energy and water from the ocean boundaries into the ocean’s interior. The Alaska Current follows the general shelf break topography around the Gulf of Alaska in a counterclockwise sense. It turns westward at the apex of the Gulf of Alaska, then south-westward in the western Gulf of Alaska where it behaves as a western boundary current, intensifying into an organized flow known as the Alaskan Stream (Figures 1 and 2). This current is highly sheared vertically due to the density field (baroclinic), with a cross-current density gradient manifested by the offshore gradient in salinity. Surface salinities range from o30 PSU at the coast, to 431 PSU on the shelf, to 432.5 PSU over the central Gulf of Alaska. The baroclinic current transport near Kodiak Island in the western gulf (see Figure 2) is approximately 10 Sv in the upper 155°W

150°

145°

1500 m, with maximum speeds exceeding 1.0 m s1. Most of the flow (480%) is found within 60 km offshore of the shelf break. Although the seasonal variability of the atmospheric forcing is large (Figure 3, upper panel), the seasonal changes in transport are relatively small – less than 10%. As the topography turns more westward, some of the Alaskan Stream turns southward where it rejoins the eastward flowing Subarctic Current, completing the circuit around the Alaska Gyre. However, most of the flow continues eastward and some fraction enters the Bering Sea, subsequently returning to the North Pacific in the Kamchatka Current that becomes the Oyashio Current (Figure 1). The Alaska Gyre Coastal Currents

The coastal fresh water runoff affects the nearshore salinity in the Gulf of Alaska (Figure 3, lower panel) and causes a strong offshore gradient in density. The winds driving downwelling confine the fresh water to the coast and also drive currents along the coast. The resulting cross-shelf density gradient, through a balance between pressure forces and the earth’s rotational forces, causes the coastal current in the Gulf of Alaska to be strongly baroclinic. The currents driven by these processes have been given specific

140°

135°

130°

125° 60°N

BRITISH COLUMBIA

55°

50°

Figure 4 A composite sea surface temperature image for the Gulf of Alaska from 1, 2, 3 and 10 March 1995. The mean north–south temperature gradient has been removed to better elucidate eddy structure. Oceanic temperatures differed from the mean by 31C (dark blue) to þ 31C (red). Clouds and land are white. The dotted curve is the 1000 m depth contour. (Reproduced with permission from Thompson and Gower, 1998; copyright by the American Geophysical Union.)



names locally (the Vancouver Island Coastal Current off Vancouver Island, the Haida Current off central British Columbia, and the Alaska Coastal Current off the Alaskan coast). It is presently unknown whether the coastal current is continuous throughout the Gulf of Alaska at a given time. The coastal current likely begins with the Davidson Current, a wintertime current in the California Current system (see below) that links the two major current systems, at least in winter. The northward tending Alaska Coastal Current is strongest in winter and weakest in summer, exactly out of phase with the southward tending coastal currents in the California Current (see later sections). The buoyancy and wind-driven coastal current flows along the coast with the coast on its right around the Gulf of Alaska, eventually flowing through Unimak Pass into the Bering Sea (see Figure 2). The flow has a typical width of about 30 km, a depth of 100–200 m and speeds 41.0 m s1 have been reported. The mean transport is about 0.6 Sv with a seasonal variation of about 0.2 Sv. Transports of more than 2 Sv have been reported. The coastal current is usually constrained along the coast by the downwelling winds. As it passes coastal openings at the entrance to Prince William Sound and Cook Inlet, some of the current enters these enclosures. Near Kayak Island (about 1441W) the coastal current is diverted across the shelf and some of it merges with the Alaska Current. In the western Gulf of Alaska, downwelling winds are less dominant and precipitation rates are lower. Here the current tends to diverge from the coast and spread across the shelf. The coastal current is believed to be important for marine mammals and fish including salmon. Salmon might use chemical tracers carried by this current to navigate back to their original spawning streams. Similarly, the current is capable of carrying pollutants alongshore. In March 1989, more than 242 000 barrels of crude oil were spilled into Prince William Sound from T/V Exxon Valdez. The oil was carried westward in the Alaska Coastal Current more than 800 km in the span of about 2 months (about 0.15 m s1). The reduction in North Pacific sea surface salinity by precipitation and runoff (Figure 3, lower panel) creates a surface low-density lens that restricts vertical mixing in the Alaska Gyre and tends to prevent deep water formation. This is in sharp contrast to the deep circulation of other high latitude regions of the world’s oceans such as the North Atlantic or Weddell Sea.

The California Current System The California Current system (Figure 5) includes the southward California Current, the wintertime

459

northward Davidson Current, the northward California Undercurrent (which flows over the continental slope beneath the southward flowing upper layers), the Southern California Countercurrent (or Eddy) as well as ‘nameless’ shelf and slope currents with primarily shorter-than-seasonal time scales. The California Current system includes one major river plume (the Columbia), several smaller estuaries, and (primarily in the north) numerous submarine canyons. The dominant scales and dynamics of the circulation over much of the California Current system are set by several characteristics of the physical environment; namely, strong winds, large alongshore scales for both the winds and the bottom topography, and a relatively narrow and deep continental shelf. Because of these characteristics, coastal-trapped waves (disturbances that travel northward along the shelf and slope) are efficiently generated and travel long distances toward the North Pole along the continental margins of much of western North America. Typical speeds are about 300–500 km d1. These waves are integral to the current patterns observed in the California Current system and to their variability. The California Current flows southward yearround offshore of the US and Mexican west coast from the shelf break to a distance of several hundred kilometers from the coast (Figure 5). The current is strongest at the sea surface, and generally extends over the upper 500 m of the water column. Seasonal mean speeds are B0.1 m s1. The California Current in summer carries relatively colder, fresher water southward along the coast. South of Point Conception (the major indentation in the coastline near 351N), a portion of the California Current turns southeastward and then shoreward and northward. This feature is known either as the ‘Southern California Countercurrent’ during periods when the flow successfully rounds Point Conception or the ‘Southern California Eddy’ when the flow recirculates within the southern California Bight (the indented region south of Point Conception; see inset map in Figure 5). The California Undercurrent is a relatively narrow feature (B10–40 km) flowing northward over the continental slope of the California Current system at depths of about 100–400 m as a nearly continuous feature, transporting warmer, saltier water of more southern origin northward along the coast. The Undercurrent has a jet-like structure, with the core of the jet usually located just seaward of the shelf break and peak speeds B0.3–0.5 m s1. The Undercurrent divides into two components within the Southern California Bight, one flowing northwestward through the Santa Barbara Channel, the other flowing westward south of the island chain that


460

CALIFORNIA AND ALASKA CURRENTS 140°

120°

130°

110°

40°

50° 30°

Winter

45°

DC

40° 35°

SCC

20° 30° 25°

CC

20° 135° 130° 125° 120° 115° 110° 50°

50°

Spring

Summer

50°

Fall DC

45°

45°

40°

40°

40°

35°

35°

45°

DC Coastal Jet

30°

SCE

30°

CC

35° 30°

25°

25°

20° 135° 130° 125° 120° 115° 110°

20° 135° 130° 125° 120° 115° 110°

25°

CC

CC

20° 135° 130° 125° 120° 115° 110°

Figure 5 Schematic illustrating seasonal variation of large-scale boundary currents and coastal currents off the west coast of North America as well as important landmarks and bottom topography. (Adapted with permission from Strub and James, 2000.) An enlargement of the Southern California Bight, including the offshore islands, is given in the upper left panel. CC, California Current; DC, Davidson Current; SCC, Southern California Countercurrent; SCE, Southern California Eddy.

forms the southern side of the Santa Barbara Channel. A southward undercurrent occurs over the continental slope in winter at some latitudes. This undercurrent occurs at deeper depths than the northward undercurrent (B300–500 m). The existence of this southward undercurrent, like that of the northward undercurrent, likely depends on the cooccurrence of opposing local alongshore winds and alongshore sea level slope. The Davidson Current

flows northward in fall and winter from Point Conception (B351N) to at least Vancouver Island (B501N) and may connect with the coastal currents that flow around the Alaska Gyre. This northward flow is generally broader (B100 km in width) and sometimes stronger than the corresponding subsurface northward flow in other seasons (the ‘Undercurrent’) and extends seaward of the slope. Poleward shelf flow, in the sense of a monthly mean



phenomenon, is sometimes described as an expression of ‘the Davidson Current’. Winds in the California Current system are governed by atmospheric pressure systems – on average a low pressure system in winter and a high pressure system in summer, producing northward winds in winter and southward winds in summer over much of the California Current system. Because of the dramatic seasonal reversal in winds along much of the coast, currents and water properties of the California Current system also undergo large seasonal fluctuations. The southward flowing California Current and the northward flowing California Undercurrent are strongest in summer to early fall and weakest in winter. The majority of the seasonal variability occurs along the coastal boundaries rather than in the central basin. The seasonal signal in transport near the coast migrates northward and offshore in both the California and Alaska Current systems. The North Pacific Current, which feeds both the California Current and the Alaska Current (see Figure 1) has little seasonal variability in either strength or position. Much of the variability in the California Current is related to coastal wind variability. However, remote forcing (disturbances that are caused farther south and travel along the coastal margins as waves) is also likely important, particularly off California. The northward flowing Davidson Current is strongest in winter, as is the Southern California Countercurrent. Shelf currents along the coast from Point Conception to the Strait of Juan de Fuca are generally southward in the upper water column from early spring to summer and northward the rest of the year. The seasonal duration of southward flow usually increases with distance offshore and with proximity to the sea surface. A northward undercurrent is commonly observed on shelves during the summer and early fall. A strong tendency for northward flow throughout the water column exists over the inner shelf in all but the spring season. The transition of currents and water properties over the shelf and slope between winter and spring, the ‘spring transition’, is a sudden and dramatic event in the California Current system. Along much of the coast sea level drops at least 10 cm during the transition, currents reverse from predominantly northward to predominantly southward within a period of several days and isopycnals tilt upward towards the coast. The transition is driven by changes in the large-scale wind field and these changes are a result of changes in the large-scale atmospheric pressure field over the Northeast Pacific and the California Current system. The transition includes both a local and a remotely forced response to the

461

change in wind conditions. The fall transition is not as rapid or as dramatic as the spring transition. Variability on Shorter-than-Seasonal Timescales

Seasonal conditions are often reversed for shorter periods of time in the California Current system. Changes in currents, water properties, and sea level over the shelf at most locations are dominated by wind forcing, with typical timescales of 3–10 days. Regions seaward of the shelf are dominated by jets, eddies, and in some locations, wave-like propagating disturbances, with typical timescales of 10–40 days. Along-shelf flow on the inner to mid shelf is primarily wind-driven and can be predicted with numerical models. However, the amplitude of current fluctuations is generally underpredicted, and the amount of variability predicted decreases offshore toward the shelf break (to o20% over the upper slope off northern California). Predictive capability for both temperature fluctuations and cross-shelf flow is very poor at the present time. The alongshelf currents include a response to both local wind forcing and wind forcing all along the coast south of a particular latitude (‘remote’ forcing). At any given time and location, the ratio of remote and local forcing varies. In winter, local wind forcing of currents dominates in regions where winter storms are accompanied by strong northward winds that increase northward (in the direction of propagating coastal-trapped waves). In summer, when winds increase southward, freely propagating coastal-trapped waves often contribute to current variability in the coastal regions of the Pacific Northwest. Off northern California, where wind stress is generally strongest in summer, both local and remote forcing are almost always important. Wind-driven currents in the Southern California Bight typically have much smaller alongshelf scales than in the region north of the Bight (20 km versus 500 km), weaker amplitudes, and weaker seasonal variation; likely a result of the much reduced winds and the narrow, more irregular shelves in this region. The eddy/meander field in the regions seaward of the shelves in the California Current system has a seasonal variation, with maximum energy in summer, when upwelling is also at a maximum (Figure 6). Meandering jets extend from the sea surface to depths of over 200 m and separate fresher, warmer, chlorophyll-depleted water from colder, saltier, chlorophyll-rich, recently upwelled water. Jets are characterized by core speeds exceeding 0.5 m s1 at the surface, widths of 50–75 km and total baroclinic transports of about 4 Sv. These filaments are particularly evident in regions which contain


462


Columbia River

46

Washington Oregon

44

Cape Blanco

Oregon California

42

T°C 18 17 Cape Mendocino

16 15 14

40

13 12 11 Pt. Arena

38

36

−128

−126

−124

−122

Figure 6 Satellite-derived image of sea surface temperature in the California Current system. The figure illustrates the jets, eddies, and meanders, many of which originate near coastal promontories. (AVHRR data collected and processed at Ocean Imaging, Inc., archived and made available at COAS/Oregon State University with funding from NSF and NASA in the US GLOBEC Northeast Pacific program.)



coastline irregularities, such as southern Oregon, northern California, and the Baja peninsula, and appear to be tied to these irregularities. One meandering jet can be traced continuously from southern Oregon, where it separates from the shelf, to southern California. Such separated coastal jets account for much of the energy, as well as seasonal and interannual variability in the California Current system. The jets and meanders are a dominant feature in satellite-derived images of sea surface temperature in the summer season (Figure 6). Changes in currents with timescales similar to those of the eddy/meander field (15–40 days) dominate current variance over the slope within the southern California Bight. The majority of this variability is the signature of freely propagating coastal-trapped waves or disturbances, with an as yet unknown source along the coast of Baja California or even farther south. Such waves are likely to be important in other slope regions, but to date have not been separated from other sources of variability. Water Properties and Upwelling

The California Current system contains waters of three primary types: Pacific Subarctic, North Pacific Central and Southern (sometimes termed ‘Equatorial’). Pacific Subarctic water, characterized by relatively low salinity and temperature and high oxygen and nutrients, is advected southward in the outer edges of the California Current system and into the southern California Bight. Colder, fresher water from British Columbia and the US Pacific Northwest coastal region is also advected southward near the boundaries. North Pacific Central water, characterized by high salinity and temperature and low oxygen and nutrients, enters the California Current system from the west. Southern water, characterized by high salinity, temperature and nutrients, and low oxygen, enters the California Current system from the south with the northward flowing California Undercurrent. In general, salinity and temperature increase southward in the California Current system and also with depth. Upwelling along the coast brings colder, saltier and more nutrient-rich water to the surface adjacent to the coast. In general, maximum upwelling (as seen at the sea surface) occurs off northern California, consistent with the alongshore maximum in alongshelf wind stress and hence mass transport away from the coast that causes deeper water to ‘upwell’ to fill the void (Ekman transport) (Figure 7). Maximum upwelling occurs in spring or summer in the California Current system. Stratification in the California Current system is remarkably similar at most locations

463

and is largely controlled by the large-scale advection and upwelling of the water masses as described above. With the possible exception of catastrophic storms (a ‘10-year storm’) most river plumes on the coast between the tip of Baja California and the Strait of Juan de Fuca are relatively small, and their effects are likely confined to within one tidal excursion of the mouth of the river or estuary. The only large river plume in the California Current system is that from the Columbia River. The Columbia provides over 77% of the drainage between the Strait of Juan de Fuca and San Francisco Bay. As mentioned above, water from the Strait of Juan de Fuca, with origins in the Fraser River, also contributes to fresh water in the nearshore California Current in spring and summer. The plume from the Columbia River is a dominant feature in near surface salinity (and, in some seasons, temperature) off the US west coast. On a seasonal basis, the plume flows northward over the shelf and slope in fall and winter, and southward well offshore of the shelf in spring and summer. In winter the Columbia plume provides a substantial fraction of fresh water to the Davidson Current, and, ultimately, the coastal currents of the Alaska Gyre. In summer, the Columbia provides fresh water to the California Current, giving rise to the low salinity signal and associated front used to trace the meandering coastal jet that separates from the shelf at Cape Blanco. Most other smaller rivers on the Pacific coast have significant river plumes only during major floods. In winter and spring, the Columbia plume has a dramatic effect on the Washington coast, producing time-variable currents as large as the wind-driven currents and flooding local estuaries with fresh water (fresher than that in much of the estuary). Short-term variability in both fresh water volume and plume location can have significant effects on shelf and slope currents as well as water properties.

Topographic Effects in the Alaska/California Current Systems The outer edge of the continental shelf along the US west coast from California to the Aleutian Islands is intersected at many locations by submarine canyons. Canyons enhance the transfer of particles and water between the shelf and the deeper ocean. Upwelling of colder, nutrient-rich water can be enhanced by more than a factor of 10 over regions with a straight slope rather than a canyon indentation. For example, upwelling from a spur of Juan de Fuca canyon is responsible for enhanced seasonal productivity in that region. In the Alaska Current system, Hinchinbrook


464

CALIFORNIA AND ALASKA CURRENTS 50°N Apr

Jan

Columbia R.

10

45°

10

10

Cape Mendocino

40°

Pt. Conception 35°

15

15

14 30°

20

25° 50°N

15

Sep

Jul

45°

15 13 15

15

13

40°

14

35°

20

17 15 17

17 20

30° 25 25°

Figure 7 Maps of monthly mean sea surface temperature in the California Current System. The corresponding profile of cross-shore transport due to local winds is shown next to each map. Each tick represents a mass transport of 50 103 kg s1 per 100 m of coastline. Offshore transport, indicative of upwelling conditions, is shaded. (Adapted with permission from Huyer, 1983.)

Canyon, which crosses the shelf adjacent to Prince William Sound, is a potential conduit for the exchange of deep waters with this coastal feature. Moreover, counterclockwise circulation patterns are generally observed both within and over submarine canyons (although not necessarily extending to the sea surface). Such eddies provide an effective mechanism for trapping particles such as suspended sediment or food for the biomass. Several major promontories and one major ridge (near Cape Mendocino) also occur along the coast adjacent to the California Current system (see Figure 5). Flow in the vicinity of such features is highly three-dimensional, generally producing offshore tending jets as well as counterclockwise eddies in their lee.

Interannual and Interdecadal Variability in the Alaska/California Current Systems Year-to-year variability in the California and Alaska Current systems is significant in both physical and biological parameters. Surface transport in the two

systems varies out of phase on El Ninõ timescales: when the Alaska Gyre strengthens, as during an El Ninõ, the California Current system weakens. As with seasonal changes, the majority of the changes in transport occur along the ocean boundaries in both systems. On interannual scales, some change in the North Pacific Current also occurs, but these changes follow those along the boundary. For the California Current system property anomalies of B2–41C and B0.3 PSU have been observed at depths of 50–200 m from the sea surface, and these anomalies extend several hundred kilometers from the coast. Much of this variability is related to the El Ninõ (ENSO) phenomenon, occurring at periods of 3–7 years (Figure 8, right panels). In the Alaska Current system, subsurface temperature anomalies of 41.51C have been observed at the coast 7–8 months after an El Ninõ has occurred at the equator. El Ninõ has been shown to affect the west coast both via an atmospheric route (i.e., by changes in the local winds which then cause changes in the local ocean currents and water properties) and by an oceanic route (i.e., by transmission of signals from the equator along the continental margin as propagating coastal-trapped



465

El Niño Southern Oscillation

Pacific Decadal Oscillation °C

0.8 0.4 0.2 0.0 _ 0.2 _ 0.6

monthly PDO index values 4 2 °C 0 _2 _4

monthly Nino3.4 index values

4 2 0 _2

1900

1920

1940

1960

1980

2000

_4

1900

1920

1940

1960

1980

2000

Figure 8 October-to-March average surface climate anomalies associated with a þ 1 standard deviation value in the Pacific Decadal Oscillation (PDO) index (left panel) and El Ninõ (cold-tongue) Index (right panel). Sea surface temperature anomalies are depicted by the color shading in degrees Celsius; surface wind stress anomalies are shown with vectors, with the largest vectors representing 10 m2 s2 anomalies. Thin solid contours depict positive sea level pressure (SLP) anomalies; dashed contour lines depict negative SLP anomalies, with contour intervals at 70.5, 1, 2 and 3 mb. The heavy solid contour depicts the SLP anomaly zero-line. (Courtesy of Steven Hare at the International Pacific Halibut Commission and Nathan Mantua at the Joint Institute for the Study of the Atmosphere and Oceans, University of Washington.)

waves). Local wind effects are more dominant in the Pacific Northwest and Alaska; remote forcing via waves is more dominant from central America to California. There is also evidence of increased formation of eddies along the shelf break in the eastern Gulf of Alaska during El Ninõ conditions. The Northeast Pacific also has lower frequency fluctuations – periods of about 22 and 52 years, associated with the Pacific Decadal Oscillation (PDO) (Figure 8, left panels). PDO has oceanic and atmospheric patterns similar to those of El Ninõ, but has much longer duration. PDO is a pattern of low sea surface temperatures in the central North Pacific and high sea surface temperatures along the eastern Pacific boundary. The major pattern reverses at 25–30 year intervals. PDO was mostly positive from 1922 to 1942 and mostly negative (warm in the central gulf, cold nearshore) from 1942 to 1976 and has been mostly positive again through 1998. From fall 1998 until summer 2001 the PDO has been negative. The coastal precipitation follows a similar pattern and could reinforce the PDO locally. High rates of coastal precipitation would increase cross-shelf pressure differences, enhancing northward flow in the coastal currents as well as the Alaska Current through the balance between pressure and rotational forces. Relatively warm water from more southern latitudes would thus be advected northward,

reinforcing the positive PDO sea surface temperature pattern. The ecosystem seems to respond to the PDO – salmon production in the northern Gulf of Alaska is positively correlated with PDO; Washington/Oregon salmon production is negatively correlated with PDO.

See also Arctic Ocean Circulation. Coastal Trapped Waves. Ocean Circulation. River Inputs. Salmon Fisheries, Pacific. Wave Generation by Wind.

Further Reading Alverson DL and Pruter AL (eds.) (1972) Bioenvironmental Studies of the Columbia River Estuary and Adjacent Ocean Regions. Seattle: University of Washington Press. Chavez FP and Collins CA (eds.) (1998) Studies of the California Current System Part 1. Deep-Sea Research 45(8–9): 1407–1904. Chavez FP and Collins CA (eds.) (2000) Studies of the California Current System Part 2. Deep-Sea Research 47(5–6): 761–1176. Divin VA (1993) The Great Russian Navigator, A.I. Chirikov. Fairbanks: University of Alaska Press.


466


Dodimead AJ, Favorite F, and Hirano T (1963) Review of the oceanography of the Subarctic Pacific. In: Salmon of the North Pacific. Vancouver: International North Pacific Fisheries Commission. Hare SR, Mantua NJ, and Francis RC (1999) Inverse production regimes: Alaska and West Coast Pacific Salmon. Fisheries 24: 6--14. Hickey BM (1979) The California Current System: hypotheses and facts. Progress in Oceanography 8: 191--279. Hickey BM (1998) Coastal Oceanography of Western North America from the tip of Baja California to Vancouver Is. In: Brink KH and Robinson AR (eds.) The Sea, vol. 11, pp. 345--393. New York: Wiley and Sons. Huyer A (1983) Upwelling in the California Current System. Progress in Oceanography 12: 259--284. Landry MR and Hickey BM (eds.) (1989) Coastal Oceanography of Washington and Oregon. Amsterdam: Elsevier Science. Lentz SJ and Beardsley RC (1991) Introduction to CODE (Coastal Ocean Dynamics Experiment) A: Collection of Reprints. Woods Hole, MA: Woods Hole Oceanographic Institution. Lynn RS and Simpson JJ (1987) The California Current System: the seasonal variability of its physical characteristics. Journal of Geophysical Research 92(C12): 12 947--12 966. Reed RK and Schumacher JD (1986) Physical oceanography. In: Hood DW and Zimmerman ST (eds.) The Gulf of Alaska: Physical Environment and Biological Resources. NOAA OCS Minerals Management Service MMS-860995. Springfield, VA. Royer TC (1983) Observations of the Alaska Coastal Current. In: Gade H, Edwards A, and Svendsen H (eds.) Coastal Oceanography, pp. 9--30. New York: Plenum.

Royer TC (1998) Coastal ocean processes in the northern North Pacific. In: Brink KH and Robinson AR (eds.) The Sea, vol. 11, pp. 395--414. New York: John Wiley Sons. Strub PT and James C (2000) Altimeter-derived variability of surface velocities in the California Current Systems: 2. Seasonal circulation and eddy statistics. Deep-Sea Research II 47(56): 831--870. Strub PT and James C (2000) Altimeter-derived surface circulation in the large scale NE Pacific Gyres: Part 1. Annual Variability. Progress in Oceanography (in press). Strub PT, Allen JS, Huyer A, Smith RL, and Beardsley RC (1987) Seasonal cycles of currents, temperatures, winds and sea level over the northeast Pacific continental shelf. Journal of Geophysical Research 92(C2): 1507--1526. Strub PT, Kosro PM, Huyer A, et al. (1991) The nature of cold filaments in the California Current system. Journal of Geophysical Research 96(C8): 14 743--14 768. Tabata S (1991) Annual and interannual variability of baroclinic transports across Line P in the northeast Pacific. Deep-Sea Research 38(supplement 1): S221--S245. Thomson RE (1981) Oceanography of the British Columbia Coast. Canadian Spec. Pub. Fish. and Aquatic Sciences 86: 291 pp. Thomson RE and Gower JFR (1998) A basin-scale oceanic instability in the Gulf of Alaska. Journal of Geophysical Research 103(C2): 3033--3040. Wilson JG and Overland JE (1986) Meteorology. In: Hood DW and Zimmerman ST (eds.) The Gulf of Alaska: Physical Environment and Biological Resources, pp. 31--54. Springfield, VA: NOAA OCS Minerals Management Service MMS-86-0995.


CANARY AND PORTUGAL CURRENTS E. D. Barton, University of Wales, Bangor, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 380–389, & 2001, Elsevier Ltd.

Introduction The Canary and Portugal Currents form the eastern limb of the North Atlantic Subtropical Gyre. Detailed knowledge of the currents is still surprisingly sparse in some respects and is largely based on indirect methods of estimating the flow. The entire eastern boundary of the gyre is affected by the process of coastal upwelling, driven by the seasonally varying Trade Winds. Upwelling is intimately related to the currents on the continental shelf, and varies on timescales from several days upwards. This phenomenon has been studied intensively at different places and times and long-term sampling has only begun in recent years.

Large-scale Circulation The low to midlatitudes of the North Atlantic Ocean are occupied by the clockwise rotating subtropical gyre (Figure 1). The western boundary of this system is made up by the Gulf Stream, which feeds into the North Atlantic Current and the Azores Current. The latter flows eastward to supply the eastern subtropical boundary region. Branches of the Azores Current loop gently into the Portugal Current and C.

ian

N

Subpolar Gyre r th

tic

sC .

60˚N

C.

A

l C.

Curren

Subtropical Gyre

North Equatorial Current

90˚N

t P or tug a

Az ore

St

Caribbean C.

n tla

No

m rea

C.

eg orw

30˚N

Canary

. or C rad Lab

lf Gu

d nlan

Latitude

ree

tG Eas

n ya

Gu aC

90˚W

60˚W

Equatorial Co un

ter C.

30˚W Longitude

0˚

0˚ 30˚E

Figure 1 Sketch of the general near-surface circulation of the North Atlantic Ocean.

further south into the Canary Current. The latter separates from the African coast at around 201N to become the North Equatorial Current, which eventually feeds into the Caribbean Current and back to the Gulf Stream. Using all available hydrographic data, the longterm average geostrophic flow for the region has been calculated assuming a level of no motion near 1200 m. Scarcity of data limits the resolution of the analysis to a grid of 31 31; near the coast, where deep data are even fewer, results are more uncertain. Where the eastward-flowing Azores Current turns south as it nears the eastern boundary, two branches of the Canary Current are formed (Figure 2A) separated by Madeira. West of Iberia only weak nearsurface flow toward the eastern boundary is indicated, while nearer to shore the Portugal Current carries about 2 106 m3 s1 equatorward in the layers above 200 m depth. Only some of this continues southward into the Canary Current, while the rest apparently enters the Mediterranean in a shallow surface layer. The total amount of water carried equatorward above 200 m in the Canary Current, including input from the Portugal Current, was estimated at about 4 106 m3 s1 between 351W and the African coast. Near 201N the Canary Current breaks away from the African coast to turn westward as the North Equatorial Current near 151N. South of this separation point, a recirculation cell around the ‘Guinea Dome’ lies east of the Cape Verde Islands between the coast and the equatorward flow. The Canary Current varies seasonally by slight changes in position but not greatly in transport (Figure 3). Areas of larger uncertainty, shaded in the figure, correspond to few or no hydrographic samples in that study. The near-shore branch of the current migrates seasonally across the Canary Islands, closer to Africa in summer and farther offshore in winter. As it does so, the Azores Current oscillates south in summer and north in winter, so that the eastern part of the gyre has an annual ‘wobble.’ Some streamlines intersect the African coast in the northern half of the area and leave in the south, suggestive of a narrow, intense equatorward flow in the under-sampled coastal band between 201 and 301N, particularly in spring and summer. Recirculation south of 201N is clearest in winter and spring, but indications of northward flow are also seen there during other seasons. In autumn, nearshore northward flow extends as far as the Canaries, although again in the coastal under-sampled band.


467

468

CANARY AND PORTUGAL CURRENTS

41˚N

41˚N Azores

38˚N

38˚N

35˚N

35˚N

33˚N

32˚N

Madeira

29˚N

29˚N

26˚N

Latitude

Latitude

Canaries 26˚N

23˚N

20˚N

17˚N Cape Verde Islands

1 cm s

14˚N

_1

11˚N

11˚N

8˚N 35˚W 32˚W 29˚W 26˚W 23˚W 20˚W 17˚W 14˚W 11˚W

(A)

AFRICA

20˚N

17˚N

14˚N

23˚N

8˚N

35˚W 32˚W 29˚W 26˚W 23˚W 20˚W 17˚W 14˚W 11˚W

8˚W

(B)

Longitude

8˚W

Longitude

Figure 2 (A) Total transport of volume calculated from the long-term mean density field with the geostrophic assumption and summed from 200 m depth to the sea surface. Between any pair of streamlines the volume transport is 0.5 106 m3 s1. The calculations have less uncertainty inside the dashed line. (B) Geostrophic current vectors (thin arrows) at 200 m depth calculated from the long-term mean density field and mean observed currents (thick arrows) near 200 m at the sites marked by dots. The circled dot indicates the longest record, Kiel276.

The few available long-term moored observations are mainly well away from shore and the shallowest records are at about 200 m depth. Most lasted 1–2 years, but one mooring (Kiel276) is continuous since 1980 on the southern edge of the Azores Current (331N, 221W). In general, fluctuations were more than 5 times more energetic than the mean flow and so most records do not provide a reliable statistical estimate of the average. Nevertheless, the measured mean currents agree broadly with the large-scale geostrophic flow (Figure 2B) but indicate that the transport may be 450% more than indicated by the geostrophic calculations. No subsurface record indicated significant seasonal variability. However, the Kiel276 record demonstrated the dominance of meanders and eddies in the Azores Current and variability on the scale of a decade. Since the Azores Current feeds into the eastern boundary system, similar long-term variability likely occurs in the Canary and Portugal Currents as well, although no long-term current monitoring has yet been achieved there.

Several year-long moorings, deployed recently between the Canary Islands and the African shelf, showed highly variable flow, with maximum velocities near 0.3–0.4 m s1, but seasonal mean currents 10 times smaller. In mid-channel the upper 200 m layer flowed southward with variable strength. The bulk of the upper 600 m flowed mainly southward in winter and spring, but northward in summer and autumn. Despite increased Trade Winds in summer the Canary Current decreased east of the Canary Islands, at the same time as the flow through the western islands increased. These observations are compatible with the suggested seasonal variation in the gyre and the extension of near-surface poleward flow to the latitude of the Canaries, although the timing is out of phase with the long-term seasonal average results. Of course, any single year is not necessarily representative of the long-term mean. There was no evidence of strong equatorward flow along the continental margin in spring and summer, though measurements did not extend over the upper slope and shelf.



469

40˚N

Latitude

30˚N

20˚N

10˚N (A)

30˚W

20˚W

10˚W

30˚W

20˚W Longitude

10˚W

(B)

30˚W

20˚W

10˚W

40˚N

Latitude

30˚N

20˚N

10˚N (C)

30˚W (D)

20˚W Longitude

10˚W

Figure 3 Seasonal variation of the geostrophic volume transport above 200 m depth in the eastern boundary region, calculated as before for (A) spring, (B) summer, (C) autumn, and (D) winter. Areas of larger uncertainty are shaded. Flow between any pair of streamlines is 0.5 106 m3 s1.

Coastal Upwelling For some part of the year the north-east Trade Winds blow along every part of the subtropical Eastern boundary with a strong alongshore component that produces offshore Ekman transport in the surface layers and therefore upwelling at the coast. The strength of upwelling is conventionally expressed in terms of the upwelling (or Bakun) index, which is simply the Ekman transport TE ¼ ðt=rf Þ where t is

the component of wind stress parallel to shore, r is the density of sea water, and f is the Coriolis parameter. The index is routinely calculated from coastal surface wind data on a daily or monthly basis. It represents the rate at which water is removed from shore in the surface layer and, in a two-dimensional system, the amount of water upwelled to replace it. The annual cycle of upwelling for the region (Figure 4) shows the coastal temperature anomaly


470


with respect to central ocean alongside the monthly mean upwelling index. Summer Trade Winds affect the Iberian and African coasts north of 201N. Off the West Coast of Iberia, upwelling generally starts in May or June and lasts only until September. The southern, Algarve, coast of Portugal and north coast of Morocco (33–371N) are oriented at a large angle to the Trade Winds and so upwelling there is intermittent and short-lived. In winter the Trades shift

southward to provoke upwelling between 301 and 121N. South of 201N upwelling starts in December and lasts until April or May. Between 201 and 301N the coast is subject to year round upwelling that peaks in July and August. Wind forcing and strength of upwelling, as represented by coastal temperature anomaly, show variations up to a factor of 2 between years and decades. The 1960s were typified by upwelling of about half

Figure 4 Annual cycle of upwelling represented by (A) temperature difference between the coast and midocean determined by satellite sea surface temperature estimates, and (B) the upwelling index calculated from surface atmospheric pressure analyses.



of the average intensity off West Africa. Upwelling increased through the 1970s, only to weaken again in the 1980s and increase in the 1990s. An intriguing finding is that the strength of upwelling appears to be increasing in the long term – a trend common to all the upwelling regions of the world over the last 40 years. This may be linked to global warming, because increased summer heating deepens the continental low-pressure systems so increasing the atmospheric pressure difference with the oceanic highs to intensify the Trade Winds. On short timescales the Trade Winds typically remain nearly constant over periods of 7–10 days and then relax to near zero or weakly northward for several days. Observations over the continental shelf of NW Africa have shown how the system responds (Figure 5). Favorable winds drive offshore Ekman transport above about 30 m depth and raise the pycnocline near shore so that within one day colder,

0

100

471

less salty subsurface water breaches the sea surface. The upwelled water, which can originate from as deep as 200 m, is denser, and there is a slight downward slope of the sea surface toward the coast because of the offshore transport. These give rise to an equatorward geostrophic flow that weakens with depth over the shelf. The jet is strongest where the pycnocline intersects the surface as a boundary between the denser upwelled waters and the less dense oceanic waters, where it reaches velocities of up to 0.8 m s1 in the example shown. As the wind varies, so do the slopes of the sea surface and the pycnocline, the strength of upwelling, and the speed of the alongshore jet. The scale of the upwelling region, i.e., the distance within which the isosurfaces are uplifted is given by the Rossby radius of deformation l ¼ OðgDr=rhÞ=f where g is the gravitational acceleration, h is the undisturbed depth of the surface layer above the pycnocline, r is the density of the deep

25.7 36.4

36.6 > 36.6

26.0 26.5 26.7

Depth (m)

26.8 200 36.2 300

36.0

27.0

35.8 400 35.6

Bojador 26˚N 27.2

500

_

Density anomaly (kg m 3)

Salinity (PSU) 0

80

21

_ 60 _ 40

18 100 _ 20 Depth (m)

16 200 _ 10 300

14

400

12

0

20 km

>0

500 Temperature (˚C)

_

Alongshore velocity (cm s 1)

Figure 5 Sections of temperature, salinity, density, and alongshore current off north-west Africa during favorable winds in August. Contours are elevated from B200 m to the sea surface. The current field is based on a combination of moored current meter (solid dots), profiling current meter (solid vertical lines), and geostrophic estimates (dashed vertical lines).


472


water, and Dr is the density contrast between the surface and deep layers. The characteristic upwelling velocity, w, is given by the upward transport divided by the width of the upwelling zone w ¼ T/l. For a typical situation lB10–20 km, and wB10 m d1. The onshore–offshore flow observed off NW Africa conforms to the classic picture of upwelling (Figure 6) where the surface layer offshore flow is compensated by deeper onshore flow. When the wind relaxes, the whole water column on the shelf moves shoreward, carrying with it warm oceanic waters. Detailed comparisons of the time-varying offshore and onshore transports with the Ekman index have shown good agreement between offshore flow and the index, in this and other upwelling regions. Poorer agreement is found between offshore and onshore transport, which often appear to be locally out of balance. Satellite imagery of sea surface temperature fields shows that the boundary between upwelled waters and open ocean is often strongly contorted into long filaments of cooler water stretching hundreds of kilometers out to sea. Filaments identified off the Iberian and African coasts appear to develop from instabilities of the along-frontal flow, often triggered by coastal or topographic irregularities such as capes or ridges. Some appear associated with offshore eddies, which may themselves be topographically anchored. Filaments tend to recur in the same positions year after year, as, for example, the one off northwest Spain at 421N in Figure 7. They are associated with localized areas of net offshore flow (therefore

local imbalance in the cross-shelf transport) that take the form of a narrow jet where the alongshore flow is diverted seaward. The figure shows a surface drifter following the filament offshore at a mean rate of 0.3 m s1. This offshore transport can help effect exchange of water properties between shelf and deep ocean because of mixing along the filament boundary and interaction with the deeper waters beneath. As upwelled waters move offshore in the filaments, they gradually warm and become indistinguishable from surrounding waters so that any return flow to the coast is not easily identified in satellite images.

Poleward Undercurrent Beneath the near-surface equatorward flow of the Canary Current, a subsurface current flows poleward, counter to the general circulation and tightly bound to the continental slope. It has been documented along the entire continental margin between the Gulf of Guinea and north-west Spain, and is a common feature of all eastern boundaries. Despite many direct and indirect observations of the flow, there are remarkably few systematic observations. The structure of the undercurrent is shown by measurements made close to 201N (Figure 8). Its maximum speed is close to 0.1 m s1 at about 150 m below the surface. The core extends about 300– 400 m vertically and apparently less than 50 km horizontally (although its offshore limit was not directly observed). Above the undercurrent, shallow

0

Depth (m)

20

40

60

27 March 80 _ 20 (A)

_ 10

30 March 0 10 _ E ← Flow (cm s 1) → W

20

_ 20

_ 10

(B)

0 10 _ E ← Flow (cm s 1) → W

20

Figure 6 Profiles of cross-shelf flow measured by current meters (solid dots) in 76 m of water on the African continental shelf during (A) near-zero winds and (B) upwelling favorable wind.



473

Temperature (˚C)

15

17

16

20

19

18

43˚N

Latitude

42.5˚N

42˚N

41.5˚N

41˚N 12˚W

11.5˚W

11˚W

10.5˚W 10˚W Longitude

9.5˚W

9˚W

8.5˚W

Figure 7 Satellite sea surface temperature image in August showing an upwelling filament extending 200 m offshore. Clouds obscure the image near to shore. A surface layer drift buoy traces the current along the filament. The dots mark daily positions starting on 14 August near shore. White curves mark the 50, 100, 200, 500, and 1000 m isobaths.

equatorward flow predominates, while in layers deeper than 500 m, Antarctic Intermediate Water is carried northward at depths around 900 m. Note also the weak undercurrent in Figure 5. Along most of the eastern subtropical gyre the poleward flow is restricted to the subsurface layers, though it may surface when the Trade Winds weaken or turn northward. Off Iberia and south of 201N it appears to extend to the sea surface for more of the annual cycle. In the latter area it forms the inshore loop of the cyclonic recirculation. Where it meets the Canary Current separating from the coast, some of the poleward flow continues northward as the undercurrent, carrying with it the typically warmer, fresher, and higher nutrient content South Atlantic Central Water of this region. The anomalous water can be traced as far as the Canary Islands (281N) before mixing with the surrounding cooler and saltier North Atlantic Central Water dilutes it beyond recognition. The seasonal analysis showed that during autumn the poleward flow may occur at the sea surface, again reaching the Canary Islands, and the few available direct observations appear to corroborate this. Off Iberia, most of the water column flows poleward, although the surface layer is flowing

equatorward above 200 m depth in the long-term mean. During winter, all of the water column moves northward over the continental slope, but in summer, when the equatorward Trade Winds are present, the currents in the upper few hundred meters are driven equatorward. The undercurrent is known to extend deeply off Iberia, where it carries Mediterranean Intermediate Water at levels between 600 and 1500 m depth. This is water that has escaped through the Strait of Gibraltar and is constrained by the Earth’s rotation to flow northward, hugging the continental slope. As it travels northward it tends to separate intermittently from the coast in various locations to form subsurface eddies known as Meddies. Three preferred paths are reported to carry the Mediterranean Water away from the Iberian continental slope: northward where the undercurrent extends beyond Cape Finisterre, north-westward west of the Galicia Bank, and south-westward off the Gorringe Bank. In each case the topographic feature seems to trigger the formation of the Meddies, which then migrate away from the boundary and can maintain their identity for up to 4 years. One unresolved question is whether the poleward flow off NW Africa is in any sense continuous with that off Iberia. There are few observations of the


474


0 _ 30 _ 10 _ 20

_ 10

_5

0

0 +2

100 + 10 > +8

200 Depth (m)

+8

+6

300

+6

+4 +4

400

_ V

_ V

+2

500

30

20 0 10 Distance (km)

30

0 20 10 Distance (km)

Figure 8 Two sections of alongshore flow measured by current meters (solid dots) near 201N off north-west Africa, showing structure of the poleward undercurrent. Speeds are given in cm s1, northward positive.

undercurrent off northern Morocco and none that might indicate how the flow interacts with the deepening Mediterranean Intermediate Water. The latter is dense and sinks from shallow levels on leaving the Strait of Gibraltar to its equilibrium level around 1200 m. It therefore must pass through any undercurrent continuing north from Morocco to the Iberian slope in the Gulf of Cadiz.

Spatial Variability Textbook pictures like Figure 1 tend to show broad currents of weak unidirectional flows. However, measurements almost always indicate currents highly variable in both strength and direction. Figure 9 shows a near-synoptic view of the currents, derived from the combined TOPEX and ERS-1 altimetry on 14 August 1993. The small sea surface height slope anomalies measured by satellites have been added to the mean summer surface elevations calculated with respect to a 400 m reference level from hydrography to compensate in part for the lack of the mean signal in the altimetry. The Azores Current meanders along latitude 321N, gradually turning southward into the

Canary Current. Little flow seems to come from the weak Portugal Current southwards. Near the African coast from 301N the flow is southward as far as 201N, where it turns abruptly offshore on meeting poleward flow from farther south. A large number of eddies are seen throughout the region, especially associated with the Azores Current and farther south. The resolution of the altimetry is limited by the ground track separation (B50 km or more) and the repeat interval (415 days). A survey of currents just south of the Canary Islands made near the same time (10–18 August 1993) shows patterns of flow on shorter scales as complex as in the satellite view (Figure 10). One might ask where the Canary Current is in this complexity. It lies, of course, in the average flow over the area of the survey, about 0.05 m s1 toward the south west as expected. However, the instantaneous current in any location can have almost any direction and speed. The flow field is composed of narrow, strong jets of current, which may meander through the area changing their position with time, and eddy circulations that drift with the background flow. Here the alongshore, equatorward flow appears diverted around a 100 km diameter cyclonic eddy generated in the trough of bottom topography between Gran Canaria, Fuerteventura, and Africa. This eddy was only just resolved in the satellite analysis. This level of detail of field observation has only become available in the last two decades with the introduction of acoustic Doppler methods of determining upper-level currents from a moving research vessel. In this way rapid surveys can be made to reveal the intricate patterns of ocean currents, though still in an area limited in size by ship speed.

Numerical Models Numerical modeling techniques for ocean circulation are improving rapidly as computer power allows more detailed calculations on finer grids. Recent results for the Canary Current region, representing the near-surface flow in early September, are shown in Figure 11. The model realistically depicts a meandering Azores Current at 351N, which separates into branches entering the Gulf of Cadiz and flowing south-east into the Canary Current. Note the southward flow near-shore off NW Africa, typical of late summer upwelling conditions, and the prevalence of meanders and eddies throughout the field. Also noteworthy is the near-shore poleward current off Iberia, similar to conditions observed there in September with the cessation of upwelling and onset of the winter season.



475

40˚N

36˚N

Latitude

32˚N

28˚N

24˚N

20˚N

16˚N 32˚W

28˚W

24˚W

12˚W

16˚W

20˚W Longitude

1ms

8˚W _1

Sea surface height (cm)

20

30

40

50

60

70

80

Latitude

Figure 9 Surface geostrophic currents superimposed on sea surface height fields from TOPEX/ERS-1 altimeter (14 August 1993). The August mean sea surface height calculated from density fields has been added to compensate for lack of altimetric mean.

27˚N

50 cm s

26˚N

_ 16˚W

_ 15˚W

_ 14˚W

_1

_ 13˚W

Longitude Figure 10 Field of current observed by acoustic Doppler current profiler near the time of Figure 9. Sea surface height contours are shown. Note the southward shelf flow turning offshore to describe a cyclonic eddy B100 km diameter.


476


Trade Wind, which varies on timescales of weeks, seasons, decades, and longer. The Trades directly force coastal upwelling and continental shelf currents throughout the region. Unresolved questions include the reality of the intense equatorward flow alongshore between 301 and 201N suggested by the seasonal analyses, and the continuity of the undercurrent along the continental margin.

42˚N

Latitude

35˚N

30˚N

20 cm s 25˚N 25˚W

20˚W

15˚W Longitude

10˚W

_1

5.5˚W

Figure 11 Numerical model results of calculated surface currents for September. (Johnson J and Stevens I 2000 Deep Sea Research I, 47(5): 875–900.)

This ‘state of the art’ numerical model is run using climatic winds, i.e., monthly averaged winds that vary smoothly through the year, and has a limited number of layers in the vertical. Nevertheless, the fine horizontal resolution allows the realistic reproduction of oceanic features often only partially sampled because of ship time and equipment constraints. Even this model is on a coarse scale compared to the size of many important features like islands or coastal capes. Entire islands are represented by one or two model grid points so the level of detail does not yet reproduce features on the scales seen in Figure 10. In the not too distant future, models driven by actual or forecast winds and incorporating ocean observations from monitoring networks will provide ocean forecasts that will rival in accuracy present-day meteorological models.

Conclusions The overall features of the Canary and Portugal currents, such as the mean pattern and seasonal variation, are well established despite relatively few systematic observations. The Canary Current is fed from the Azores Current to a lesser extent from the weak Portugal Current. To north and south of the Canary Current proper, regions of predominantly northward flow persist through most of the year. These are apparently connected by a narrow undercurrent trapped to the continental slope near 300 m depth. Variability of the system is dominated by the

See also Benguela Current. California and Alaska Currents. Ekman Transport and Pumping. Meddies and SubSurface Eddies. Mesoscale Eddies. Regional and Shelf Sea Models. Satellite Altimetry. Satellite Remote Sensing of Sea Surface Temperatures. Upwelling Ecosystems. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Barton ED (1989) The poleward undercurrent on the eastern boundary of the Subtropical North Atlantic. In: Neshyba S, Smith RL, and Mooers CNK (eds.) Poleward Flows Along Eastern Ocean Boundaries, pp. 82--95. (Springer Lecture Note Series). Berlin: Springer-Verlag. Barton ED (1998) Eastern boundary of the North Atlantic: Northwest Africa and Iberia. In: Brink KH and Robinson AR (eds.) The Sea, vol. 11: The Global Coastal Ocean: Regional Studies and Syntheses, ch. 22. New York: Wiley. Brink KH (1997) Wind driven currents over the continental shelf. In: Brink KH and Robinson AR (eds.) The Sea, vol. 10: The Global Coastal Ocean: Processes and Methods, ch. 1. New York: Wiley. Johnson J and Stevens I (2000) A fine resolution model of the eastern North Atlantic between the Azores, the Canary Islands and the Gibraltar Strait. Deep-Sea Research I 47(5): 875--900. Krauss W (ed.) (1996) The Warmwatersphere of the North Atlantic Ocean chs 10–12. Berlin: Borntraeger. Mann KH and Lazier JRN (eds.) (1991) Dynamics of Marine Ecosystems. Boston: Blackwell. Mittelstaedt E (1983) The upwelling area off Northwest Africa – a description of phenomena related to coastal upwelling. Progress in Oceanography 12: 307--331. Stramma L and Siedler G (1988) Seasonal changes in the North Atlantic Subtropical Gyre. Journal of Geophysical Research 93(C7): 8111--8118. Tomczak M and Godfrey JS (1994) Regional Oceanography: An Introduction. Oxford: Elsevier.


CARBON CYCLE C. A. Carlson, University of California, Santa Barbara, CA, USA N. R. Bates, Bermuda Biological Station for Research, St George’s, Bermuda, USA D. A. Hansell, University of Miami, Miami FL, USA D. K. Steinberg, College of William and Mary, Gloucester Pt, VA, USA

the mechanisms of carbon exchange between the ocean and atmosphere; (3) how carbon is redistributed throughout the ocean by ocean circulation; and (4) the roles of the ‘solubility’, ‘biological,’ and ‘carbonate’ pumps in the ocean carbon cycle.

Global Carbon Cycle

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 390–400, & 2001, Elsevier Ltd.

Introduction Why is carbon an important element? Carbon has several unique properties that make it an important component of life, energy flow, and climate regulation. It is present on the Earth in many different inorganic and organic forms. Importantly, it has the ability to form complex, stable carbon compounds, such as proteins and carbohydrates, which are the fundamental building blocks of life. Photosynthesis provides marine plants (phytoplankton) with an ability to transform energy from sunlight, and inorganic carbon and nutrients dissolved in sea water, into complex organic carbon materials. All organisms, including autotrophs and heterotrophs, then catabolize these organic compounds to their inorganic constituents via respiration, yielding energy for their metabolic requirements. Production, consumption, and transformation of these organic materials provide the energy to be transferred between all the trophic states of the ocean ecosystem. In its inorganic gaseous phases (carbon dioxide, CO2; methane, CH4; carbon monoxide, CO), carbon has important greenhouse properties that can influence climate. Greenhouse gases in the atmosphere act to trap long-wave radiation escaping from Earth to space. As a result, the Earth’s surface warms, an effect necessary to maintain liquid water and life on Earth. Human activities have led to a rapid increase in greenhouse gas concentrations, potentially impacting the world’s climate through the effects of global warming. Because of the importance of carbon for life and climate, much research effort has been focused on understanding the global carbon cycle and, in particular, the functioning of the ocean carbon cycle. Biological and chemical processes in the marine environment respond to and influence climate by helping to regulate the concentration of CO2 in the atmosphere. We will discuss (1) the importance of the ocean to the global carbon cycle; (2)

The global carbon cycle describes the complex transformations and fluxes of carbon between the major components of the Earth system. Carbon is stored in four major Earth reservoirs, including the atmosphere, lithosphere, biosphere, and hydrosphere. Each reservoir contains a variety of organic and inorganic carbon compounds ranging in amounts. In addition, the exchange and storage times for each carbon reservoir can vary from a few years to millions of years. For example, the lithosphere contains the largest amount of carbon (1023 g C), buried in sedimentary rocks in the form of carbonate minerals (CaCO3, CaMgCO3, and FeCO3) and organic compounds such as oil, natural gas, and coal (fossil fuels). Carbon in the lithosphere is redistributed to other carbon reservoirs on timescales of millions of years by slow geological processes such as chemical weathering and sedimentation. Thus, the lithosphere is considered to be a relatively inactive component of the global carbon cycle (though the fossil fuels are now being added to the biologically active reservoirs at unnaturally high rates). The Earth’s active carbon reservoirs contain approximately 43 1018 g of carbon, which is partitioned between the atmosphere (750 1015 g C), the terrestrial biosphere (2190 1015 g C), and the ocean (39 973 1015 g C; Figure 1). While the absolute sum of carbon found in the active reservoirs is maintained in near steady state by slow geological processes, more rapid biogeochemical processes drive the redistribution of carbon among the active reservoirs. Human activities, such as use of fossil fuels and deforestation, have significantly altered the amount of carbon stored in the atmosphere and perturbed the fluxes of carbon between the atmosphere, the terrestrial biosphere, and the ocean. Since the emergence of the industrial age 200 years ago, the release of CO2 from fossil fuel use, cement manufacture, and deforestation has increased the partial pressure of atmospheric CO2 from 280 ppm to present day values of 360 ppm; an increase of 25% in the last century (Figure 2). Currently, as a result of human activities, approximately 5.5 1015 g of


477

CARBON CYCLE

Atmosphere 750

0

n5

io irat esp ary

duc

Pro

m Pri

6

1.

Ch an gin g

Vegetation 610

bu r

us e

R

1

5 tion

lan d

478

Fo s

sil

nin

g

fu

el

5.

5

90

92

5

0.

Soil and detritus 1580 50

Biota 3

Surface ocean 1020

40

2

100

DOC < 700

91.6

8 2

Deep ocean 38 100

C in Earth′s crust 90 000 000

0.2 Surface sediments 150

Recoverable fossil fuel 10 000

Figure 1 The global carbon cycle. Arrows indicate fluxes of carbon between the various reservoirs of the atmosphere, lithosphere, terrestrial biosphere, and the ocean. All stocks are expressed as 1015 g C. All fluxes are decadal means and expressed as 1015 g C y1. (Adapted with permission from Sigenthaler and Sarmiento, 1993), copyright 1993, Macmillan Magazines Ltd.). Data used to construct this figure came from Sigenthaler and Sarmiento (1993), Hansell and Carlson (1998), and Sarmiento and Wofsy (1999).

‘anthropogenic’ carbon is added to the atmosphere every year. About half of the anthropogenic CO2 is retained in the atmosphere, while the remaining carbon is transferred to and stored in the ocean and the terrestrial biosphere. Carbon reservoirs that remove and sequester CO2 from the atmosphere are referred to as carbon ‘sinks’. The partitioning of anthropogenic carbon between oceanic and terrestrial sinks is not well known. Quantifying controls on the partitioning is necessary for understanding the dynamics of the global carbon cycle. The terrestrial biosphere may be a significant sink for anthropogenic carbon, but scientific understanding of the causative processes is hindered by the complexity of terrestrial ecosystems. Global ocean research programs such as Geochemical Ocean Sections (GEOSEC), the Joint Global Ocean Flux Study (JGOFS), and the JGOFS/ World Ocean Circulation Experiment (WOCE) Ocean CO2 Survey have resulted in improvements in

our understanding of physical circulation and biological processes of the ocean. These studies have also allowed oceanographers to better constrain the role of the ocean in CO2 sequestration compared to terrestrial systems. Based on numerical models of ocean circulation and ecosystem processes, oceanographers estimate that 70% (2 1015 g C) of the anthropogenic CO2 is absorbed by the ocean each year. The fate of the remaining 30% (0.75 1015 g) of anthropogenic CO2 is unknown. Determining the magnitude of the oceanic sink of anthropogenic CO2 is dependent on understanding the interplay of various chemical, physical, and biological factors.

Oceanic Carbon Cycle The ocean is the largest reservoir of the Earth’s active carbon, containing 39 973 1015 g C. Oceanic carbon occurs as a variety of inorganic and organic


CARBON CYCLE

380

PCO2 (ppmv)

360

Siple Mauna Loa 100-year running mean

340

320

300 280 1850

1900

1950

2000

Year Figure 2 Atmospheric CO2 concentrations from 1850 to 1996. These data illustrate an increase in atmospheric CO2 concentration from pre-industrial concentration of 280 ppmv to present-day concentrations of 360 ppmv. Human activities of fossil fuel burning and deforestation have caused this observed increase in atmospheric CO2. (Adapted from Houghton et al. (1996) with permission from Intergovernmental Panel on Climate Change (IPCC). The original figure was constructed from Siple ice core data and (from 1958) data collected at the Mauna Loa sampling site.)

forms, including dissolved CO2, bicarbonate 2 (HCO 3 ), carbonate (CO3 ) and organic compounds. CO2 is one of the most soluble of the major gases in sea water and the ocean has an enormous capacity to buffer changes in the atmospheric CO2 content. The concentration of dissolved CO2 in sea water is relatively small because CO2 reacts with water to form a weak acid, carbonic acid (H2CO3), which rapidly dissociates (within milliseconds) to form 2 HCO 3 and CO3 (eqn [I]). CO2 ðgasÞ þ H2 O"H2 CO3 ðaqÞ"Hþ ðaqÞ 2 þ þHCO 3 ðaqÞ"2H ðaqÞ þ CO3 ðaqÞ

½I

For every 20 molecules of CO2 absorbed by the ocean, 19 molecules are rapidly converted to HCO 3 and CO2 3 ; at the typical range of pH in sea water (7.8–8.2; see below), most inorganic carbon is found in the form of HCO 3 . These reactions (eqn [I]) provide a chemical buffer, maintain the pH of the ocean within a small range, and constrain the amount of atmospheric CO2 that can be taken up by the ocean. The amount of dissolved CO2 in sea water cannot be determined analytically but can be calculated after measuring other inorganic carbon species. Dissolved inorganic carbon (DIC) refers to the total amount of 2 CO2, HCO 3 plus CO3 in sea water, while the partial pressure of CO2 (PCO2) measures the contribution of CO2 to total gas pressure. The alkalinity of sea water (A) is a measure of the bases present in sea

479

2 water, consisting mainly of HCO 3 and CO3 2 (A[HCO3 ] þ 2[CO3 ]) and minor constituents such as borate (BO4) and hydrogen ions (Hþ). Changes in DIC concentration and alkalinity affect the solubility of CO2 in sea water (i.e., the ability of sea water to absorb CO2) (see below). The concentrations of inorganic carbon species in sea water are controlled not only by the chemical reactions outlined above (i.e., eqn [I]) but also by various physical and biological processes, including the exchange of CO2 between ocean and atmosphere; the solubility of CO2; photosynthesis and respiration; and the formation and dissolution of calcium carbonate (CaCO3). Typical surface sea water ranges from pH of 7.8 to 8.2. On addition of acid (i.e., Hþ), the chemical reactions shift toward a higher concentration of CO2 in sea water (eqn [IIa]) and pH decreases from 8.0 to 7.8 and then pH will rise from 8.0 to 8.2.

Hþ þ HCO 3 -H2 CO3 -CO2 ðaqÞ þ H2 O

½IIa

If base is added to sea water (eqn [IIb]), then pH will rise. H2 CO3 -Hþ þ HCO3

½IIb

Solubility and Exchange of CO2 between the Ocean and Atmosphere The solubility of CO2 in sea water is an important factor in controlling the exchange of carbon between the ocean and atmosphere. Henry’s law (eqn [1]) describes the relationship between solubility and sea water properties, where S equals the solubility of gas in liquid, k is the solubility constant (k is a function mainly of temperature) and P is the overlying pressure of the gas in the atmosphere. S ¼ kP

½1

Sea water properties such as temperature, salinity, and partial pressure of CO2 determine the solubility of CO2. For example, at 01C in sea water, the solubility of CO2 is double that in sea water at 201C; thus colder water will tend to absorb more CO2 than warmer water. Henry’s law also describes the relationship between the partial pressure of CO2 in solution (PCO2) and its concentration (i.e., [CO2]). Colder waters tend to have lower PCO2 than warmer waters: for every 11C temperature increase, sea water PCO2 increases by B4%. Sea water PCO2 is also influenced by complicated thermodynamic relationships


480

CARBON CYCLE

Alkalinity

DIC

p CO2

Temperature

DIC

Alkalinity

p CO2

Temperature

Alkalinity

DIC

p CO2

Temperature

Figure 3 The response of PCO2 to changes in the sea water properties of (A) DIC concentration, (B) alkalinity and (C) temperature. Each panel describes how PCO2 will respond to the changes in the relevant sea water property. The blue arrows illustrate the response of PCO2 to an increase in the sea water property and the red arrows illustrate the response to a decrease in the sea water property. For example, as DIC or temperature increases, PCO2 increases; whereas an increase in alkalinity results in a decrease in PCO2.

between the different carbon species. For example, a decrease in sea water DIC or temperature acts to decrease PCO2, while a decrease in alkalinity acts to increase PCO2 (Figure 3). Carbon dioxide is transferred across the air–sea interface by molecular diffusion and turbulence at the ocean surface. The flux (F) of CO2 between the atmosphere and ocean is driven by the concentration difference between the reservoirs (eqn [2]). F ¼ DPco2 KW

½2

In eqn [2] DPCO2 is the difference in PCO2 between the ocean and atmosphere and Kw is the transfer coefficient across the air–sea interface, termed the piston velocity. In cold waters, sea water PCO2 tends to be lower than atmospheric PCO2, thus driving the direction of CO2 gas exchange from atmosphere to ocean (Figure 3). In warmer waters, sea water PCO2 is greater than atmospheric PCO2, and CO2 gas exchange occurs in the opposite direction, from the ocean to the atmosphere. The rate at which CO2 is transferred between the ocean and the atmosphere depends not only on the PCO2 difference but on turbulence at the ocean surface. The piston velocity of CO2 is related to solubility and the strength of the wind blowing on the sea surface. As wind speed increases, the rate of air– sea CO2 exchange also increase. Turbulence caused by breaking waves also influences gas exchange because air bubbles may dissolve following entrainment into the ocean mixed layer. Ocean Structure

Physically, the ocean can be thought of as two concentric spheres, the surface ocean and the deep

ocean, separated by a density discontinuity called the pycnocline. The surface ocean occupies the upper few hundred meters of the water column and contains approximately 1020 1015 g C of DIC (Figure 1). The absorption of CO2 by the ocean through gas exchange takes place in the mixed layer, the upper portion of the surface ocean that makes direct contact with the atmosphere. The surface ocean reaches equilibrium with the atmosphere within one year. The partial pressure of CO2 in the surface ocean is slightly less than or greater than that of the atmosphere, depending on the controlling variables as described above, and varies temporally and spatially with changing environmental conditions. The deeper ocean represents the remainder of the ocean volume and is supersaturated with CO2, with a DIC stock of 38 100 1015 g C (Figure 1), or 50 times the DIC contained in the atmosphere. CO2 absorbed by the ocean through gas exchange has a variety of fates. Physical and biological mechanisms can return the CO2 back to the atmosphere or transfer carbon from the surface ocean to the deep ocean and ocean sediments through several transport processes termed the ‘solubility’, ‘biological’, and ‘carbonate’ pumps. The Solubility Pump, Oceanic Circulation, and Carbon Redistribution

The ‘solubility pump’ is defined as the exchange of carbon between the atmosphere and the ocean as mediated by physical processes such as heat flux, advection and diffusion, and ocean circulation. It assists in the transfer of atmospheric CO2 to the deep ocean. This transfer is controlled by circulation patterns of the surface ocean (wind-driven


CARBON CYCLE

circulation) and the deep ocean (thermohaline circulation). These circulation patterns assist in the transfer of atmospheric CO2 to the deep ocean and help to maintain the vertical gradient of DIC found in the ocean (Figure 4). The ability of the ocean to take up anthropogenic CO2 via the solubility pump is limited by the physical structure of the ocean, the distribution of oceanic DIC, ocean circulation patterns, and the exchange between the surface and deep ocean layers. To be an effective sink for anthropogenic carbon, CO2 must be transferred to the deep ocean by mixing and biological processes (see below). Wind-driven circulation occurs as a consequence of friction and turbulence imparted by wind blowing over the sea surface. This circulation pattern is primarily horizontal in movement and is responsible for transporting warm water from lower latitudes (warm) to higher latitudes (cold). Surface currents move water and carbon great distances within ocean DIC (μmol kg–1) 0

Biological and carbonate pump

Depth (m)

100

500

Solubility pump

1000 Figure 4 Illustration of the vertical gradient of DIC in the ocean. The uptake of DIC by phytoplankton and conversion into sinking organic matter (‘biological pump’; gray arrow) and sinking calcium carbonate skeletal matter (‘carbonate pump’; gray arrow) contributes to the maintenance of the vertical gradient. Introduction of DIC to the deep waters via the ‘solubility pump’ at high latitudes and subsequent deep water formation also helps maintain this vertical gradient (black arrow; see Figure 5).

481

basins on timescales of months to years. As surface sea water moves from low latitudes to high latitudes, the increasing solubility of CO2 in the sea water (due to sea surface cooling) allows atmospheric CO2 to invade the surface mixed layer (Figure 5 and 6A). Exchange of surface waters with the deep ocean through wind-driven mixing is limited because of strong density stratification of the water column over the majority of the world’s oceans. However, thermohaline (overturning) circulation at high latitudes provides a mechanism for surface waters to exchange with the deep ocean. Passage of cold and dry air masses over high-latitude regions, such as the Greenland and Labrador Seas in the North Atlantic or the Weddell Sea in the Southern Ocean, forms cold and very dense sea water (‘deep water’ formation). Once formed, these dense water masses sink vertically until they reach a depth at which water is of similar density (i.e., 2000–4000 m deep). Following sinking, the dense waters are transported slowly throughout all of the deep ocean basins by advection and diffusion, displacing other deep water that eventually is brought back to the surface by upwelling (Figure 5). Because of the smaller volume and faster circulation, the residence time of the surface ocean is only one decade compared to 600–1000 years for the deep ocean. The process of deep water formation transfers CO2, absorbed from the atmosphere by the solubility pump, into the deep ocean. The effect is that DIC concentration increases with depth in all ocean basins (Figure 4, 5 and 6A,). As a result of the long residence time of the deep ocean, carbon, once removed from the surface ocean to the deep ocean through the effects of solubility and deep water formation, is stored without contact with the atmosphere for hundreds to thousands of years. At present, deep water formed at the surface that is in equilibrium with the atmosphere (sea water PCO2 of B360 ppm), carries more CO2 to depth than deep water formed prior to the industrial age (e.g., B280 ppm). Furthermore, PCO2 of upwelled deep water is less than that in the recently formed deep water, indicating that the deep water formation and the ‘solubility pump’ allow the ocean to be a net sink for anthropogenic CO2. The vertical gradient in DIC (Figure 4) and the ability of the ocean to take up atmospheric CO2 is augmented by biological processes known as the ‘biological pump’.

The Biological Pump Although the standing stock of marine biota in the ocean is relatively small (3 1015 g C), the activity


482

CARBON CYCLE

CO2 escapes

Wind stress and cooling

CO2 invades

Pole

Equator

Upper ocean layer

Pycnocline

SINKING

Deep ocean layer

Figure 5 Conceptual model of the ‘solubility pump’. White arrows represent movement of water; black arrows represent movement of CO2 within, and into and out of, the ocean. Cooling increases the solubility of CO2 and results in a flux of CO2 from the atmosphere to the surface ocean. At subpolar latitudes the water density increases and the CO2-enriched water sinks rapidly. At depth, the CO2enriched water moves slowly as is it is dispersed throughout the deep ocean. The sinking water displaces water that is returned to the surface ocean in upwelling regions. As the water warms, PCO2 increases, resulting in escape of CO2 from the surface water to the atmosphere.

associated with the biota is extremely important to the cycling of carbon between the atmosphere and the ocean. The largest and most rapid fluxes in the global carbon cycle are those that link atmospheric CO2 to photosynthetic production (primary production) on the land and in the ocean. Globally, marine phytoplankton are responsible for more than one-third of the total gross photosynthetic production (50 1015 g C y1). In the sea, photosynthesis is limited to the euphotic zone, the upper 100–150 m of the water column where light can penetrate. Photosynthetic organisms use light energy to reduce CO2 to high-energy organic compounds. In turn, a portion of these synthesized organic compounds are utilized by heterotrophic organisms as an energy source, being remineralized to CO2 via respiration. Eqn [III] represents the overall reactions of photosynthesis and respiration. Light energy (photosynthesis)

CO2(gas) ⫹ H2O

(CH2O)n Metabolic energy (respiration)

½III

⫹ O2(gas)

In the sea, net primary production (primary production in excess of respiration) converts CO2 to

organic matter that is stored as particulate organic carbon (POC; in living and detrital particles) and as dissolved organic carbon (DOC). In stratified regions of the ocean (lower latitudes), net primary production results in a drawdown of DIC and an accumulation of organic matter as POC and DOC (Figure 6B). However, it is the portion of organic carbon production that can be exported from the surface ocean and remineralized in the deep ocean that is important in the exchange of CO2 between the atmosphere and the ocean. The biological pump refers to the processes that convert CO2 (thereby drawing down DIC) to organic matter by photosynthesis, and remove the organic carbon to depth (where it is respired) via sinking, mixing, and active transport mechanisms (Figure 7). Once at great depth, it is effectively removed from exchange with the atmosphere. As living biomass is produced, some particles becomes senescent and form sinking aggregates, while other particles are consumed by herbivores and sinking fecal pellets (POC) are formed. These sinking aggregates and pellets remove carbon from the surface to be remineralized at depth via decomposition by bacteria or consumption by zooplankton and fish (Figure 7). In addition, DOC produced by phytoplankton or by animal excretion


CARBON CYCLE

400

2100

600

2000

800

DIC (μmol kg–1)

2200

200 Depth (m)

483

1800 1000 (A)

10°S

20°S

30°S

40°S

50°S

60°S

Depth (m)

200

70

400 60 600 50

800

40

1000 (B)

DOC (μmol kg–1)

80

10°S

20°S

30°S

40°S

50°S

60°S

Latitude Figure 6 Contour plot of (A) DIC and (B) DOC along a transect line in the South Pacific between the equator (01; 1701 W) and the Antarctic Polar Front (661 S; 1701 W). Note that in the low-latitude stratified waters DIC concentrations are depleted in surface water relative to deep water, as a result of net primary production and air–sea exchange. DOC concentrations are elevated relative to deep water. In high-latitude regions, DIC concentration are elevated in the surface water as a result of increased solubility of cooler surface waters.

in surface waters can also be transported downward by subduction or convective mixing of surface waters (Figure 7). Finally, vertically migrating zooplankton that feed in the surface waters at night and return to deep waters during the day actively transport dissolved and particulate material to depth, where a portion is metabolized (Figure 7). Production via photosynthesis can occur only in the surface ocean, whereas remineralization can occur throughout the water column. The biological pump serves to spatially separate the net photosynthetic from net respiratory processes. Thus, the conversion of DIC to exportable organic matter acts to reduce the DIC concentration in the surface water and its subsequent remineralization increases DIC concentration in the deep ocean (Figure 6). The biological pump is important to the maintenance of a vertical DIC profile of undersaturation in the surface and supersaturation at depth (Figure 4 and 5A). Undersaturation of DIC in the surface mixed layer, created by the biological pump, allows for the influx

of CO2 from the atmosphere (see Henry’s law above; Figure 7). Gross export of organic matter out of the surface waters is approximately 10 1015 g C y1 (Figure 1). Less than 1% of the organic matter exported from the surface waters is stored in the abyssal sediment. In fact, most of the exported organic matter is remineralized to DIC in the upper 500 m of the water column. It is released back to the atmosphere on timescales of months to years via upwelling, mixing, or ventilation of high-density water at high latitudes. It is that fraction of exported organic matter that actually reaches the deep ocean (>1000 m) that is important for long-term atmospheric CO2 regulation. Once in the deep ocean, the organic matter either remains as long-lived DOC or is remineralized to DIC and is removed from interaction with the atmosphere on timescales of centuries to millennia. Thus, even though less than 1% of the exported carbon is stored in marine sediments, the activities of the biological pump are very important in mediating


484

CARBON CYCLE

CO2

Fixation of carbon by phytoplankton

Respiration

Grazing

Excretion

2

Aggregate formation

Physical mixing of DOC

Egestion

3

Break up Base of euphotic zone 1

Active vertical migration

Passive sinking of POC, PIC Consumption, repackaging

Respiration

(Zooplankton)

Excretion

Decomposition (Bacteria)

Seabed Figure 7 Conceptual diagram depicting components of the ‘biological pump’. CO2 is taken up by phytoplankton and organic matter is produced. As this organic matter is processed through the marine food web, fecal pellets or aggregates are produced, a portion of which sink from the surface waters to depth (1). As organic matter is processed through the food web, DOC is also produced. DOC is removed from the surface waters to depth via physical mixing of the water by convective overturn (2). DOC and DIC are also actively transported to depth by vertically migrating organisms such as copepods that feed in surface waters and excrete and respire the consumed organic carbon at depth (3).

the air–sea transfer of CO2. Without this pump in action, atmospheric CO2 concentration might be as high as 500 to 1000 ppm versus the 360 ppm observed today. Contribution of POC Versus DOC in the Biological Pump

Historically, sinking particles were thought of as the dominant export mechanism of the biological pump and the primary driver of respiration in the ocean’s interior. However, downward mixing of surface water can also transport large quantities of DOC trapped within the sinking water mass. In order for DOC to be an important contributor to the biological pump two sets of conditions must exist. First, the producer–consumer dynamics in the surface waters must yield DOC of a quality that is resistant to rapid remineralization by bacteria and lead to net DOC production. Second, the physical system must undergo periods of deep convective mixing or

subduction in order to remove surface waters and DOC to depth. Although approximately 80% of the globally exported carbon is in the form of POC, DOC can represent 30–50% of carbon export in the upper 500 m of the water column at specific ocean sites. The biological/physical controls on DOC export are complex and are currently being assessed for various regions of the worlds’ ocean. Factors That Affect the Efficiency of the Biological Pump

An efficient biological pump means that a large fraction of the system’s net production is removed from the surface waters via export mechanisms. Factors that affect the efficiency of the biological pump are numerous and include nutrient supply and plankton community structure. Nutrient supply Does an increased partial pressure of atmospheric CO2 lead to a more


CARBON CYCLE

efficient biological pump? Not necessarily, since net primary production is limited by the availability of other inorganic nutrients such as nitrogen, phosphorus, silicon and iron. Because these inorganic nutrients are continuously being removed from the surface waters with vertical export of organic particles, their concentrations are often below detection limits in highly stratified water columns. As a result, primary production becomes limited by the rate at which these nutrients can be re-supplied to the surface ocean by mixing, by atmospheric deposition, or by heterotrophic recycling. Primary production supported by the recycling of nutrients in the surface ocean is referred to as ‘regenerative’ production and contributes little to the biological pump. Primary production supported by the introduction of new nutrients from outside the system, via mixing from below or by atmospheric deposition (e.g. dust), is referred to as ‘new’ production. New nutrients enhance the amount of net production that can be exported (new production). Because CO2 is not considered to

485

be a limiting nutrient in marine systems, the increase in atmospheric CO2 is not likely to stimulate net production for most of the world’s ocean unless it indirectly affects the introduction of new nutrients as well. Community structure Food web structure also plays an important role in determining the size distribution of the organic particles produced and whether the organic carbon and associated nutrients are exported from or recycled within the surface waters. The production of large, rapidly settling cells will make a greater contribution to the biological pump than the production of small, suspended particles. Factors such as the number of trophic links and the size of the primary producers help determine the overall contribution of sinking particles. The number of trophic steps is inversely related to the magnitude of the export flux. For example, in systems where picoplankton are the dominant primary producers there may be 4–5 steps before reaching a trophic level capable of producing

Simplified Carbonate Pump CO2

Aggregation

Calcification

–

HCO3

CO2

CaCO3

Base of euphotic zone Upwelling

CO2

Passive sinking of CaCO3

CaCO3

–

HCO3

1–4 km

Dissolution of CaCO3

Seabed

Figure 8 Conceptual diagram of a simplified ‘carbonate pump’. Some marine organisms form calcareous skeletal material, a portion of which sinks as calcium carbonate aggregates. These aggregates are preserved in shallow ocean sediments or dissolve at greater depths (3000–5000 m), thus increasing DIC concentrations in the deep ocean. The calcium and bicarbonate are returned to the surface ocean through upwelling.


486

CARBON CYCLE

sinking particles. With each trophic transfer, a percentage (50–70%) of the organic carbon is respired, so only a small fraction of the original primary production forms sinking particles. Although picoplankton may dominate primary production in oceanic systems, their production is considered ‘regenerative’ and contributes little to the production of sinking material. Alternatively, production by larger phytoplankton such as diatoms (>20 mm in size) may represent a smaller fraction of primary production, but their contribution to the biological pump is larger because fewer trophic steps are taken to produce sinking particles. The Carbonate Pump

A process considered part of the biological pump (depending how it is defined) is the formation and sinking of calcareous skeletal material by some marine phytoplankton (e.g., coccolithophores) and animals (e.g., pteropods and foraminifera). Calcification is the process by which marine organisms combine calcium with carbonate ions to form hard body parts. The resulting calcium carbonate (CaCO3) is dense and sinks out of the surface water with export production (Figure 8). The global mean ratio for carbon sinking from the surface ocean as CaCO3 or organic carbon is 1:4. However, unlike organic matter, CaCO3 is not remineralized as it sinks; it only begins to dissolve in intermediate and deep waters, waters undersaturated with respect to CaCO3. Complete dissolution of CaCO3 skeletons typically occurs at depths of 1–4 km (in the north Pacific Ocean) to 5 km (in the North Atlantic). This depth zone is known as the carbonate compensation depth. CaCO3 is only found in sediments shallower than the carbonate compensation depth. Globally, the CO2 sink in sedimentary rock is four times greater than the sink in organic sediments.

Summary In summary, the biological and physical processes of the oceanic carbon cycle play an important role in the regulation of atmospheric CO2. However, the intricacies of the oceanic carbon cycle are vast and continued ocean research is essential to better understand the controls of the Earth’s climate.

See also Atmospheric Input of Pollutants. Carbon Dioxide (CO2) Cycle. Network Analysis of Food Webs.

Further Reading Bates NR, Michaels AF, and Knap AH (1996) Seasonal and interannual variability of oceanic carbon dioxide species at the U.S. JGOFS Bermuda Atlantic Time-series Study (BATS) site. Deep-Sea Research II 43: 347--383. Bolin B (ed.) (1983) The Major Biogeochemical Cycles and Their Interactions: SCOPE 21. New York: Wiley. Carlson CA, Ducklow HW, and Michaels AF (1994) Annual flux of dissolved organic carbon from the euphotic zone in the northwestern Sargasso Sea. Nature 371: 405--408. Denman K, Hofman H, and Marchant H (1996) Marine biotic responses to environmental change and feedbacks to climate. In: Houghton JT, Meira Filho LG, and Callander BA, et al. (eds.) Climate Change 1995: The Science of Climate Change. New York: Cambridge University Press. Follows MJ, Williams RG, and Marshall JC (1996) The solubility pump of carbon in the subtropical gyre of the North Atlantic. Journal of Marine Research 54: 605--630. Hansell DA and Carlson CA (1998) Net community production of dissolved organic carbon. Global Biogeochemical Cycles 12: 443--453. Holmeń K (1992) The global carbon cycle. In: Butcher SS, Charlson RJ, Orians GH, and Wolfe GV (eds.) Global Biogeochemical Cycles, pp. 239--262. New York: Academic Press. Houghton JT, Meira Filho LG, and Callander BA, et al. (eds.) (1996) Climate Change 1995: The Science of Climate Change. New York: Cambridge University Press. Michaels AF and Silver MW (1988) Primary producers, sinking fluxes and the microbial food web. Deep-Sea Research 35: 473--490. Sarmiento JL and Wofsy (eds.) (1999) A U.S. Carbon Cycle Science Plan. Washington, DC: U.S. Global Change Research Program. Sarmiento JL, Hughes TMC, Stouffer RJ, and Manabe S (1998) Simulated response of the ocean carbon cycle to anthropogenic climate warming. Nature 393: 245--249. Schlesinger WH (1997) Biogeochemistry: An Analysis of Global Change. New York: Academic Press. Siegenthaler U and Sarmiento JL (1993) Atmospheric carbon dioxide and the ocean. Nature 365: 119--125. Steinberg DK, Carlson CA, Bates NR, Goldthwait SA, Madin LP, and Michaels AF (2000) Zooplankton vertical migration and the active transport of dissolved organic and inorganic carbon in the Sargasso Sea. DeepSea Research I 47: 137--158. Takahashi T, Tans PP, and Fung I (1992) Balancing the budget: carbon dioxide sources and sinks, and the effect of industry. Oceanus 35: 18--28. Varney M (1996) The marine carbonate system. In: Summerhayes CP and Thorpe SA (eds.) Oceanography an Illustrated Guide, pp. 182--194. London: Manson Publishing.


CARBON DIOXIDE (CO2) CYCLE T. Takahashi, Lamont Doherty Earth Observatory, Columbia University, Palisades, NY, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 400–407, & 2001, Elsevier Ltd.

Introduction The oceans, the terrestrial biosphere, and the atmosphere are the three major dynamic reservoirs for carbon on the earth. Through the exchange of CO2 between them, the atmospheric concentration of CO2 that affects the heat balance of the earth, and hence the climate, is regulated. Since carbon is one of the fundamental constituents of living matter, how it cycles through these natural reservoirs has been one of the fundamental questions in environmental sciences. The oceans contain about 50 times as much carbon (about 40 000 Pg-C or1015 g as carbon) as the atmosphere (about 750Pg-C). The terrestrial biosphere contains about three times as muchcarbon (610 Pg-C in living vegetation and 1580 Pg-C in soil organic matter) as the atmosphere. The air–sea exchange of CO2 occurs via gas exchange processes across the sea surface; the natural air-to-sea and sea-to-air fluxes have been estimated to be about 90 Pg-C y1 each. The unperturbed uptake flux of CO2 by global terrestrial photosynthesis is roughly balanced with the release flux by respiration, and both have been estimated to be about 60 Pg-C y1. Accordingly, atmospheric CO2 is cycled through the ocean and terrestrial biosphere with a time scale of about 7 years. The lithosphere contains a huge amount of carbon (about 100 000 000 Pg-C) in the form of limestones ((Ca, Mg) CO3), coal, petroleum, and other formsof organic matter, and exchanges carbon slowly with the other carbon reservoirs via such natural processes as chemical weathering and burial of carbonate andorganic carbon. The rate of removal of atmospheric CO2 by chemical weathering has been estimated to be of the order of 1 Pg-Cy1. Since the industrial revolution in the nineteenth century, the combustion of fossil fuels and the manufacturing of cement have transferred the lithospheric carbon into the atmosphere at rates comparable to the natural CO2 exchange fluxes between the major carbon reservoirs, and thus have perturbed the natural balance significantly (6 Pg-C y1 is about an order of magnitude less than the natural exchanges with the oceans (90 Pg-C y1 and land (60 Pg-C y1)).

The industrial carbon emission rate has been about 6 Pg-C y1 for the 1990s, and the cumulative industrial emissions since the nineteenth century to the end of the twentieth century have been estimated to be about 250 Pg-C. Presently, the atmospheric CO2 content is increasing at a rate ofabout 3.5 Pg-C y1 (equivalent to about 50% of theannual emission) and the remainder of the CO2 emitted into the atmosphere is absorbed by the oceans and terrestrial biosphere in approximately equal proportions. These industrial CO2 emissions have caused the atmospheric CO2 concentration to increase by as much as 30% from about 280 ppm (parts per million mole fraction in dry air) in the pre-industrial year 1850 to about 362 ppm in the year 2000. Theatmospheric CO2 concentration may reach 580 ppm, double thepre-industrial value, by the mid-twenty first century. This represents a significant change that is wholly attributable to human activities on the Earth. It is well known that the oceans play an important role in regulating our living environment by providing water vapor into the atmosphere and transporting heat from the tropics to high latitude areas. In addition to these physical influences, the oceans partially ameliorate the potential CO2-induced climate changes by absorbing industrial CO2 in the atmosphere. Therefore, it is important to understand howthe oceans take up CO2 from the atmosphere and how they store CO2 in circulating ocean water. Furthermore, in order to predict the future course of the atmospheric CO2 changes, we need to understand how the capacity of the ocean carbon reservoir might be changed in response to the Earth’s climate changes, that may, in turn, alter the circulation of ocean water. Since the capacity of the ocean carbon reservoir is governed by complex interactions of physical, biological, and chemical processes, it is presently not possible to identify and predict reliably various climate feedback mechanisms that affect the ocean CO2 storage capacity. Units

In scientific and technical literature, the amount of carbon has often been expressed in three different units: giga tons of carbon (Gt-C), petagrams of carbon (Pg-C) and moles of carbon or CO2. Their relationships are: 1Gt-C ¼ 1 Pg-C ¼ 1 1015 g of carbon ¼ 1000 million metric tonnes of carbon ¼ (1/12) 1015 moles of carbon. The equivalent quantity as CO2 may be obtained by multiplying the above numbers by 3.67 ( ¼ 44/12 ¼ the molecular weight of CO2 divided by the atomic weight of carbon).


487

488

CARBON DIOXIDE (CO2) CYCLE

The magnitude of CO2 disequilibrium between the atmosphere and ocean water is expressed by the difference between the partial pressure of CO2 of ocean water, (pCO2)sw, and that in the overlying air, (pCO2)air. This difference represents the thermodynamic driving potential for CO2 gas transfer across the sea surface. The pCO2 in the air may be estimated using the concentration of CO2 in air, that is commonly expressed in terms of ppm (parts per million) in mole fraction of CO2 in dry air, in the relationship:

equatorial Pacific was identified as a major CO2 source area. The GEOSECS Program of the International Decade of Ocean Exploration, 1970–80, produced a global data set that began to show systematic patterns for the distribution of CO2 sink and source areas over the global oceans.

Methods The net flux of CO2 across these a surface, Fs-a, may be estimated by:

pðCO2 Þair ¼ ðCO2 conc:Þair ðPb pH2 OÞ ½1 where Pb is the barometric pressure and pH2O is the vapor pressure of water at the sea water temperature. The partial pressure of CO2 in sea water, (pCO2)sw, may be measured by equilibration methods or computed using thermodynamic relationships. The unit of microatmospheres (matm) or 106 atm is commonly used in the oceanographic literature.

History The air–sea exchange of CO2was first investigated in the 1910s through the 1930s by a group of scientists including K. Buch, H. Wattenberg, and G.E.R. Deacon. Buch and his collaborators determined in land-based laboratories CO2 solubility, the dissociation constants for carbonic and boric acids in sea water, and their dependence on temperature and chlorinity (the chloride ion concentrationin sea water). Based upon these dissociation constants along with the shipboard measurements of pH and titration alkalinity, they computed the partial pressure of CO2 in surface ocean waters. The Atlantic Ocean was investigated from the Arctic to Antarctic regions during the period 1917–1935, especially during the METEOR Expedition 1925–27, in the North and South Atlantic. They discovered that temperate and cold oceans had lower pCO2 than air (hence the sea water was a sink for atmospheric CO2), especially during spring and summer seasons, due to the assimilation of CO2 by plants. They also observed that the upwelling areas of deep water (such as African coastal areas) had greater pCO2 than the air (hence the sea water was a CO2 source) due to the presence of respired CO2 in deep waters. With the advent of the high-precision infrared CO2 gas analyzer, a new method for shipboard measurements of pCO2 in sea water and in air was introduced during the International Geophysical Year, 1956–59. The precision of measurements was improved by more than an order of magnitude. The global oceans were investigated by this new method, which rapidly yielded high precision data. The

Fs-a ¼ E ½ðpCO2 Þsw ðpCO2 Þair ¼ k a ½ðpCO2 Þsw ðpCO2 Þair

½2

where E is the CO2 gas transfer coefficient expressed commonly in (moles CO2/m2/y/uatm); k is the gas transfer piston velocity (e.g. in (cmh1)) and a is the solubility of CO2 in sea water at a given temperature and salinity (e.g. (moles CO2 kg-sw1 atm1)). If (pCO2)swo(pCO2) air, the net flux of CO2 is from the sea to the air and the ocean is a source of CO2; if (pCO2)swo(pCO2) air, the ocean water is a sink for atmospheric CO2. The sea–air pCO2 difference may be measured at sea and a has been determined experimentally as a function of temperature and salinity. However, the values of E and k that depend on the magnitude of turbulence near the air–water interface cannot be simply characterized over complex ocean surface conditions. Nevertheless, these two variables have been commonly parameterized in terms of wind speed over the ocean. A number of experiments have been performed to determine the wind speed dependence under various wind tunnel conditions as well as ocean and lake environments using different nonreactive tracer gases such as SF6 and 222Rn. However, the published results differ by as much as 50% over the wind speed range of oceanographic interests. Since 14C is in the form of CO2 in the atmosphere and enters into the surface ocean water as CO2 in a timescale of decades, its partition between the atmosphere and the oceans yields a reliable estimate for the mean CO2 gas transfer rate over the global oceans. This yields a CO2 gas exchange rate of 2073 mol CO2 m 2 y1 that corresponds to a sea–air CO2 transfer coefficient of 0.067 mol CO2 m2 y1 uatm1. Wanninkhof in 1992 presented an expression that satisfies the mean global CO2 transfer coefficient based on 14C and takes other field and wind tunnel results into consideration. His equation for variable wind speed conditions is: k cm h1 ¼ 0:39 ðuav Þ2 ðSc=660Þ0:5


½3


where uav is the average wind speed in ms1 corrected to 10 m above sea surface; Sc(dimensionless) is the Schmidt number (kinematic viscosity ofwater)/ (diffusion coefficient of CO2 gas inwater); and 660 represents the Schmidt number for CO2 in seawater at 201C. In view of the difficulties in determining gas transfer coefficients accurately, direct methods for CO2 flux measurements aboard the ship are desirable. Sea– air CO2 flux was measured directly by means of the shipboard eddy-covariance method over the North Atlantic Ocean by Wanninkhof and McGillis in 1999. The net flux of CO2 across the sea surface was determined by a covariance analysis of the tri-axial motion of air with CO2 concentrations in the moving air measured in short time intervals (Bms) as a ship moved over the ocean. The results obtained over awind speed range of 2–13.5 m s1 are consistent with eqn [3] within about 720%. If the data obtainedin wind speeds up to 15 m s1 are taken into consideration, they indicate that the gas transfer piston velocity tends to increase as a cubeof wind speed. However, because of a large scatter (735%) of the flux values at high wind speeds, further work is needed to confirm the cubic dependence. In addition to the uncertainties in the gas transfer coefficient (or piston velocity), the CO2 fluxestimated with eqn [2] is subject to errors in (pCO2)sw caused by the difference between the bulk water temperature and the temperature of the thin skin of ocean water at the sea–air interface. Ordinarily the (pCO2)sw is obtained at the bulk seawater temperature, whereas the relevant value for the flux calculation is (pCO2)sw at the ‘skin’temperature, that depends on the rate of evaporation, the incoming solar radiation, the wind speed, and the degree of turbulence near the interface. The ‘skin’ temperature is often cooler than the bulk water temperatureby as much as 0.51C if the water evaporates rapidly to a dry air mass, but is not always so if a warm humid air mass covers over the ocean. Presently, the time–space distribution of the ‘skin’ temperature is not well known. This, therefore, could introduce errors in (pCO2)sw up to about 6 matm or 2%.

CO2 Sink/Source Areas of the Global Ocean The oceanic sink and source areas for atmospheric CO2 and the magnitude of the sea–air CO2 flux over the global ocean vary seasonally and annually as well as geographically. These changes are the manifestation of changes in the partial pressure of sea water,

489

(pCO2)sw, which are caused primarily by changes in the water temperature, in the biological utilization of CO2, and in the lateral/vertical circulation of ocean waters including the upwelling of deep water rich in CO2. Over the global oceans, sea water temperatures change from the pole to the equator by about 321C. Since the pCO2 in sea water doubles with each 161C of warming, temperature changes should cause a factor of 4 change in pCO2. Biological utilization of CO2 over the global oceans is about 200 mmol CO2 kg1, which should reduce pCO2 in sea water by a factor of 3. If this is accompanied with growths of CaCO3-secreting organisms, the reduction of pCO2 could be somewhat smaller. While these effects are similar in magnitude, they tend to counteract each other seasonally, since the biological utilization tends to be large when waters are warm. In subpolar and polar areas, winter cooling of surface waters induces deep convective mixing that brings high pCO2 deep waters to the surface. The lowering effect on CO2 by winter cooling is often compensated for or some times over compensated for by the increasing effect of the upwelling of high CO2 deep waters. Thus, in high latitude oceans, surface waters may become a source for atmospheric CO2 during the winter time when the water is coldest. In Figure 1, the global distribution map of the sea– air pCO2 differences for February and August 1995, are shown. These maps were constructed on the basis of about a half million pairs of atmospheric and seawater pCO2 measurements made at sea over the 40-year period, 1958–98, by many investigators. Since the measurements were made in different years, during which the atmospheric pCO2 was increasing, they were corrected to a single reference year (arbitrarily chosen to be 1995) on the basis of the following observations. Warm surface waters in subtropical gyres communicate slowly with the underlying subsurface waters due to the presence of a strong stratification at the base of the mixed layer. This allows a long time for the surface mixed-layerwaters (B75 m thick) to exchange CO2 with the atmosphere. Therefore, their CO2 chemistry tends to follow the atmospheric CO2 increase. Accordingly, the pCO2 in the warm water follows the increasing trend of atmospheric CO2, and the sea–air pCO2 difference tends to be independent of the year of measurements. On the other hand, since surface waters in high latitude regions are replaced partially with subsurface waters by deep convection during the winter, the effect of increased atmospheric CO2 is diluted to undetectable levels and their CO2 properties tend to remain unchanged from year to year. Accordingly, the sea–air pCO2 difference measured in a given year increases as the atmospheric CO2


490


80˚

0˚ 20˚ 40˚ 60˚ 80˚ 100˚ 120˚ 140˚ 160˚ 180˚ 160˚ 140˚ 120˚ 100˚ 80˚ 60˚ 40˚ 20˚ 0˚

80˚

70˚

70˚

60˚

60˚

50˚

50˚

40˚

40˚

30˚

30˚

20˚

20˚

10˚

10˚

0˚

0˚

10˚

10˚

20˚

20˚

30˚

30˚ 40˚

40˚ 50˚

50˚ 60˚

60˚ 70˚

70˚ 80˚

0˚ 20˚ 40˚ 60˚ 80˚ 100˚ 120˚ 140˚ 160˚ 180˚ 160˚ 140˚ 120˚ 100˚ 80˚ 60˚ 40˚ 20˚ 0˚

_ 150 _ 135 _ 120 _105 _ 90 _ 75 _ 60 _ 45 _ 30 _ 15 0 15 30 GMGMT T 09:16 Apr 04 2000 _ Seawater Air pCO2 (µatm)

45

60

75

90

80˚

105 120 135

(A)

80˚

0˚ 20˚ 40˚ 60˚ 80˚ 100˚ 120˚ 140˚ 160˚ 180˚ 160˚ 140˚ 120˚ 100˚ 80˚ 60˚ 40˚ 20˚ 0˚

80˚

70˚

70˚

60˚

60˚

50˚

50˚

40˚

40˚

30˚

30˚

20˚

20˚

10˚

10˚

0˚

0˚

10˚

10˚

20˚

20˚

30˚

30˚ 40˚

40˚ 50˚

50˚ 60˚

60˚ 70˚

70˚ 80˚

0˚ 20˚ 40˚ 60˚ 80˚ 100˚ 120˚ 140˚ 160˚ 180˚ 160˚ 140˚ 120˚ 100˚ 80˚ 60˚ 40˚ 20˚ 0˚

_ 150 _ 135 _ 120 _105 _ 90 _ 75 _ 60 _ 45 _ 30 _ 15 G MGMT T 09:16 Apr 04 2000

0

15

30

45

60

75

90

80˚

105 120 135

Seawater_Air pCO2 (µatm)

(B) Figure 1 The sea–air pCO2 difference in matm (DpCO2) for (A) February and (B) August for the reference year 1995. The purpleblue areas indicate that the ocean is a sink for atmospheric CO2, and the red-yellow areas indicate that the ocean is source. The pink lines in the polar regions indicate the edges of ice fields.



concentration increases with time. This effect was corrected to the reference year using the observed increase in the atmospheric CO2 concentration. During El Ninõ periods, sea–air pCO2 differences over the equatorial belt of the Pacific Ocean, which are large in normal years, are reduced significantly and observations are scarce. Therefore, observations made between 101N and 101S in the equatorial Pacific for these periods were excluded from the maps. Accordingly, these maps represent the climatological means for non-El Ninõ period oceans for the past 40 years. The purple-blue areas indicate that the ocean is a sink for atmospheric CO2, and the red-yellow areas indicate that the ocean is a source. Strong CO2 sinks (blue and purple areas) are present during the winter months in the Northern (Figure 1A) and Southern (Figure 1B) Hemispheres along the poleward edges of the subtropical gyres, where major warm currents are located. The Gulf Stream in the North Atlantic and the Kuroshio Current in the North Pacific are both major CO2 sinks (Figure 1A) due primarily to cooling as they flow from warm tropical oceans to subpolar zones. Similarly, in the Southern Hemisphere, CO2 sink areas are formed by the cooling of poleward-flowing currents such as the Brazil Current located along eastern South America, the Agulhus Current located south of South Africa, and the East Australian Current located along south-eastern Australia. These warm water currents meet with cold currents flowing equator ward from the Antarctic zone along the northern border of the Southern (or Antarctic) Ocean. As the sub Antarctic waters rich in nutrients flow northward to more sunlit regions, CO2 is drawn down by photosynthesis, thus creating strong CO2 sink conditions, as exemplified by the Falkland Current in the western South Atlantic (Figure 1A). Confluence of subtropical waters with polar waters forms broad and strong CO2 sink zones as a result of the juxta position of the lowering effects on pCO2 of the cooling of warm waters and the photosynthetic drawdown of CO2 in nutrient-rich subpolar waters. This feature is clearly depicted in azone between 401S and 601S in Figure 1A representing the austral summer, and between 201S and 401S in Figure 1B representing the austral winter. During the summer months, the high latitude areas of the North Atlantic Ocean (Figure 1A) and the Weddell and Ross Seas, Antarctica(Figure 1B), are intense sink areas for CO2. This is attributed to the intense biological utilization of CO2 within the strongly stratified surface layer caused by solar warming and ice melting during the summer. The winter convective mixing of deep waters rich in CO2 and nutrient seliminates the strong CO2 sink and

491

replenishes the depleted nutrients in the surface waters. The Pacific equatorial belt is a strong CO2 source which is caused by the warming of upwelled deep waters along the coast of South America as well as by the upward entrainment of the equatorial under current water. The source strengths are most intense in the eastern equatorial Pacific due to the strong upwelling, and decrease to the west as a result of the biological utilization of CO2 and nutrients during the westward flow of the surface water. Small but strong source areas in the north-western subArctic Pacific Ocean are due to the winter convective mixing of deep waters (Figure 1A). The lowering effect on pCO2 of cooling in the winter is surpassed by the increasing effect of highCO2 concentration in the upwelled deep waters. During the summer (Figure 1B), however, these source areas become a sink for atmospheric CO2 due to the intense biological utilization that overwhelms the increasing effect on pCO2 of warming. A similar area is found in the Arabian Sea, where upwelling of deepwaters is induced by the south-west monsoon during July–August(Figure 1B), causing the area tobecome a strong CO2 source. This source area is eliminated by the photosynthetic utilization of CO2 following the end of the upwelling period (Figure 1A). As illustrated in Figure 1A and B, the distribution of oceanic sink and source areas for atmospheric CO2 varies over a wide range in space and time. Surface ocean waters are out of equilibrium with respect to atmospheric CO2 by as much as 7200 matm (or by760%).The large magnitudes of CO2 disequilibrium between the sea and theair is in contrast with the behavior of oxygen, another biologically mediated gas, that shows only up to 710% sea–air disequilibrium. The large CO2 disequilibrium may be attributed to the fact that the internal ocean processes that control pCO2 in sea water, such as the temperature of water, the photosynthesis, and the upwelling of deep waters,occur at much faster rates than the sea–air CO2 transfer rates. The slow rate of CO2 transfer across the sea surface is due to the slow hydration rates of CO2 as well as to the large solubility of CO2 in sea water attributable to the formation of bicarbonate and carbonate ions. The latter effect does not exist at all for oxygen.

Net CO2 Flux Across the Sea Surface The net sea–air CO2 flux over the global oceans may be computed using eqns [2] and [3]. Figure 2 shows the climatological mean distribution of the annual sea–air CO2 flux for the reference year 1995 using


492


the following set of information. (1) The monthly mean DpCO2 values in 41 51 pixel areas for the reference year 1995 (Figure 1A and B for all other months); (2) the Wanninkhof formulation, eqn [3], for the effect of wind speed on the CO2 gas transfer coefficient; and (3) the climatological mean wind speeds for each month compiled by Esbensen and Kushnir in 1981. This set yields a mean global gas transfer rate of 0.063 mole CO2 m2 matm1 y1, that is consistent with 20 moles CO2 m2 y1 estimated on the basis of carbon-14 distribution in the atmosphere and the oceans. Figure 2 shows that the equatorial Pacific is a strong CO2 source. On the other hand, the areas along the poleward edges of the temperate gyres in both hemispheres are strong sinks for atmospheric CO2. This feature is particularly prominent in the southern Indian and Atlantic Oceans between 401S and 601S, and is attributable to the combined effects of negative sea–air pCO2 differences with strong winds (‘the roaring 40 s’) that accelerate sea–air gas transfer rates. Similarly strong sink zones are formed in the North Pacific and North Atlantic between 451N and 601N. In the high latitude Atlantic, strong

sink areas extend into the Norwegian and Greenland Seas. Over the high latitude Southern Ocean areas, the sea–air gas transfer is impeded by the field of ice that covers the sea surface for X6 months in a year. The net sea–air CO2 fluxes computed for each ocean basin for the reference year of 1995, representing non-El Ninõ conditions, are summarized in Table 1. The annual net CO2 uptake by the global ocean is estimated to be about 2.0 Pg-C y1. This is consistent with estimates obtained on the basis of a number of different ocean–atmosphere models including multi-box diffusion advection models and three-dimensional general circulation models. The uptake flux for the Northern Hemisphere ocean (north of 141N) is 1.2 Pg-C y1, whereas that for the Southern Hemisphere ocean (south of 141S) is 1.7 Pg-C y1. Thus, the Southern Hemisphere ocean is astronger CO2 sink by about 0.5 Pg-C y1. This is due partially to the much greater oceanic areas in the Southern Hemisphere. In addition, the Southern Ocean south of 501S is an efficient CO2 sink, for it takes up about 26% of the global ocean CO2 uptake, while it has only 10% of the global ocean area. Cold temperature and moderate photosynthesis are both

Annual flux (u2 wind) 80˚

0˚ 20˚ 40˚ 60˚ 80˚ 100˚ 120˚ 140˚ 160˚ 180˚ 160˚ 140˚ 120˚ 100˚ 80˚ 60˚ 40˚ 20˚ 0˚

80˚

70˚

70˚

60˚

60˚

50˚

50˚

40˚

40˚

30˚

30˚

20˚

20˚

10˚

10˚

0˚

0˚

10˚

10˚

20˚

20˚

30˚

30˚ 40˚

40˚ 50˚

50˚ 60˚

60˚ 70˚

70˚ 80˚

_ 9.5 GM T 15:27 Apr 03 2000 GMT

0˚ 20˚ 40˚ 60˚ 80˚ 100˚ 120˚ 140˚ 160˚ 180˚ 160˚ 140˚ 120˚ 100˚ 80˚ 60˚ 40˚ 20˚ 0˚

_ 6.0 _ 5.0

_ 4.0

_ 3.0

_ 2.0 _ 1.0 _2

Net flux (moles CO2 m

0.0

1.0

2.0

3.0

80˚

5.1

_1

y )

Figure 2 The mean annual sea–air flux of CO2 in the reference year 1995. The red-yellow areas indicate that the flux is from sea to air, whereas blue-purple areas indicate that the flux is from air to sea. The flux is given in moles of CO2 m2 y1. The map gives a total annual air-to-sea flux of 2.0 Pg-C y1.



493

Table 1 The net sea–air flux of CO2 estimated for a reference year of 1995 using the effect of wind speed on the CO2 gas transfer coefficient, eqn [3], of Wanninkhof and the monthly wind field of Esbensen and Kushnir Latitudes

Pacific Ocean

Atlantic Ocean

Indian Ocean

Southern Ocean

Global Oceans

Sea–air flux in 1015g Carbon y 1 North of 501N 501N–141N 141N–141S 141S–501S South of 501S

0.02 0.47 þ 0.64 0.37 —

0.44 0.27 þ 0.13 0.20 —

— þ 0.03 þ 0.09 0.60 —

— — — — 0.52

0.47 0.73 þ 0.86 1.17 0.52

Total %Uptake

0.23 11%

0.78 39%

0.47 24%

0.52 26%

2.00 100%

Area (106 km2) Area (%)

151.6 49.0%

72.7 23.5%

53.2 17.2%

31.7 10.2%

309.1 100%

Positive values indicate sea-to-air fluxes, and negative values indicate air-to-sea fluxes.

responsible for the large uptake by the Southern Ocean. The Atlantic Ocean is the largest net sink for atmospheric CO2 (39%); the Southern Ocean (26%) and the Indian Ocean (24%) are next; and the Pacific Ocean (11%) is the smallest. The intense biological drawdown of CO2 in the high latitude areas of the North Atlantic and Arctic seasduring the summer months is responsible for the Atlantic being a major sink. This is also due to the fact that the upwelling deep waters in the North Atlantic contain low CO2 concentrations, which are in turn caused primarily by the short residence time (B80y) of the North Atlantic Deep Waters. The small uptake flux of the Pacific can be attributed to the fact that the combined sink flux of the northern and southern subtropical gyres is roughly balanced by the source flux from the equatorial Pacific during non-El Ninõ periods. On the other hand, the equatorial Pacific CO2 source flux is significantly reduced or eliminated during El Ninõ events. As a result the equatorial zone is covered with the eastward spreading of the warm, low pCO2 western Pacific waters in response to the relaxation of the trade wind. Although the effects of El Ninõ and Southern Ocean Oscillation may be far reaching beyond the equatorial zone as far as to the polar areas, the El Ninõ effects on the equatorial Pacific alone could reduce the equatorial CO2 source. Hence, this could increase the global ocean uptake flux by up to 0.6 PgC y1 during an El Ninõ year. The sea–air CO2 flux estimated above is subject to three sources of error: (1) biases in sea–air DpCO2 values interpolated from relatively sparse observations, (2) the ‘skin’ temperature effect, and(3) uncertainties in the gas transfer coefficient estimated on

the basis of the wind speed dependence. Possible biases in DpCO2 differences have been tested using sea surface temperatures (SST) as a proxy. The systematic error in the global sea–air CO2 flux resulting from sampling and interpolation has been estimated to be about 730% or 70.6 Pg-C y1. The‘skin’ temperature of ocean water may affect DpCO2 by as much as 76 matm depending upon time and place, as discussed earlier. Although the distribution of the ‘skin’ temperature over the global ocean is not known, it may be cooler than the bulk water temperature by a few tenths of a degree on the global average. This may result in an under estimation of the ocean uptake by 0.4 PgC y1. The estimated global sea–air flux depends on the wind speed data used. Since the gas transfer rate increases nonlinearly with wind speed, the estimated CO2 fluxes tend to be smaller when mean monthly wind speeds are used instead of high frequency wind data. Furthermore, the wind speed dependence on the CO2 gas transfer coefficient in high wind speed regimes is still questionable. If the gas transfer rate is taken to be a cubic function of wind speed instead of the square dependence as shown above, the global ocean uptake would be increased by about 1 Pg-C y1. The effect is particularly significant over the high latitude oceans where the winds are strong. Considering various uncertainties discussed above, the global ocean CO2 uptake presented in Table 1 is uncertain by about 1 Pg-C y1.

See also Air–Sea Gas Exchange. Arctic Ocean Circulation. Carbon Cycle. Ocean Carbon System, Modeling of.


494


Primary Production Distribution. Radiocarbon. Stable Carbon Isotope Variations in the Ocean. Wind Driven Circulation.

Further Reading Broecker WS and Peng TH (1982) Tracers in the Sea. Palisades, NY: Eldigio Press. Broecker WS, Ledwell JR, Takahashi, et al. (1986) Isotopic versus micrometeorologic ocean CO2 fluxes a: serious conflict. Journal of Geophysical Research 91: 10517--10527. Keeling R, Piper SC, and Heinmann M (1996) Global and hemispheric CO2 sinks deduced from changes in atmospheric O2 concentration. Nature 381: 218--221. Sarmiento JL, Murnane R, and Le Quere C (1995) Air–sea CO2 transfer and the carbon budget of the North Atlantic. Philosophical Transactions of the Royal Society of London, series B 343: 211--219. Sundquist ET (1985) Geological perspectives on carbon dioxide and carbon cycle. In: Sundquist ET and Broecker WS (eds.) The Carbon Cycle and Atmospheric

CO2 N:atural Variations, Archean to Present, Geophysical Monograph 32, pp. 5--59. Washington, DC: American Geophysical Union. Takashahi T, Olafsson J, Goddard J, Chipman DW, and Sutherland SC (1993) Seasonal variation of CO2 and nutrients in the high-latitude surface oceans a: comparative study. Global Biogeochemical Cycles 7: 843--878. Takahashi T, Feely RA, Weiss R, et al. (1997) Global air– sea flux of CO2 a:n estimate based on measurements of sea–air pCO2 difference. Proceedings of the National Academy of science USA 94: 8292--8299. Tans PP, Fung IY, and Takahashi T (1990) Observational constraints on the global atmospheric CO2 budget. Sciece 247: 1431--1438. Wanninkhof R (1992) Relationship between wind speed and gas exchange. Journal of Geophysical Research 97: 7373--7382. Wanninkhof R and McGillis WM (1999) A cubic relationship between gas transfer and wind speed. Geophysical Research Letters 26: 1889--1893.


CARBON SEQUESTRATION VIA DIRECT INJECTION INTO THE OCEAN E. E. Adams, Massachusetts Institute of Technology, Cambridge, MA, USA K. Caldeira, Stanford University, Stanford, CA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Global climate change, triggered by a buildup of greenhouse gases, is emerging as perhaps the most serious environmental challenge in the twenty-first century. The primary greenhouse gas is CO2, whose concentration in the atmosphere has climbed from its preindustrial level of c. 280 to 4380 ppm. Stabilization at no more than 500–550 ppm is a target frequently discussed to avoid major climatic impact. The primary source of CO2 is the burning of fossil fuels – specifically gas, oil, and coal – so stabilization of atmospheric CO2 concentration will clearly require substantial reductions in CO2 emissions from these sources. For example, one commonly discussed scenario to stabilize at 500 ppm by the mid-twentyfirst century suggests that about 640 Gt CO2 (c. 175 Gt C) would need to be avoided over 50 years, with further emission reductions beyond 50 years. As references, a 1000 MW pulverized coal plant produces 6–8 Mt CO2 (c. 2 Mt C) per year, while an oil-fired single-cycle plant produces about two-thirds this amount and a natural gas combined cycle plant produces about half this amount. Thus the above scenario would require that the atmospheric emissions from the equivalent of 2000–4000 large power plants be avoided by approximately the year 2050. Such changes will require a dramatic reduction in our current dependence on fossil fuels through increased conservation and improved efficiency, as well as the introduction of nonfossil energy sources like solar, wind, and nuclear. While these strategies will slow the buildup of atmospheric CO2, it is probable that they will not reduce emissions to the required level. In other words, fossil fuels, which currently supply over 85% of the world’s energy needs, are likely to remain our primary energy source for the foreseeable future. This has led to increased interest in a new strategy termed carbon capture and storage, or sequestration. The importance of this option for mitigating climate change is highlighted by the recent

Special Report on Carbon Dioxide Capture and Storage published by the Intergovernmental Panel on Climate Change, to which the reader is referred for more information. Carbon sequestration is often associated with the planting of trees. As they mature, the trees remove carbon from the atmosphere. As long as the forest remains in place, the carbon is effectively sequestered. Another type of sequestration involves capturing CO2 from large, stationary sources, such as a power plant or chemical factory, and storing the CO2 in underground reservoirs or the deep ocean, the latter being the focus of this article. There has been much attention paid recently to underground storage with several large-scale field sites in operation or being planned. Conversely, while there have been many studies regarding use of the deep ocean as a sink for atmospheric carbon, there have been only a few small-scale field studies. Why is the ocean of interest as a sink for anthropogenic CO2? The ocean already contains an estimated 40 000 Gt C compared with about 800 Gt C in the atmosphere and 2200 Gt C in the land biosphere. As a result, the amount of carbon that would cause a doubling of the atmospheric concentration would only change the ocean concentration by about 2%. In addition, natural chemical equilibration between the atmosphere and ocean would result in about 80% of present-day emissions ultimately residing in the ocean. Discharging CO2 directly to the ocean would accelerate this slow, natural process, thus reducing both peak atmospheric CO2 concentrations and their rate of increase. It is noted that a related strategy for sequestration – not discussed here – would be to enhance the biological sink using nutrients such as iron to fertilize portions of the world’s oceans, thus stimulating phytoplankton growth. The phytoplankton would increase the rate of biological uptake of CO2, and a portion of the CO2 would be transported to ocean depths when the plankton die. The indirect flux of CO2 to the ocean from the atmosphere is already quite apparent: since preindustrial times, the pH of the surface ocean has been reduced by about 0.1 units, from an initial surface pH of about 8.2. Figure 1 illustrates what could happen to ocean pH under conditions of continued atmospheric release of CO2. Under the conditions simulated, the pH of the surface would drop by over 0.7 units. Conversely, by injecting some of the CO2


495

20

2000

10

is buffered by the fact that total alkalinity is conserved, which results in carbonate ion being converted into bicarbonate. Thus, the principal reactions occurring when CO2 is dissolved in seawater are

Emissions

0 pCO2

1000 0

−0.7

1 2 −0.1

3

−0.2

−0.3

−0.4

4 27 50

25 00

22 50

20 00

17 50

ΔpH 30 00

Depth (km)

Gt C yr−1

CARBON SEQUESTRATION VIA DIRECT INJECTION INTO THE OCEAN

ppm

496

Year Figure 1 Model simulations of long-term ocean pH changes, averaged horizontally, as a result of atmospheric CO2 emissions shown in the top panel. Reprinted from Caldeira K and Wickett ME (2003) Anthropogenic carbon and ocean pH. Nature 425: 365.

to the deep ocean, the change in pH could be more uniformly distributed. Ocean sequestration of CO2 by direct injection assumes that a relatively pure CO2 stream has been generated at a power plant or chemical factory and transported to an injection point. To better understand the role the ocean can play, we address the capacity of the ocean to sequester CO2, its effectiveness at reducing atmospheric CO2 levels, how to inject the CO2, and possible environmental consequences and issues of public perception.

Capacity How much carbon can the ocean sequester? At over 70% of the Earth’s surface and an average depth of 3800 m, the ocean has enormous storage capacity; based on physical chemistry, the amount of CO2 that could be dissolved in the deep ocean far exceeds the estimated available fossil energy resources of 5000– 10 000 Gt C. However, a more realistic criterion needs to be based on an understanding of ocean biogeochemistry and expected environmental impact. CO2 exists in seawater in various forms as part of the carbonate system: CO2 ðaqÞ þ H2 O2H2 CO3 ðaqÞ 2 Hþ þ HCO3 22Hþ þ CO3 2

½1

Dissolving additional CO2 increases the hydrogen ion concentration (lowering the pH), but the change

CO2 þ H2 O þ CO3 2 -2HCO3

½2

CO2 þ H2 O þ Hþ -HCO3

½3

which result in a decrease in pH and carbonate ion, and an increase in bicarbonate ion. Reduced pH is one of the principal environmental impacts threatening marine organisms, the other being the concentration of CO2 itself. At short travel times from the injection point, the changes in pH and CO2 concentration will be greatest, which suggests that injection schemes should achieve the maximum dilution possible to minimize potential acute impacts in the vicinity of injection. See further discussion below. At longer travel times, injected carbon would be distributed widely in the oceans and any far-field impact of the injected CO2 on the oceans would be similar to the impact of anthropogenic CO2 absorbed from the atmosphere. As indicated above, such changes are already taking place within the surface ocean, where the pH has been reduced by about 0.1 unit. Adding about 2000 Gt CO2 to the ocean would reduce the average ocean pH by about 0.1 unit, while adding about 5600 Gt CO2 (about 200 years of current emissions) would decrease the average ocean pH by about 0.3 units. (It should be noted that with stabilization of atmospheric CO2 at 550 ppm, natural chemical equilibration between the atmosphere and ocean will result in eventual storage of over 6000 Gt CO2 in the ocean.) The impacts of such changes are poorly understood. The deep-ocean environment has been relatively stable and it is unknown to what extent changes in dissolved carbon or pH would affect these ecosystems. However, one can examine measured spatial and temporal variation in ocean pH to understand how much change might be tolerated. The spatial variability within given zoogeographic regions and bathymetric ranges (where similar ecosystems might be expected), and the temporal variability at a particular site, have both been found to vary by about 0.1 pH unit. If it is assumed that a change of 0.1 unit is a threshold tolerance, and that CO2 should be stored in the bottom half of the ocean’s volume (to maximize retention), nearly 1000 Gt CO2 might be stored, which exceeds the 640 Gt CO2 over 50 years estimated above. It is important to recognize that the long-term changes in ocean pH would ultimately be much the same



whether the CO2 is released into the atmosphere or the deep ocean. However, in the shorter term, releasing the CO2 in the deep ocean will diminish the pH change in the near-surface ocean, where marine biota are most plentiful. Thus, direct injection of CO2 into the deep ocean could reduce adverse impacts presently occurring in the surface ocean. In the long run, however, a sustainable solution to the problem of climate change must ultimately entail a drastic reduction of total CO2 emissions.

Effectiveness Carbon dioxide is constantly exchanged between the ocean and atmosphere. Each year the ocean and atmosphere exchange about 350 Gt CO2, with a net ocean uptake currently of about 8 Gt CO2. Because of this exchange, questions arise as to how effective ocean sequestration will be at keeping the CO2 out of the atmosphere. Specifically, is the sequestration permanent, and if not, how fast does the CO2 leak back to the atmosphere. Because there has been no long-term CO2 direct-injection experiment in the ocean, the long-term effectiveness of direct CO2 injection must be predicted based on observations of other oceanic tracers (e.g., radiocarbon) and on computer models of ocean circulation and chemistry. As implied earlier, because the atmosphere and ocean are currently out of equilibrium, most CO2 emitted to either media will ultimately enter the ocean. The percentage that is permanently sequestered depends on the atmospheric CO2 concentration, through the effect of atmospheric CO2 on surface ocean chemistry (see Table 1). At today’s concentration of c. 380 ppm, nearly 80% of any Table 1 Percent of injected CO2 permanently sequestered from the atmosphere as a function of atmospheric CO2 stabilization concentration Atmospheric carbon dioxide concentration (ppm)

Percentage of carbon dioxide permanently sequestered

350 450 550 650 750 1000

80 77 74 72 70 66

Based on data in IPCC (2005) Special Report on Carbon Dioxide Capture and Storage. Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press. http://arch.rivm.nl/env/int/ipcc/ pages_media/SRCCS-final/IPCCSpecialReportonCarbondioxide CaptureandStorage.htm (accessed Mar. 2008) and references therein.

497

carbon emitted to either the atmosphere or the ocean would be permanent, while at a concentration of 550 ppm, 74% would be permanent. Of course, even at equilibrium, CO2 would continue to be exchanged between the atmosphere and oceans, so the carbon that is currently being injected is not exactly the same carbon that will reside in equilibrium. For CO2 injected to the ocean today, the net quantity retained in the ocean ranges from 100% (now) to about 80% as equilibrium between the atmosphere and oceans is approached. (A somewhat greater percentage will ultimately be retained as CO2 reacts with ocean sediments over a timescale of thousands of years.) The nomenclature surrounding ocean carbon storage can be somewhat confusing. The percentage retained in the ocean shown in Figure 2 is the fraction of injected CO2 that has never interacted with the atmosphere. Table 1 shows the fraction of CO2 that contacts the atmosphere that remains permanently in the ocean. So, for example, for a 550 ppm atmosphere, even as the ‘retained fraction’ approaches zero (Figure 2), the amount permanently stored in the ocean approaches 74% (see Table 1). The exact time course depends on the location and depth of the injection. Several computer modeling studies have studied the issue of retention. The most comprehensive summary is the Global Ocean Storage of Anthropogenic Carbon (GOSAC) intercomparison study of several ocean general circulation models (OGCMs). In this study a number of OGCMs simulated the fate of CO2, injected over a period of 100 years at seven locations and three depths, for a period of 500 years. The CO2 retained as a function of time, averaged over the seven sites, is shown in Figure 2. While there is variability among models, they all show that retention increases with injection depth, with most simulations predicting over 70% retention after 500 years for an injection depth of 3000 m. The time required for injected carbon to mix from the deep ocean to the atmosphere is roughly equal to the time required for carbon to mix from the atmosphere to the deep ocean. This can be estimated through observations of radiocarbon (carbon-14) in the ocean. Correcting for mixing of ocean waters from different sources, the age of North Pacific deep water is in the range of 700–1000 years, while other basins, such as the North Atlantic, have overturning times of 300 years or more. These estimates are consistent with output from OGCMs and, collectively, suggest that outgassing of the 20% of injected carbon would occur on a timescale of 300–1000 years. It is important to stress that leakage to the atmosphere would take place gradually and over large areas of the ocean. Thus, unlike geological sequestration, it


498


Retained fraction

1.0 MPIM PIUB PRINCE PRINC2 SOC

0.8 0.6 AWI CSIRO IPSL(HOR) IPSL(GM) LLNL

0.4 0.2

800 m

0.0 2000

2500

Retained fraction

1.0 0.8 0.6 0.4 0.2 1500 m

0.0 2000

2500

Retained fraction

1.0 0.8 0.6 0.4 0.2 3000 m 0.0 2000

2500 Year

Figure 2 Model-intercomparison study reported by Orr in 2004 showing fraction of CO2, injected from 2000 through 2100, that remains isolated from the atmosphere as a function of time and injection depth. Results are averaged over seven injection locations. Most of the CO2 that does interact with the atmosphere remains in the ocean (see Table 1), so the amount of CO2 remaining in the ocean is much greater than shown here. Reprinted with permission from IEA Greenhouse Gas R&D Programme.

would not be possible to produce a sudden release that could lead to harmful CO2 concentrations at the ocean or land surface.

Injection Methods The first injection concept was proposed by the Italian physicist Cesare Marchetti, who thought to introduce CO2 into the outflow of the Mediterranean

Sea, where the relatively dense seawater would cause the CO2 to sink as it entered the Atlantic Ocean. As illustrated in Figure 3, a number of options have been considered since then. Understanding these methods requires some background information on the CO2–seawater system. Referring to Figure 4, at typical ocean pressures and temperatures, pure CO2 would be a gas above a depth of 400–500 m and a liquid below that depth. Liquid CO2 is more compressible than seawater, and would be positively buoyant (i.e., it will rise) down to about 3000 m, but negatively buoyant (i.e., it will sink) below that depth. At about 3700 m, the liquid becomes negatively buoyant compared to seawater saturated with CO2. In seawater–CO2 systems, CO2 hydrate (CO2 nH2O, n B5.75) can form below c. 400 m depth depending on the relative compositions of CO2 and H2O. CO2 hydrate is a solid with a density about 10% greater than that of seawater. The rising droplet plume has been the most studied and is probably the easiest scheme to implement. It would rely on commercially available technology to inject the CO2 as a stream of buoyant droplets from a bottom manifold. Effective sequestration can be achieved by locating the manifold below the thermocline, and dilution can be increased by increasing the manifold length. Even better dilution can be achieved by releasing the CO2 droplets from a moving ship whose motion provides additional dispersal. Although the means of delivery are different, the plumes resulting from these two options would be similar, each creating a vertical band of CO2-enriched seawater over a prescribed horizontal region. Another promising option is to inject liquid CO2 into a reactor where it can react at a controlled rate with seawater to form hydrates. While it is difficult to achieve 100% reaction efficiency, laboratory and field experiments indicate that negative buoyancy, and hence sinking, can be achieved with as little as about 25% reaction efficiency. The hydrate reactor could be towed from a moving ship to encourage dilution, or attached to a fixed platform, where the large concentration of dense particles, and the increased seawater density caused by hydrate dissolution, would induce a negatively buoyant plume. The concept of a CO2 lake is based on a desire to minimize leakage to the atmosphere and exposure to biota. This would require more advanced technology and perhaps higher costs, as the depth of the lake should be at least 3000 m, which exceeds the depths at which the offshore oil industry currently works. The CO2 in the lake would be partly in the form of solid hydrates. This would limit the CO2 dissolution into the water column, further slowing leakage to the atmosphere from that shown in Figure 2, which



CO2 /CaCO3 reactor Flue gas

Captured and compressed CO2

Dispersal of CO2 /CaCO3 mixture

Dispersal of CO2 by ship

499

Refilling ship Rising CO2 plume

m

Sinking CO2 plume

CO2 lake

3k

CO2 lake

Figure 3 Different strategies for ocean carbon sequestration. Reprinted with permission from IPCC (2005) Special Report on Carbon Dioxide Capture and Storage, figure TS-9 (printed as Special Report on Safeguarding the Ozone Layer and the Global Climate System, Figure 6.1). Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press. http://arch.rivm.nl/env/int/ipcc/pages_media/SRCCS-final/IPCCSpecialReportonCarbondioxideCaptureandStorage. htm, with permission from the Intergovernmental Panel on Climate Change.

200

Depth (m)

400 Gas

Hydrate stability zone

Liquid

600

800 0

4

8 12 Temperature (°C)

16

Figure 4 Phase diagram for CO2 including typical ocean temperature profile (solid line). Reprinted from Brewer PG, Peltzer E, Aya I, et al. (2004) Small scale field study of an ocean CO2 plume. Journal of Oceanography 60(4): 751.

assumes that CO2 is injected into the water column. It is also possible that various approaches could be engineered to physically contain CO2 on the seafloor and isolate the CO2 from the overlying water column (and perhaps the sediments); however, this would entail an additional cost. Another method that has received attention is injecting a dense CO2–seawater mixture at a depth of 500–1000 m, forming a sinking bottom gravity current. CO2-enriched seawater is less than 1% heavier than seawater, but this is sufficient to promote a sinking density current, especially if the current were formed along a submarine canyon. However, the environmental impacts would be greater with this option due to the concentrated nature of the plume, and its contact with the seafloor. As discussed earlier, the deep ocean equilibrates with the surface ocean on the scale of 300–1000 years, and by injecting anthropogenic CO2 into the deep ocean, the surface-to-deep mixing timescale is effectively bypassed. Anthropogenic CO2 also equilibrates with carbonate sediments, but over a much longer time, about 6000 years. Technical


500


means could also be used to bypass this timescale, thereby increasing the effectiveness and diminishing the environmental impacts of intentional storage of carbon dioxide in the ocean. For example, CO2 reacts with carbonate sediments to form bicarbonate ions (HCO3 ) as indicated by eqn [2]. Power plant CO2 could be dissolved in seawater, then reacted with crushed limestone, either at the power plant or at the point of release, thus minimizing changes in plume pH. Or an emulsion of liquid CO2-in-water could be stabilized by fine particles of pulverized limestone; the emulsion would be sufficiently dense to form a sinking plume, whose pH change would be buffered by the limestone. Drawbacks of these approaches include the cost to mine and transport large quantities of carbonate minerals.

Local Environmental Impacts and Public Perception Environmental impacts may be the most significant factor determining the acceptability of ocean storage, since the strategy is predicated on the notion that impacts to the ocean will be significantly less than the avoided impacts of continued emission to the atmosphere. Earlier, environmental impacts were discussed from the global viewpoint. Here, we examine the environmental impacts near the injection point. A number of studies have summarized potential impacts to different types of organisms, including adult fish, developmental fish, zooplankton, and benthic fauna. While earlier studies focused mainly on lethal impacts to coastal fauna exposed to strong acids, recent data have focused on deep-water organisms exposed to CO2, and have included sublethal effects. Impacts include respiratory stress (reduced pH limits oxygen binding and transport of respiratory proteins), acidosis (reduced pH disrupts an organism’s acid/basis balance), and metabolic depression (elevated CO2 causes some animals to reach a state of torpor). Data generally show that CO2 causes greater stress than an equivalent change in pH caused by a different acid, that there are strong differences in tolerance among different species and among different life stages of the same species, and that the duration of stress, as well as the level of stress, are important. While some studies imply that deep organisms would be less tolerant than surface organisms, other studies have found the opposite. Likewise, some animals are able to avoid regions of high CO2 concentration, while others appear less able to. Results generally suggest that lethal effects can be avoided by achieving high near-field dilution. However, more research

is needed to resolve impacts, especially at the community level (e.g., reduced lifespan and reproduction effects). The viability of ocean storage as a greenhouse gas mitigation option hinges on social, political, and regulatory considerations. In view of public precaution toward the ocean, the strategy will require that all parties (private, public, nongovernmental organizations) be included in ongoing research and debate. But the difficulty in this approach is highlighted by the recent experience of an international research team working on ocean carbon sequestration research. A major part of their collaboration was to have included a field experiment involving release of 5 t of CO2 off the coast of Norway. Researchers would have monitored the physical, chemical, and biological effects of the injected CO2 over a period of about a week. However, lobbying from environmental groups caused the Norwegian Minister of Environment to rescind the group’s permit that had previously been granted. Such actions are unfortunate, because field experiments of this type are what is needed to produce data that would help policymakers decide if full-scale implementation would be prudent.

See also Abrupt Climate Change. Air–Sea Gas Exchange. Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens. Air–Sea Transfer: N2O, NO, CH4, CO. Carbon Cycle. Carbon Dioxide (CO2) Cycle. Organic Carbon Cycling in Continental Margin Environments. Ocean Carbon System, Modeling of. Paleoceanography: the Greenhouse World.

Further Reading Alendal G and Drange H (2001) Two-phase, near field modeling of purposefully released CO2 in the ocean. Journal of Geophysical Research 106(C1): 1085--1096. Brewer PG, Peltzer E, Aya I, et al. (2004) Small scale field study of an ocean CO2 plume. Journal of Oceanography 60(4): 751--758. Caldeira K and Rau GH (2000) Accelerating carbonate dissolution to sequester carbon dioxide in the ocean: Geochemical implications. Geophysical Research Letters 27(2): 225--228. Caldeira K and Wickett ME (2003) Anthropogenic carbon and ocean pH. Nature 425: 365. Giles J (2002) Norway sinks ocean carbon study. Nature 419: 6. Golomb D, Pennell S, Ryan D, Barry E, and Swett P (2007) Ocean sequestration of carbon dioxide: Modeling the



deep ocean release of a dense emulsion of liquid CO2-in-water stabilized by pulverized limestone particles. Environmental Science and Technology 41(13): 4698--4704. Haugan H and Drange H (1992) Sequestration of CO2 in the deep ocean by shallow injection. Nature 357(28): 1065--1072. IPCC (2005) Special Report on Carbon Dioxide Capture and Storage. Prepared by Working Group III of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge University Press. http://arch.rivm.nl/env/ int/ipcc/pages_media/SRCCS-final/IPCCSpecialReporton CarbondioxideCaptureandStorage.htm (accessed Mar. 2008). Ishimatsu A, Kikkawa T, Hayashi M, and Lee KS (2004) Effects of CO2 on marine fish: Larvae and adults. Journal of Oceanography 60: 731--741. Israelsson P and Adams E (2007) Evaluation of the Acute Biological Impacts of Ocean Carbon Sequestration. Final Report for US Dept. of Energy, under grant DE-FG2698FT40334. Cambridge, MA: Massachusetts Institute of Technology. Kikkawa T, Ishimatsu A, and Kita J (2003) Acute CO2 tolerance during the early developmental stages of four marine teleosts. Environmental Toxicology 18(6): 375--382. Ohsumi T (1995) CO2 storage options in the deep-sea. Marine Technology Society Journal 29(3): 58--66. Orr JC (2004) Modeling of Ocean Storage of CO2 – The GOSAC Study, Report PH4/37, 96pp. Paris: Greenhouse Gas R&D Programme, International Energy Agency.

501

Ozaki M, Minamiura J, Kitajima Y, Mizokami S, Takeuchi K, and Hatakenka K (2001) CO2 ocean sequestration by moving ships. Journal of Marine Science and Technology 6: 51--58. Po¨rtner HO, Reipschla¨ger A, and Heisler N (2004) Biological impact of elevated ocean CO2 concentrations: Lessons from animal physiology and Earth history. Journal of Oceanography 60(4): 705--718. Riestenberg D, Tsouris C, Brewer P, et al. (2005) Field studies on the formation of sinking CO2 particles for ocean carbon sequestration: Effects of injector geometry on particle density and dissolution rate and model simulation of plume behavior. Environmental Science and Technology 39: 7287--7293. Sato T and Sato K (2002) Numerical prediction of the dilution process and its biological impacts in CO2 ocean sequestration. Journal of Marine Science and Technology 6(4): 169--180. Vetter EW and Smith CR (2005) Insights into the ecological effects of deep-ocean CO2 enrichment: The impacts of natural CO2 venting at Loihi seamount on deep sea scavengers. Journal of Geophysical Research 110: C09S13 (doi:10.1029/2004JC002617). Wannamaker E and Adams E (2006) Modeling descending carbon dioxide injections in the ocean. Journal of Hydraulic Research 44(3): 324--337. Watanabe Y, Yamaguchi A, Ishida H, et al. (2006) Lethality of increasing CO2 levels on deep-sea copepods in the western North Pacific. Journal of Oceanography 62: 185--196.


CENOZOIC CLIMATE – OXYGEN ISOTOPE EVIDENCE J. D. Wright, Rutgers University, Piscataway, NJ, USA

expressed in delta notation, d18O, where:


d18 O ¼


Discoveries of fossil remains of 50 million year old alligators on Ellesmere Island and 30–40 million year-old forests on Antarctica contrast sharply with our present vision of polar climates. These are not isolated discoveries or quirks of nature. An evergrowing body of faunal, floral, and geochemical evidence shows that the first half of the Cenozoic Era was much warmer than the present time. What maintained such a warm climate and could it be an analog for future global warming? To address these and other questions, one needs more than a qualitative estimate of planetary temperatures. Quantitative temperature estimates (both magnitudes and rates of change) are required to depict how the Earth’s climate has changed through time. One of the most powerful tools used to reconstruct past climates during the Cenozoic (the last 65 million years of Earth’s history) is the analysis of oxygen isotopes in the fossil shells of marine organisms. The calcium carbonate shells of the protist foraminifera are the most often analyzed organisms because the different species are distributed throughout surface (planktonic) and deep (benthic) marine environments.

Oxygen Isotope Systematics The stable isotopes of oxygen used in paleooceanographic reconstructions are 16O and 18O. There are about 500 16O atoms for every 18O atom in the ocean/atmosphere environment. During the 1940s, Harold Urey at the University of Chicago predicted that the 18O/16O ratio in calcite (CaCO3) should vary as a function of the temperature at which the mineral precipitated. His prediction spurred on experiments by himself and others at the University of Chicago who measured 18O/16O ratios in CaCO3 precipitated in a wide range of temperatures, leading to the use of stable oxygen isotope measurements as a paleothermometer. To determine oxygen isotopic ratios, unknown 18 O/16O ratios are compared to the known 18O/16O ratio of a standard. The resulting values are

502

18

O=16 Osample 18 O=16 Ostandard 18 O=16 O

1000

½1

standard

Carbonate samples are reacted in phosphoric acid to produce CO2. To analyze water samples, CO2 gas is equilibrated with water samples at a constant temperature. Given time, the CO2 will isotopically equilibriate with the water. For both the carbonate and water samples, the isotopic composition of CO2 gas is compared with CO2 gas of known isotopic composition. There are two standards for reporting d18O values. For carbonate samples, the reference standard is PDB, which was a crushed belemnite shell (Belemnitella americana) from the Peedee formation of Cretaceous age in South Carolina. The original PDB material has been exhausted, but other standards have been calibrated to PDB and are used as an intermediate reference standard through which a PDB value can be calculated. For measuring the isotopic composition of water samples, Standard Mean Ocean Water (SMOW) is used as the reference. The SMOW reference was developed so that its d18Owater value is 0.0% (parts per thousand) and approximates the average oxygen isotopic composition of the whole ocean. Deep ocean d18Owater values are close to the SMOW value, ranging from 0.2 to 0.2%. In contrast, surface ocean d18Owater values exhibit a much greater variability, varying between 0.5 and þ 1.5%.

Oxygen Isotope Paleothermometry Early studies into the natural variations in oxygen isotopes led to the development of a paleotemperature equation. The temperature during the precipitation of calcite can be estimated by measuring the d18O value in calcite-secreting organisms (foraminifera, corals, and mollusks) and the value of the water in which the organisms live. The various paleotemperature equations all follow the original proposed by Sam Epstein and his colleagues (University of Chicago): T ¼ 16:5 4:3 d18 Ocalcite d18 Owater 2 þ0:14 d18 Ocalcite d18 Owater

½2

where T and d18Owater are the temperature (1C) and oxygen isotope value of the water in which the


CENOZOIC CLIMATE – OXYGEN ISOTOPE EVIDENCE

d18O Variation in the Natural Environment d18Owater values in the ocean/atmosphere system vary both spatially and temporally because fractionation between the H2 18 O and H2 16 O molecules is temperature-dependent in the hydrologic cycle and follows the Rayleigh Distillation model (Figure 1). In general, water vapor evaporates at low latitudes and

1 18

d O calcareous deposits are commonly reported relative to a carbonate standard, PDB (Peedee belemnite), and not SMOW (Standard Mean Ocean Water). PDB is 22% relative to SMOW.

0

Liquid

_ 10 Water v ap

or

˚

_ 20

18

O%

organism lived1 and d18Ocalcite is the oxygen isotope value of calcite measured in the mass spectrometer. Eqn [2] shows that the changes in d18Ocalcite are a function of the water temperature and d18Owater value. A one-to-one relationship between d18Ocalcite and d18Owater values dictates that a change in the d18Owater term will cause a similar change in the measured d18Ocalcite value. However, an inverse relationship between d18Ocalcite and T changes dictates that for every 11C increase in temperature, there is a 0.23% decrease in the measured d18Ocalcite value. These relationships enable us to interpret d18Ocalcite changes generated from foraminifera, corals, and mollusks. For many years, the convention was to plot d18Ocalcite values with the axis reversed (higher values to the left or bottom) so that d18O records reflect climate changes (e.g., colder to the left or bottom). More recently, there has been a trend among some scientists to plot d18Ocalcite values without reversing the axis. The paleotemperature equation contains two unknowns (temperature, d18Owater). Although temperature is the main target in reconstructions, one cannot ignore the d18Owater term. In the modern ocean, the equator-to-pole gradient measured in planktonic foraminifera d18Ocalcite values is B5.0% and largely reflects the temperature gradient (B281C). However, if temperature were the sole influence on modern d18Ocalcite values, the equator-topole gradient would be B6.5% (281C 0.23%/1C). The attenuated d18Ocalcite gradient measured in planktonic foraminifera reflects the surface ocean d18Owater variability. Therefore, a key to using d18Ocalcite records as indicators of past climates is to understand the hydrographic parameters that produce the modern d18Ocalcite gradient. For instance, ignoring the d18Owater term results in a 5–61C underestimation compared to the observed temperature gradient. This occurs largely because tropical temperature estimates will be too cold (B41C) whereas polar estimates will be warm (B1–21C).

503

_ 30 _ 40 _ 50 _ 60 1.0

0.8

0.6 0.4 Fraction remaining

0.2

0

Figure 1 Rayleigh distillation model showing the effects of evaporation and precipitation on the d18O values in the vapor and liquid phases. The initial conditions are a temperature of 251C and d18Owater value of 0%. This model also assumes that it is a closed system, meaning that water vapor is not added once the cloud moves away from the source regions. As clouds lose moisture, fractionation during the condensation further lowers the d18Owater value in the water vapor.

precipitates at higher latitudes. Fractionation during evaporation concentrates the lighter H2 16 O molecule in the water vapor, leaving the water enriched in H2 18 O and H2 16 O. On average, the d18Owater value of water vapor is 9% lower than its source water (Figure 1). Fractionation during condensation concentrates the H2 18 O molecules in the precipitation (rain/snow) by B9%. Therefore, if all of the water evaporated in the tropics rained back into the tropical oceans, there would be no net change in the d18Owater term. However, some water vapor is transported to higher latitudes. If the clouds remain a closed system (i.e., mid-to-high latitude evaporation does not influence the d18Owater value in the clouds2), then precipitation will further deplete the clouds (water vapor) in H2 18 O relative to H2 16 O. Consequently, the d18O value of water vapor decreases from the original value as water vapor condenses into precipitation (Figure 1) and the cloud that formed from the evaporation in the tropics will eventually lose moisture, fractionating the d18Owater value of the remaining water vapor (Figure 1 and 2).

2 Many island or coastal regions have significantly higher d18Owater values relative to continental locations at similar latitudes. This occurs because local evaporation increases the d18Owater values, thus resetting the initial conditions for Rayleigh distillation to occur.


504


100% of water vapor remaining 18 O water vapor = _ 9%

10% of water vapor remaining 18 Owater vapor = _ 30%

50% of water vapor remaining 18 O water vapor = _ 15%

˚

˚

˚

Evaporation

18 Osnow = _ 20%

˚

18

Orain = _ 6% 18

Orain = 0%

˚

˚

18

Ice sheet O _ 20 to _ 60%

˚

20˚

EQ

40˚ Latitude

80˚

60˚

Figure 2 Illustration of the Rayleigh distillation process on d18O values as clouds move over land and into the polar regions. Decreasing air temperatures cause moisture to rain/snow out of the cloud. Fractionation of the oxygen isotopes during condensation further decreases values. By the point that a cloud reaches the high latitudes, less than 10% of the original water vapor remains. Snowfall on Antarctica has values between 20 and 60%.The average d18O value for ice on Antarctica is B 40%.

5

0

˚

O of precipitation SMOW (% )

0 _5 _ 10

18

O%

˚

_ 15

_ 20 _ 25

_ 35 _ 40 90 S (A)

_ 20

_ 30

18

_ 30

_10

70

50

30 10

10

30

50

_ 40 _ 40 _ 30

70 90 N

Latitude

(B)

_ 20

_ 10

0

10

20

30

Surface air temperature (˚C)

Figure 3 Mean annual d18O water of precipitation (rain/snow) versus mean annual temperatures. The correlation between d18O values and latitude (A) is a function of temperature (B). The rainout/fraction of water remaining, and hence the fraction of d18O values, is determined by the cloud temperatures. Latitude is the dominant effect shown here. The scatter among sites at similar latitude results from elevation differences as well as differences in the distance from the ocean.

By the time 50% of the initial moisture precipitates, the d18O value of the water vapor will be B 15%, while precipitation will be B 6%. Once the cloud reaches the poles, over 90% of the initial water vapor will have been lost, producing d18O values of snow less than 20%. Snow at the South Pole approaches 60%. There is a strong relationship between d18O values in precipitation and air mass temperatures because air temperature dictates how much water vapor it can hold, and the d18O values of

the precipitation is a function of the amount of water remaining in the clouds (Figure 3).

Spatial Variations in d18Owater of Modern Sea Water The evaporation/precipitation process that determines the d18Owater values of precipitation (e.g., Figure 1) also controls the d18Owater values in regions



Temporal Variations Variations in the amount of water stored on land through time, usually in the form of ice, can have a significant effect on the mean ocean d18Owater value, and hence, the marine d18Ocalcite record. At present, high-latitude precipitation returns to the oceans through summer ice/snow melting. During glacial periods, snow and ice accumulate into large ice sheets. Because the difference in ice sheet and mean ocean values is large (d18Oice ¼ 35 to 40% vs. d18Owater mean ocean ¼ B0%), ice sheet fluctuations are reflected in mean oceanic d18Owater values. This relationship can be illustrated by examining how the mean ocean d18Owater value increased during the last glacial maximum (LGM) relative to the present (Figure 5). During the LGM, water stored in continental ice lowered global sea level by 120 m, removing B3% of the ocean’s volume. Thus, the mean ocean d18Owater value increased by 1.2% during the LGM relative to the present (Figure 5).

38

˚

Salinity (% )

37 36 35 34 33 32 80˚S 60˚S 40˚S 20˚S 0˚ 20˚N 40˚N 60˚N 80˚N Latitude

(A)

1.5

18

O%

˚

1.0 0.5 0.0 _ 0.5 (B)

_ 1.0 80˚S 60˚S 40˚S 20˚S 0˚ 20˚N 40˚N 60˚N 80˚N Latitude 1.5 1.0

O%

˚

18

18

in the ocean. At any one time, the volume of water being transported through the hydrologic cycle (e.g., atmosphere, lakes, rivers, and groundwater) is small compared to the volume of water in the oceans (1:130). Therefore, the hydrologic cycle can influence the whole ocean d18Owater value only by creating a new or enlarging an existing reservoir (e.g., glacier/ice sheets). In contrast, evaporation/precipitation processes will change the d18Owater and salinity values in the surface waters because only the thin surface layer of the ocean communicates with the atmosphere. As noted above, the process of evaporation enriches surface water in H2 18 O molecules and salt because the water vapor is enriched in H2 16 O molecules. For this reason, high salinity sea water has a high d18Owater value and vice versa. More specifically, tropical and subtropical surface water d18Owater values are B1% higher than mean ocean water values (Figure 4). Interestingly, subtropical d18Owater values are higher than tropical values even though evaporation is higher in the tropics. Atmospheric circulation patterns produce intense rainfall in the tropics to offset some of the evaporation, whereas very little rain falls in the subtropical regions. Because evaporation minus precipitation (E P) is greater in the subtropics, these regions have higher salinity and d18Owater values. In contrast, subpolar and polar regions have greater precipitation than evaporation; hence, highlatitude surface waters have low salinity and d18Owater values that approach 0.5% (Figure 4).

505

0.5

O= 0.5 Salinity

0.0

_ 0.5 _1.0 32 (C)

33

34

36 35 Salinity (PSU)

37

38

Figure 4 The salinity (A) and d18Owater values (B) measured in the open Atlantic () and Pacific (J) Oceans. Note the higher values in the tropical and subtropical region relative to the subpolar and polar regions. Evaporation and precipitation/runoff processes produce similar patterns in salinity and d18Owater values which is illustrated by the linearity in the d18O versus linity plot (C). The ocean-to-ocean difference between the Atlantic and Pacific results from a net transfer of fresh water from the Atlantic to the Pacific.

Pleistocene Oxygen Isotope Variations The first systematic downcore examination of the marine stable isotope record was made by Cesare´ Emiliani during the 1950s on d18Ocalcite records generated from planktonic foraminifera in Caribbean deep-sea cores. Emiliani recognized the cyclic pattern of low and high d18Ocalcite values and concluded that these represented glacial–interglacial intervals. Emiliani identified the seven most recent climate cycles and estimated that they spanned the last 280 000 years. (Current age estimates indi-


506


Present Mean depth of oceans = 3800 m 18

Sea water 18 O = 0.0%˚

Glacial ice O = _ 40%˚

O = 0.0%

˚

20 000 y BP

18

ΔSea level = 120 m Sea water 18 O = 1.2%

˚

120 m / 3800 m = 0.03 18

Δ O water = 1.2%

˚

0.0% _ ( _ 40% × 0.03)

˚

˚

Figure 5 The effect of building or removing large ice sheets on the d18O composition of the ocean can be significant. The removal of 3% of the ocean’s water during the last glacial maximum lowered sea level by 120 m. The d18O difference between the ocean and the ice is 40%, causing a whole ocean d18O change of 1.2%. The reverse process occurs during the melting of large ice sheets. If the Antarctic and Greenland ice were to melt, then sea level would rise B70 m. The volume of water stored in these ice sheets is equivalent to B2% of the water in the ocean. Therefore, the mean d18O value of the ocean would decrease by 0.7–0.8% (relative to PDB).

cate that the duration of the cycles is approximately 525 000 years.) To apply the paleotemperature equation to these records, Emiliani estimated that ice sheet-induced ocean d18Owater variability was relatively small, 0.3%. (As shown above, the maximum glacial–interglacial ice sheet signal was closer to 1.2%.) Therefore, most of the d18Ocalcite variability between glacial and interglacial intervals represented temperature changes of 5–101C. Emiliani divided the d18Ocalcite record into warm stages (designated with odd numbers counting down from the Holocene) and cold stages (even numbers). Hence, ‘Isotope Stage 1’ refers to the present interglacial interval and ‘Isotope Stage 2’ refers to the LGM (Figure 6). During the 1960s and 1970s, many argued that most of the glacial to interglacial difference in d18Ocalcite values resulted from ice volume changes. Nicholas Shackleton of Cambridge University made the key observation that benthic foraminiferal d18O values show a glacial to interglacial difference of B1.8%. If the ice volume contribution was only 0.3% as argued by Emiliani, then the deep ocean temperatures would have been 6–71C colder than the present temperatures of 0–31C. Sea water freezes at 1.81C, precluding Emiliani’s ‘low’ ice volume estimate. By the early 1970s, numerous d18O records had been generated and showed a cyclic variation through the Pleistocene and into the late Pliocene. One hundred oxygen isotope stages, representing

50 glacial–interglacial cycles, have been identified for the interval since 2.6 million years ago (Ma) (Figure 6).

Cenozoic d18O Records The first Cenozoic d18O syntheses based on foraminiferal d18O records were produced during the mid-1970s. Nicholas Shackleton and James Kennett produced a composite benthic d18O record for the Cenozoic from cores to the south of Australia. A second group led by Samuel Savin generated lowlatitude planktonic and benthic foraminiferal d18O syntheses. Both records are important to understanding Cenozoic climate changes. Benthic foraminiferal records best reflect global temperature and ice volume changes. Additional advantages of the benthic foraminiferal composite include: (1) deepocean temperatures are more uniform with respect to horizontal and vertical gradients; (2) deep-ocean d18Owater values are less variable compared to the large surface water changes; (3) the deep ocean approximates high-latitude surface water conditions where deep waters originated during the Cenozoic (i.e., Antarctica, northern North Atlantic); and (4) many benthic foraminifera taxa are long-lived so that one species can be used to construct records spanning several millions of years in contrast to planktonic taxa



507

_2

Planktonic foraminifera

3

0

Benthic foraminifera 4

18

O% benthic foraminifera ˚

18

O% planktonic foraminifera ˚

_1

2

5 123 4 5 6

7

9

11

13

15

17 19 21

23

25 27 29 31 33 35 37 39 4143 45 4749

Pleistocene

Holocene 0.5

0 (A)

1.0

1.5

Age (Ma)


_2

Benthic foraminifera

3

0

4

18

O% benthic foraminifera ˚

18

O% planktonic foraminifera ˚

_1

2

5

96 98 100 515355 57 59 61 63 65 67 69 71 7375 77 79 8183 85 87 89 919395

Pleistocene 1.5 (B)

Pliocene 2.0

2.5

3.0

Age (Ma)

Figure 6 Planktonic and benthic foraminiferal d18O records for the last 3 million years. Note the high frequency signals in the records. For the interval between 3 and 1 Ma, a 40 000 year cycle dominates the records. After 1 Ma, the beat changes to a 100 000 year cycle and the amplitudes increase (relative to PDB).

which have shorter durations and require records to be spliced together from several species. Low-latitude planktonic foraminiferal d18O records are good proxies for tropical sea surface temperatures. Tropical temperatures are an important component of the climate system because they influence evaporation, and hence, total moisture in the atmosphere. Planktonic and benthic foraminiferal d18O comparison allows one to assess equator-topole as well as vertical temperature gradients during the Cenozoic, and thus, to determine planetary temperature changes. Finally, much of the climatic change in the last 65 million years has been ascribed

to poleward heat transport or greenhouse gas fluctuations. General circulation models indicate that each mechanism should produce different temperature patterns that can be approximated with the planktonic and benthic d18O records. The first benthic d18O syntheses generated, as well as more recent compilations, show the same longterm patterns. After the Cretaceous–Tertiary (K/T) boundary events, deep-water d18O values remained relatively constant for the first 7 million years of the Paleocene (Figure 7A). At 58 Ma, benthic foraminiferal d18O values began a decrease over the next 6 My that culminated during the early Eocene


508


_2


_1

18

O%

˚

0 1 2 3

Benthic foraminifera

Pleist.

4

0

5

Oligocene

Miocene

Plio. 10

15

20

25

35 30 Age (Ma)

(A)

Eocene 40

45

Paleocene 50

55

65

60

30

25

Temperature (˚C)

20

15

10

5

Pleist.

0

0 (B)

5

Oligocene

Miocene

Plio. 10

15

20

25

35 30 Age (Ma)

Eocene 40

45

Paleocene 50

55

60

65

Figure 7 (A) Planktonic and benthic foraminiferal d18O composite records representing the tropical surface and deep ocean conditions (relative to PDB). The thick line through both records was generated using a 1 million year Gaussian filter. (B) Temperature estimates based on planktonic and benthic records and ice volume estimates discussed in the text.

with the lowest recorded value ( 0.5%) of the Cenozoic. Following this minimum at 52 Ma, d18O values increased by 5.5%, recording maximum values (B5%) during the glacial intervals of Pleistocene (Figure 6). The first part of this long-term change was a gradual increase of 2% through the end of the Eocene (52–34 Ma). The remainder of the increase was accomplished through large steps at the Eocene/ Oligocene boundary (B33.5 Ma), during the middle Miocene (ca. 15–13 Ma) and late Pliocene (ca. 3.2– 2.6 Ma). After 2.6 Ma, the amplitude of the high-

frequency signal increased to>1%, reaching 1.8% over the past 800 thousand years. Planktonic and benthic foraminiferal d18O values co-varied in general during the early Cenozoic (6.5– 34 Ma). Values averaged about 1% between 65 and 58 Ma, before decreasing to 2.5% during the early Eocene, recording the lowest values of the Cenozoic (Figure 7A). From 52 to 33 Ma, planktonic foraminiferal values increased by 2%. In spite of a break in the latest Eocene record, it appears that the tropical ocean differed from the deep ocean across



the Eocene/Oligocene boundary. For much of the Oligocene (B33–25 Ma), planktonic foraminiferal d18O values remained unusually high, averaging 0.5%. Beginning around the Oligocene/Miocene boundary (B25 Ma), planktonic foraminiferal d18O values began a long-term decrease, culminating in the Pleistocene with average values of 1.5%. In contrast, the benthic d18O record permanently changed during the middle Miocene d18O shift and late Pliocene increase. Apportioning the d18O changes recorded by the benthic and planktonic foraminifera between temperature and ice volume changes requires knowledge of, or reasonable estimates for, one of these parameters. One promising tool that may help discriminate between each effect is the Mg/Ca ratio measured in benthic foraminifera. Initial studies using Mg/Ca ratios confirmed the long-term temperature changes during the Cenozoic calculated using the d18O record and other climate proxies. If verified, this record implies that small ice sheets grew during the middle and late Eocene and fluctuated in size throughout the Oligocene to Miocene. At present, the Mg/Ca record lacks the resolution for key intervals and still requires verification of interspecies offsets before it can be applied unequivocally to isolate the ice volume-induced d18Owater component in the foraminiferal d18O records. For the discussion that follows, glacialogical evidence is used to estimate the ice volume/d18Owater variations.

The Greenhouse World The oldest unequivocal evidence for ice sheets on Antarctica, ice-rafted detritus (IRD) deposited by icebergs in the ocean, places the first large ice sheet in the earliest Oligocene. Thus, it is reasonable to assume that ice sheets were small to absent and that surface and deep-water temperature changes controlled much if not all the d18O change prior to 34 Ma. The modern Antarctic and Greenland ice sheets lock up B2% of the total water in the world’s ocean. If melted, these ice sheets would raise global sea level by B70–75 m and mean ocean d18Owater value would decrease to 0.9% PDB (see above). One can then apply eqn [2] to the benthic and planktonic foraminiferal d18O records to estimate deep- and surface-ocean temperatures for the first half of the Cenozoic (c. 65–34 Ma). During the early to middle Paleocene, deep-water temperatures remained close to 101C (Figure 7B). The 1% decrease between 58 and 52 Ma translates into a deep-water warming of 41C, reaching a high of 141C. This is in sharp contrast to the modern deep-

509

water temperatures, which range between 0 and 31C. Following the peak warmth at 52 Ma, the 2% increase in benthic foraminiferal values indicates that the deep waters cooled by 71C and were 71C by the end of the Eocene. If small ice sheets existed during the Paleocene and Eocene, then temperature estimates would be on the order of 11C warmer than those calculated for the ice-free assumption. (Some data indicate that smaller ice sheets may have existed on the inland parts of Antarctica during the late Eocene. However, these were not large enough to deposit IRD in the ocean. Therefore, their effect on the d18O values of the ocean was probably less than 0.3%.) Tropical surface water temperatures warmed from 22 to 241C, based on eqn [2], at the beginning of the Cenozoic to 281C during the early Eocene (52 Ma; Figure 7B). The higher estimate is similar to temperatures in the equatorial regions of the modern oceans. Planktonic foraminiferal d18O values recorded a long-term increase of by 2% ( 2.5 to 0.5%) through the remainder of the Eocene. Just prior to the Eocene/Oligocene boundary, tropical surface water temperatures were B211C, ending the long-term tropical cooling of 71C from 52 to 34 Ma.

The Ice House World of the Last 33 Million Years As mentioned above, southern ocean cores contain IRD at and above the Eocene/Oligocene boundary. Widely distributed IRD and glacial tills on parts of the Antarctic continental margin representing the Oligocene to Recent mark the onset of large ice sheets. Whether these sediments represent persistent or periodic ice cover is uncertain. At least some ice was present on Antarctica during the Oligocene to early Miocene. The Antarctic ice sheet has been a fixture since the middle Miocene (B15 Ma). Our record of Northern Hemisphere ice sheets suggests that they were small or nonexistent prior to the late Pliocene. For the purpose of estimating surface and deep temperatures, an ice volume estimate slightly lower than the modern will be applied for the interval that spans from the Oligocene into the middle Miocene (33–15 Ma). For the interval between 15 and 3 Ma, ice volumes were probably similar to those of today. From 3 Ma, ice volumes ranged between the modern and LGM. Using these broad estimates for ice volumes, mean ocean d18Owater values for those three intervals were 0.5, 0.22, and 0.4% PDB, respectively. The 0.4% estimate reflects the average between the maximum and minimum conditions during the Plio-Pleistocene. As noted


510


above, the largest portion of the high frequency signal is controlled by ice volume changes. The benthic foraminiferal d18O increase at the Eocene/Oligocene boundary occurred rapidly (B10 000 years; Figure 8A). At the peak of the Eocene/Oligocene boundary event, benthic foraminifera recorded d18O values similar to modern values. Using the ice volume assumption from above, deepwater temperatures approached modern deep-ocean temperatures (31C). This marks an important transition from the relatively warm oceans of the Paleocene and Eocene to the cold deep waters of the Oligocene to present. This switch to a cold ocean where bottom waters formed at near-freezing temperatures heralded the development of the psychrosphere. Following the Eocene/Oligocene boundary, deep-water temperatures began a long-term warming over the next 18 million years (33–15 Ma). The coldest deep-water temperatures of 31C were recorded at 33 Ma, while temperatures reached 91C at B25 and B15 Ma (Figure 7B).

There is a gap in the planktonic foraminiferal d18O record for the latest Eocene that hampers our assessment of tropical response during Eocene/ Oligocene climate event. However, it is clear from the data that do exist that the planktonic response across the Eocene/Oligocene boundary differed from the benthic response. The planktonic foraminiferal d18O values for the early Oligocene are similar to late Eocene values, whereas the benthic values recorded a 1.5% increase. Planktonic foraminiferal records from other regions that span the Eocene/Oligocene boundary indicate that the surface water d18O increase was on the order of 0.5%. This change is approximately equal to the effect of the modern Antarctic ice sheet. Combined with the physical evidence, it seems probable that the planktonic foraminiferal d18O increase at the Eocene/Oligocene boundary recorded the ice volume influence with little temperature effect. Therefore, tropical surface temperatures remained around 221C while the deep ocean cooled during

1.5

18

O%

˚

2.0 2.5 3.0 33.00

33.50 Age (Ma)

33.25

(A)

33.75

34.00

14.0

14.5

1.0

O%

˚

1.5

18

2.0 2.5 3.0 12.5

13.0

13.5 Age (Ma)

(B) 2.5

3.5

18

O%

˚

3.0

phere

4.0

Hemis g No.

tion

glacia

sin

Increa

4.5 2.4 (C)

2.5

2.6

2.7

2.8 2.9 Age (Ma)

3.0

3.1

3.2

Figure 8 High-resolution d18O records representing the Eocene/Oligocene boundary (A), middle Miocene (B), and late Pliocene (C) d18O shifts (relative to PDB).



this d18O shift. Following the boundary event, planktonic foraminifera d18O record during the Oligocene and early Miocene mirrored the benthic record in many respects. For much of the Oligocene and early Miocene, the absolute values are close to 0.5%, which translates into a temperature estimate of 211C (Figure 7B). By 15 Ma, tropical surface waters had warmed to 261C. The middle Miocene d18O shift represents an increase of 1.5% in the benthic record between 15 and 13 Ma. This transition is composed of two sharp increases around 14 and 13 Ma (Figure 8B). These d18O steps occurred in less than 200 000 years with each recording an increase of B1% followed by a small decrease. During these two shifts, deep waters cooled from 9 to 51C. The planktonic foraminiferal d18O record from 15 to 13 Ma shows two increases as recorded in the benthic foraminiferal record (Figure 7). However, it does not show the large permanent shift recorded by the benthic foraminifera, indicating a small cooling from 26 to 241C. From 13 to 3 Ma, the deep ocean cooled slightly from 5 to 31C while the surface waters warmed from 24 to 261C (Figure 7B). The last of the large d18O steps in the Cenozoic was recorded during the late Pliocene from 3.2 to 2.6 Ma. This ‘step’ is better characterized as a series of d18O cycles with increasing amplitudes and values over this interval (Figures 6 and 8C). The cycles have been subsequently determined to be 40 000 year cycles related to variations in the solar radiation received in the high latitudes. This interval ushered in the large-scale Northern Hemisphere ice sheets that have since dominated Earth’s climate. At 2.6 Ma, the first IRD was deposited in the open North Atlantic and was coeval with the d18O maximum. Prior to 2.6 Ma, IRD was confined to the marginal basins to the north, Greenland’s and Iceland’s continental margins. Subsequent d18O maxima were associated with IRD. Between 2.6 and 1 Ma, large Northern Hemisphere ice sheets waxed and waned on the 40 000 year beat. Beginning around 1 Ma, the ice sheets increased in size and switched to a 100 000 year beat (Figure 6). During this interval, deep-water temperatures remained similar to those in the modern ocean (0 to 31C). The planktonic foraminiferal d18O response during the late Pliocene event shows the cyclic behavior, but not the overall increase recorded by the benthic foraminifera. As with the middle Miocene d18O shift, the late Pliocene increase represents the cyclic buildup of ice sheets accompanied by deep water cooling. The tropical surface water temperatures, however, varied between 26 and 281C.

511

Mechanisms for Climate Change Most climate change hypotheses for the Cenozoic focus on either oceanic heat transport and/or greenhouse gas concentrations. Each mechanism produces different responses in the equatorial-to-pole and surface-to-deep temperature gradients. An increase in the meridional heat transport generally cools the tropics and warms the poles. If poleward transport of heat decreases, then the tropics will warm and the poles will cool. Variations in greenhouse gas concentrations should produce similar changes in both the tropical and polar regions. Tropical surface water and deep-ocean records covaried for the first part of the Cenozoic. The warming and subsequent cooling between 65 and 34 Ma are most often ascribed to changing greenhouse gas concentrations. The interval of warming that began around 58 Ma and peaked at 52 Ma coincided with the release of large amounts of CO2 into the atmosphere as a consequence of tectonic processes. The eruption of the Thulean basalts in the northeastern Atlantic Ocean began during the Paleocene and peaked around 54 Ma. It is also recognized that there was a large-scale reorganization of the midocean ridge hydrothermal system which began during the late Paleocene and extended into the Eocene. Both tectonic processes accelerate mantle degassing which raises atmospheric levels of CO2. Recently, evidence for another potentially large CO2 reservoir was found along the eastern continental margin of North America. Methane hydrates frozen within the sediments appear to have released catastrophically at least once and possibly multiple times during the latest Paleocene and early Eocene (B58–52 Ma). One or all of these sources could have contributed to the build-up of greenhouse gases in the atmosphere between 58 and 52 Ma. Following the thermal maximum, the long-term cooling in both the surface and deep waters implies that greenhouse gas concentrations slowly decreased. Proxies for estimating pCO2 concentrations (d13C fractionation within organic carbon and boron isotopes) are still being developed and refined. However, preliminary indications are that atmospheric pCO2 levels were high (>1000 p.p.m.) during the early Eocene, dropped to B400–500 p.p.m. during the middle to late Eocene, and reached late Pleistocene concentrations (200–300 p.p.m.) by the early Oligocene. The deep-water temperature cooling across the Eocene/Oligocene boundary (Figure 7B) was not accompanied by tropical cooling, and resulted from the first step in the thermal isolation of Antarctica. In modern ocean, the Antarctic Circumpolar Current is a vigorous surface-to-bottom current that provides


512


an effective barrier to southward-flowing warm surface waters. The development of this current during the Cenozoic hinged on the deepening for the Tasman Rise and opening of the Drake Passage. Recent drilling indicates that marine connections developed across the Tasman Rise at or near the Eocene/ Oligocene boundary (33.5 Ma). Tectonic constraints on the separation of the Drake Passage are less precise. Estimates range from 35 to 22 Ma for the opening of this gateway. The uncertainty lies in the tectonic complexity of the region and what constitutes an effective opening for water to flow through. The climatic consequence of creating a circumpolar flow was to thermally isolate Antarctica and promote the growth of the Antarctic ice sheet. As noted above, the first large ice sheet grew at the beginning at the Eocene/Oligocene boundary. The most notable divergence in the d18O records occurred during the middle Miocene (B15 Ma). For the first time during the Cenozoic, the tropical surface and deep waters recorded a clear divergence in d18O values, a trend that increased in magnitude and reached a maximum in the modern ocean. Any poleward transport of heat appears to have been effectively severed from Antarctica by 15 Ma, promoting further cooling. On the other hand, the tropics have been warming over the past 15 My. A combination of different factors fueled this warming. First, less heat was being transported out of the lowand mid-latitude regions to the high southern latitudes. Second, the opening of the Southern Hemisphere gateways that promoted the formation of the circumpolar circulation led to the destruction of the circumequatorial circulation. The effects of the closure of the Tethys Ocean (predecessor to the Mediterranean), shoaling of the Panamanian Isthmus (4.5–2.6 Ma), and constriction in the Indonesian Passage (B3 Ma to present) allowed the east-to-west flowing surface waters in the tropics to ‘pile’ up and absorb more solar radiation. A consequence of the equatorial warming and high-latitude cooling was an increase in the equator-to-pole temperature gradient. As the gradient increased, winds increased, promoting the organization of the surface ocean circulation patterns that persist today.

In planktonic foraminifera, variations between species can be as great as 1.5%. For both the planktonic and benthic foraminifera, interspecific differences are as large as the glacial–interglacial signal. These interspecific d18Ocalcite variations are often ascribed to a vital effect or kinetic fractionation of the oxygen isotopes within the organism. However, some of the differences in the planktonic taxa results from different seasonal or depth habitats and therefore provides important information about properties in the upper part of the water column. It is noteworthy that the first d18O syntheses were based on mixed species analyses and yet basic features captured in these curves still persist today. This attests to the robustness of these records and method for reconstructing climate changes in the ocean. The high-frequency signal that dominates the late Pliocene to Pleistocene records is also present in the Miocene and Oligocene intervals. The cloud of points about the mean shown in Figure 7 reflects records that were sampled at a resolution sufficient to document the high frequency signal. For the interval between 35 and 1 Ma, the benthic foraminiferal d18O record has a 40 000 year frequency superimposed on the long-term means that are represented by the smoothed line. The origin of the 40 000 year cycles lies in variations in the tilt of the earth’s axis that influences the amount of solar radiation received in the high latitudes. This insolation signal is transmitted to the deep ocean because the high latitudes were the source regions for deep waters during much, if not all, of the Cenozoic. The record prior to 35 Ma is unclear with regard to the presence or absence of 40 000 year cycles.

See also Holocene Climate Variability. Icebergs. Iceinduced Gouging of the Seafloor. Oxygen Isotopes in the Ocean. Sea Ice: Overview. Sub Ice-Shelf Circulation and Processes.

Further Reading Some Caveats A concern in generating marine isotope records is that the isotopic analyses should be made on the same species. This is important because d18Ocalcite values can vary among the different species of organisms. Coexisting taxa of benthic foraminifera record d18O values that can differ by as much as 1%.

Craig H (1957) Isotopic standards for carbon and oxygen and correction factors for mass spectrometric analysis of carbon dioxide. Geochemica et Cosmochemica Acta 12: 133--149. Craig H (1965) The measurement of oxygen isotope paleotemperatures. In: Tongiorgi E (ed.) Stable Isotopes in Oceanographic Studies and Paleotemperatures, pp. 161--182. Spoleto: Consiglio Nazionale delle Ricerche, Laboratorio di Geologica Nucleare, Pisa.



Craig H and Gordon LI (1965) Deuterium and oxygen-18 variations in the oceans and marine atmosphere. In: Tongiorgi E (ed.) Stable Isotopes in Oceanographic Studies and Paleotemperatures, pp. 1--122. Spoleto: Consiglio Nazionale delle Ricerche, Laboratorio di Geologica Nucleare, Pisa. Emiliani C (1955) Pleistocene temperatures. Journal of Geology 63: 539--578. Epstein S, Buchsbaum R, Lowenstam H, and Urey HC (1953) Revised carbonate-water temperature scale. Bulletin of the Geological Society of America 64: 1315--1326. Fairbanks RG, Charles CD, and Wright JD (1992) Origin of Melt Water Pulses. In: Taylor RE, Long A, and Kra RS (eds.) Radiocarbon After Four Decades, pp. 473--500. New York: Springer-Verlag. Imbrie J, Hays JD, Martinson DG, et al. (1984) The orbital theory of Pleistocene climate: support from a revised chronology of the marine d18O record. In: Berger AL, Imbrie J, Hays JD, Kukla G, and Saltzman B (eds.) Milankovitch and Climate, part I, pp. 269--305. Dordrecht: Reidel. Lear CH, Elderfield H, and Wilson PA (1999) Cenozoic deep-sea temperatures and global ice volumes from Mg/ Ca in benthic foraminiferal calcite. Science 287: 269--272. Miller KG, Fairbanks RG, and Mountain GS (1987) Tertiary oxygen isotope synthesis, sea-level history, and continental margin erosion. Paleoceanography 2: 1--19. Miller KG, Wright JD, and Fairbanks RG (1991) Unlocking the Ice House: Oligocene–Miocene oxygen isotopes, eustasy, and margin erosion. Journal of Geophysical Research 96: 6829--6848. Pagani M, Arthur MA, and Freeman KH (1999) Miocene evolution of atmospheric carbon dioxide. Paleoceanography 14: 273--292. Palmer MR, Pearson PN, and Cobb SJ (1998) Reconstructing past ocean pH-depth profiles. Science 282: 1468--1471. Pearson PN and Palmer MR (1999) Middle Eocene seawater pH and atmospheric carbon dioxide concentrations. Science 284: 1824--1826.

513

Rozanski K, Araguas-Araguas L and Gonfiantini R (1993) Isotopic patterns in modern global precipitation. In: Swart PK, McKenzie J and Savin S (eds) Climate Change in Continental Isotopic Records. Geophysical Monograph 78, pp. 1–35. Washington, DC: American Geophysical Union. Rye DM and Sommer MA (1980) Reconstructing paleotemperature and paleosalinity regimes with oxygen isotopes. In: Rhoads DC and Lutz RA (eds.) Skeletal Growth of Aquatic Organisms, pp. 162--202. New York: Plenum. Savin SM, Douglas RG, and Stehli FG (1975) Tertiary marine paleotemperatures. Geological Society of America Bulletin 86: 1499--1510. Shackleton NJ (1967) Oxygen isotope analyses and Pleistocene temperatures re-assessed. Nature 215: 115--117. Shackleton NJ, Berger A, and Peltier WR (1990) An alternative astronomical calibration of the Lower Pleistocene time scale based on ODP Site 677. Transactions of the Royal Society of Edinburgh, Earth Science 81: 251--261. Shackleton NJ and Kennett JP (1975) Paleotemperature history of the Cenozoic and initiation of Antarctic glaciation. Oxygen and carbon isotopic analysis in DSDP Sites 277, 279, and 281. Initial Report Deep Sea Drilling Project 29: 743--755. Shackleton NJ and Opdyke ND (1973) Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28-238. Oxygen isotope temperatures and ice volumes on a 105 year and 106 year scale. Quaternary Research 3: 39--55. Tiedemann RM, Sarnthein M, and Shackleton NJ (1994) Astronomic calibration for the Pliocene Atlantic d18O and dust flux records of Ocean Drilling Program Site 659. Paleoceanography 9: 619--638. Urey HC (1947) The thermodynamic properties of isotopic substances. Journal of the Chemical Society pp. 562--581.


CENOZOIC OCEANS – CARBON CYCLE MODELS

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 426–436, & 2001, Elsevier Ltd.

Introduction The story of the Cenozoic is essentially a story of global cooling. The last 65 million years of the Earth’s history mark the transition from the Cretaceous ‘greenhouse’ climate toward the present-day ‘icehouse’ conditions. Particularly, the cooling by about 8–101C of deep ocean waters since the Cretaceous was linked to a reorganization of the oceanic circulation triggered by tectonic plate movements. These oceanic circulation changes were coeval with continental climatic change, as demonstrated by abundant evidence for global cooling (pollen, faunal assemblages, development of glaciers, etc.). For instance, most of western Europe and the western United States had a subtropical climate during the Eocene, despite the fact that they were located at the same latitude as today. Another striking feature of the changes that have occurred during Cenozoic times is the decrease of the partial pressure of atmospheric CO2 (PCO2). Since CO2 is a greenhouse gas, there might be a causal relationship between the decrease in PCO2 and the general cooling trend of Cenozoic climate. The global cooling might be the result of the changes in oceanic circulation and atmospheric CO2, both probably influencing each other and possibly initiated by tectonic processes.

Indicators of Atmospheric CO2 Change It should be kept in mind that there are no direct proxies of ancient levels of CO2 in the atmosphere. Methods rely on three indirect indicators. 1. The d13C measured in ancient soil carbonates can be directly linked to the atmospheric PCO2. This method reveals declining atmospheric PCO2 over the last 65 million years, from about 650 ppm by volume (ppmv) in the Paleocene (Figure 1). 2. The biological isotopic fractionation occurring during assimilation of carbon by the marine biosphere (ep) depends on the partial pressure of CO2 dissolved in sea water, itself directly related to the

514

atmospheric PCO2. The estimation of ep for ancient organic sediments indicates high PCO2 values in the Eocene (620 ppmv), followed by a constant decline toward the present-day pressure through the Cenozoic (Figure 1). 3. The measured boron isotopic composition of marine carbonates gives insight into the pH of ancient sea water. Assuming a plausible history for the ocean alkalinity, lower or higher pH values can be respectively related to higher or lower PCO2. This method has been applied to the last 60 million years, showing values as high as 3500 ppmv CO2 during the Paleocene. The decline in PCO2 was then roughly linear through time until the late Eocene. During the last 25 million years of the Earth’s history, PCO2 was relatively constant, possibly displaying lower values than present-day ones during Miocene. No data are available for the Oligocene (Figure 1). Despite some disagreements between the three reconstructions, they all indicate a major reduction of the atmospheric CO2 partial pressure during the Cenozoic, which might potentially play an important role in the coeval global cooling. Any exploration of the cause of the decline in PCO2 requires the identification of the sources and sinks of oceanic and atmospheric carbon, and some knowledge of their relative changes during the Cenozoic. 4000 3500 3000 Atmospheric P co2 (ppmv)

L. François and Y. Godde´ris, University of Lie`ge, Lie`ge, Belgium

2500 2000 1500 1000 500 0

Late Cr

70

Paleoc

60

Eocene

50

Oligocene Cenozoic

40 30 Age (Ma)

Miocene

20

10

Pl

0

Figure 1 The reconstructed partial pressure of atmospheric CO2. D, from boron isotopes (Pearson and Palmer (2000)) ; , from ep (Kump and Arthur, 1997); &, from paleosoils analysis (compilation by Berner (1998)). Timescale according to Harland et al. (1990) (CR ¼ Cretaceous; PALEOC ¼ Paleocene; PL ¼ Pleistocene.)



515

Carbon Cycle Changes and Processes do not contain carbon. The budget can be written as Involved CaSiO3 ðrockÞ þ 2H2 CO3 ðatmosphereÞ þ H2 O

Long-term Regulation of Atmospheric CO2

On the geological timescale, and neglecting at this point the possible impact of sedimentary organic carbon cycling, the sources of carbon for the ocean– atmosphere system are the degassing of the mantle and metamorphic processes. Carbon is injected as CO2 into the ocean–atmosphere system today through the degassing of fresh basalts along midoceanic ridges (MOR) (1.5–2.2 1012 mol y1), and through plume events and arc volcanism (1.5– 5.5 1012 mol y1). These various sources account for the total degassing flux FVOL. Once released, this carbon is rapidly (within 103 years) redistributed between the atmosphere and the ocean, reaching a steady-state repartition after a negligible time compared to the geological timescale. Carbon can leave the system mainly through the deposition of carbonate minerals on the seafloor. The rate of carbon removal through carbonate deposition is controlled by the saturation state of the ocean, and thus by the rate of supply of alkalinity by the chemical weathering of continental minerals. Carbonate and silicate minerals exposed at the continental surface weather under the corrosive action of atmospheric CO2 dissolved in rain water as carbonic acid, and eventually concentrated by the microbial respiration in soils. Regarding the chemical weathering of carbonate minerals, the net budget of the dissolution reaction can be written as follows eqn [I]. CaCO3 ðrockÞ þ H2 CO3 ðatmosphereÞ -Ca2þ ðriversÞ þ 2HCO 3 ðriversÞ

½I

In this reaction, there is a net transfer of Ca2þ (or Mg2þ in the case of magnesium carbonates) from the continental crust to the ocean, a creation of two equivalents of alkalinity and a transfer to the river system of two moles of carbon per mole of Ca2þ, one coming from the crust, the other one from the atmosphere. Once the weathering products reach the ocean, they will increase the saturation state of surface waters with respect to calcite and induce rapidly (within 103 years) the biologically driven precipitation of one mole of CaCO3 followed by its deposition on the seafloor. The precipitation– deposition reaction is the reverse of reaction [1]. The net carbon budget of the weathering of carbonate minerals followed by deposition of sedimentary carbonate is thus equal to zero. The chemical weathering of continental silicate rocks is fundamentally different, since silicate rocks

-Ca2þ ðriversÞ þ 2HCO 3 ðriversÞ þ H4 SiO4 ðriversÞ ½II Here CaSiO3 stands for a ‘generic’ silicate mineral. The weathering reaction of more realistic Ca- (or Mg-) silicate minerals, if more complex, displays the same budget in terms of alkalinity versus carbon fluxes. Again, once reaching the ocean, the excess Ca2þ will precipitate as CaCO3, thus removing one mole of carbon from the ocean per mole of weathered silicates. The net budget of this reaction, after sedimentary carbonate precipitation, is the transfer of exospheric carbon to the crust. Chemical weathering of continental silicate minerals thus acts as the main sink of carbon on the geological timescale. Today, about 6 1012 mol y1 of Ca2þ and Mg2þ are released from silicate weathering. The size of the exospheric carbon pool (ocean þ atmosphere) is about 3.2 1018 mol today. As mentioned above, the fluxes entering and leaving this reservoir are of the order of 1012–1013 mol y1. A relatively small imbalance between the input and output of carbon of 1012 mol y1, the output being higher, but persisting for several million years, will result in a drastic reduction in the exospheric content. Three million years will be sufficient to remove all the carbon from the exospheric system, thus forcing the atmospheric PCO2 to zero. There is no lithological, fossil, or geochemical record of such a dramatic event during the Cenozoic, or event during the complete Phanerozoic. To avoid the occurrence of such events for which there is no evidence, the perturbations of the carbon cycle had to be limited in time and amplitude, and thus the past exospheric carbon cycle was not strictly at, but always close to, steady state. The same considerations apply to the alkalinity budget. These steady-state conditions require that the amount of carbon removed from the atmosphere–ocean system by continental silicate weathering must always closely track the amount of carbon released by degassing. Mathematically, these conditions translate into eqn [1]. FSW ¼ FVOL

½1

The question is now how to physically force FSW to follow the degassing. The answer lies in the fact that the chemical weathering of continental silicates appears to be dependent on air temperature, the dissolution being enhanced during warmer climates. This dependence provides a negative feedback that



not only allows equilibration of the carbon and alkalinity budgets on the geological timescale, but also stabilizes the Earth’s climate. When the degassing increases suddenly, for instance as a result of a higher spreading rate of the oceanic floor, the amount of carbon in the ocean and atmosphere will first increase, increasing the atmospheric PCO2. Because CO2 is a greenhouse gas, the climate will become warmer, and this will enhance the weathering of silicate rocks. As a result, any increase in the input of carbon will be counterbalanced by an increase of the output through silicate weathering, thus stabilizing the system through a negative feedback loop. The PCO2 will stabilize at a somewhat higher level than before the perturbation. Similarly, the decline in PCO2 through the Cenozoic could be due to a decreasing degassing rate, which acts as the driving force of changes. This simple process is the basis of all existing long-term geochemical cycle models. It was first identified in 1981 by Walker, Hays, and Kasting. Breaking this feedback loop would result in fluctuations in calculated PCO2 and thus presumably in climate, that are not reflected in the geological record. Himalayan Uplift, 87Sr/86Sr Record, and Possible Implications for Weathering History

M.E. Raymo in 1991 put forward another explanation of the global Cenozoic PCO2 decline. Instead of a decreased degassing rate, she suggested that continental silicate weathering rates increased drastically over the last 40 million years, although degassing conditions remained more or less constant. This assertion was originally based on the Cenozoic carbonate record of 87Sr/86Sr. The isotope 86Sr is stable, whereas 87Sr is produced by the radioactive b-decay of 87Rb. Strontium ions easily replace calcium ions in mineral lattices, since their ionic diameters are comparable. The present-day sea water 87Sr/86Sr equals 0.709. Two main processes impinge on this ratio: the chemical weathering of continents, delivering strontium with a mean 87Sr/86Sr of 0.712, and the exchanges between seafloor basalts and sea water, resulting in the release of mantle strontium into the ocean (87Sr/86Sr ¼ 0.703). In other words, chemical weathering of continental rocks tends to increase the strontium isotopic ratio of sea water, while exchanges with seafloor basalts at low or high temperature tend to decrease it. The sea water 87Sr/86Sr recorded over the last 65 million years displays a major increase, starting about 37–38 million years ago (Figure 2). An event approximately coeval with the sea water 87Sr/86Sr upward shift is the Himalayan uplift, which was

0.7095

Sea water 87Sr/ 86Sr

516

0.7090 0.7085 0.7080 0.7075

Late Cr

70

Paleoc

60

Eocene

50

Oligocene Cenozoic

40 30 Age (Ma)

Miocene

20

10

Pl

0

Figure 2 Sea water 87Sr/86Sr recorded in ancient carbonate sediments (Ottawa-Bochum database: http://www.science. uottawa.ca/geology/isotope_data/). Timescale abbreviations as Figure 1.

initiated by the India–Asia collision some 50 million years ago. Raymo has proposed that, in an uplifted area, the mechanical breakdown of rocks increases owing to the cooling and development of glaciers, to the development of steep slopes, and to temperatures oscillating below and above the freezing point at high altitudes. Furthermore, the development of the monsoon regime, about 10 million years ago, resulted in increased runoff, and thus an enhanced water availability for weathering, over at least the southern side of the Himalayan range. All these uplift-related changes might result in an enhanced chemical dissolution of minerals, since the surface in contact with the corrosive solutions increases when rocks fragment. The consequence of the Himalayan uplift might thus be an increase in the consumption of exospheric carbon by enhanced weathering on the continents, a process recorded in the sea water 87 Sr/86Sr rise. The system depicted in this hypothesis is new, compared to the hypothesis described in the previous section. Here, tectonic processes result in uplift, followed by enhanced weathering, itself consuming atmospheric CO2, thus cooling the climate. This cooling favors the development of glaciers not only in the uplifted area, but also globally, producing a global increase in mechanical and subsequent chemical weathering, a positive feedback that further cools the Earth. In Raymo’s hypothesis, the negative feedback proposed by Walker et al., stabilizing PCO2 no longer exists. Chemical weathering is mainly controlled by tectonic processes with high rates in a cool world (Raymo’s world), while it was controlled by climate and PCO2 with high rates in a warm world in the Walker hypothesis (Walker’s world). However, as mentioned above, negative feedbacks are needed to stabilize PCO2, especially since the degassing remained more or less constant over the period of



Lysocline and Carbonate Accumulation Changes

Other indicators of a possible increase in the continental weathering rate over the course of Cenozoic exist. For instance, the global mean Calcite Compensation Depth (CCD) sank by about 1 km over the last 40 million years (Figure 3), a change possibly linked to an increased supply of alkalinity from rivers caused by the Himalayan uplift. Paradoxically, there is no evidence of major changes in the carbonate accumulation flux during the Cenozoic (Figure 4). The deepening of the CCD might thus be linked, at least partially, to the global Cenozoic marine regression, reducing the area of shallow epicontinental seas and thus the area available for the accumulation of coral reefs. In that case, carbon and alkalinity will be preferentially removed from the ocean through enhanced formation of calcitic shells in open waters, leading to the deepening of the CCD. This process might have been favored by the coeval

CCD depth (km)

3.0

3.5

4.0

4.5

5.0

Late Cr

70

Paleoc

60

Oligocene Cenozoic

Eocene

50

40 30 Age (Ma)

Miocene

20

Pl

10

0

Figure 3 Reconstructed Carbonate Compensation Depth (CCD) through the Cenozoic (Van Andel, 1975; Broecker and Peng, 1982). Timescale as Figure 1.

_

Carbonate deposition flux (1012 mol y 1)

interest. Raymo’s world has the ability to exhaust atmospheric CO2 within a few million years. In an attempt to reconcile the two approaches, François and Walker proposed in 1992 the addition of a new CO2 consumption flux to the carbon cycle, identified as the precipitation of abiotic carbonates within the oceanic crust, subsequent to its alteration at low temperature. This flux is directly dependent on deep water temperature, which has decreased by B81C over the Cenozoic. An increase in the continental weathering rate might be compatible with a constant degassing rate, since the sink of carbon through low-temperature alteration of the oceanic crust is decreasing. The balance between input and output is thus still in place. However, this additional sink of carbon is poorly constrained. The present-day consumption of carbon is estimated to be about 1.4 1012 mol y1, but the kinetics of the process is essentially unknown. This attractive hypothesis still needs experimental verification. Finally, it should be noted that Raymo’s hypothesis interprets the increase in the sea water 87Sr/86Sr in terms of an increase in the weathering fluxes. However, silicate minerals exposed in the Himalayan area, particularly in the High Himalayan Crystalline Series, display unusually high isotopic ratios (reaching 0.740). Sediments of Proterozoic age with a 87 Sr/86Sr reaching 0.8 are also exposed in the Lesser Himalaya area. For this reason, rivers draining the Himalayan area (Ganges, Brahmapoutra, etc.) display an isotopic ratio (0.725 for the Ganges) higher than the mean global value (0.712). At least part of the Cenozoic increase in the sea water 87Sr/86Sr might thus be due to changes in the isotopic composition of source rocks.

517

30 25

20

15 10

Late Cr

70

Paleoc

60

Eocene

50

Oligocene Cenozoic

40 30 Age (Ma)

Miocene

20

10

Pl

0

Figure 4 Total carbonate accumulation flux reconstructed from paleodata (Opdyke and Wilkinson, 1988). Timescale as Figure 1.

development of new foraminiferal species. The cause of the CCD deepening thus remains unresolved. Organic Carbon Subcycle

The Cenozoic history of sea water d13C recorded in marine limestones (Figure 5) is marked by an ample fluctuation in the Paleocene and early Eocene, a roughly constant background value with superimposed high-frequency variations from the middle Eocene to the middle Miocene, and a sharp decrease from the middle Miocene to the present. Since organic matter is enriched in the lighter 12C isotope with respect to sea water (owing to photosynthetic fractionation), this d13C record can be used to constrain the temporal changes in the organic fluxes of the carbon cycle. The burial of organic matter on the sea floor preferentially removes 12C from the ocean and hence tends to increase seawater d13C. Conversely, the oxidation of old organic carbon


518


35

2 30 TOC(ppt)

13

Benthic forams C (ppt PDB)

3

1

25

0 _1

Late Cr

70

Paleoc

60

Eocene

50

Oligocene Cenozoic

40 30 Age (Ma)

Miocene

20

10

20

Pl

0

Figure 5 Sea water d13C recorded in ancient carbonate sediments (Ottawa-Bochum database: http://www.science. uottawa.ca/geology/isotope_data/). Timescale as Figure 1. ppt PDB ¼ parts per thousand PeeDee Belemnite.

(kerogen) contained in weathered sedimentary rocks is a source of isotopically light carbon for the ocean, tending to decrease its d13C. The recent decrease in d13C since mid-Miocene times might thus be interpreted as the result of kerogen carbon oxidation being larger than organic carbon burial during that period. Similarly, the overall constancy of sea water d13C from the middle Eocene to the middle Miocene may suggest that the organic subcycle was essentially balanced at that time. However, the average carbon isotopic fractionation (eTOC) between total organic carbon and sedimentary carbonate (which is close to coeval sea water) has decreased from the Eocene to the present (Figure 6). With a balanced organic subcycle, this change in TOC would imply a decrease of the sea water d13C over time, as it forces the d13C of organic deposits to become closer to the sea water value than it is for kerogen carbon. For the isotopic composition of the ocean to remain constant from the middle Eocene to the middle Miocene, the trend associated with eTOC variations must be compensated for by an imbalance in the organic subcycle in which the burial of organic carbon exceeds kerogen oxidation. This imbalance was progressively reduced after mid-Miocene times, but may have persisted until very recently. The late Cenozoic was therefore a time of unusually high organic carbon deposition rates, leading to an increase in the size of the sedimentary organic carbon reservoir. The organic subcycle thus acted as a carbon sink over the course of the Himalayan uplift. There are two possible causes of this evolution. 1. The increase in chemical weathering rates in the Himalayan region during the uplift (Raymo’s world) leads to enhanced delivery of nutrients to the ocean, forcing the oceanic primary

Late Cr

70

Paleoc

60

Eocene Oligocene Cenozoic

50

40 30 Age (Ma)

Miocene

20

10

Pl

0

Figure 6 Average carbon isotopic fractionation (eTOC) between total organic carbon and sedimentary carbonate (solid line, Hayes et al. (1999); dashed line, Freeman and Hayes (1992)). Timescale as Figure 1.

productivity to increase. This might result in an increased burial of organic matter. 2. Enhanced mechanical weathering in the Himalayan region increased the sedimentation rate on the ocean floor, so that organic carbon was more easily preserved. This hypothesis does not require any increase in the chemical weathering rate in the Himalayan region. This facilitated burial might have significantly contributed to the Cenozoic PCO2 decrease, since carbon is stored in a sedimentary reservoir. C. France-Lanord and L. Derry argued in 1997 that the consumption of CO2 through organic carbon burial might be three times more important today than the amount of CO2 consumed by silicate weathering within the orogen. Even if this hypothesis still links the climatic cooling with the Himalayan uplift, the origin of the CO2 sink is quite different from that hypothesized in Raymo’s world. Observational data argue toward the second hypothesis, indicating that the Cenozoic increase in the sea water 87Sr/86Sr might be of isotopic origin. Calcium silicates are indeed not the most common mineral exposed in the Himalayan orogen, and thus cannot contribute widely to the CO2 consumption. Furthermore, reverse weathering reactions take place in the Bengal Fan, releasing CO2 and thus reducing the impact of the Himalayan silicate weathering on PCO2. It has been suggested that the emission at some time in the past of large amounts of methane from gas hydrates may have influenced the d13C of the ocean. This may invalidate the interpretation of the carbon isotopic record if the gas hydrate reservoir has had long-term as well as shorter-term effects. Organic carbon deposition on the seafloor is linked to ocean biological productivity, itself


Phosphorus accumulation rate (mg cm

_2

per 1000 y)

CENOZOIC OCEANS – CARBON CYCLE MODELS N N X X dqi ¼ Fji Fij ði ¼ 1; y; NÞ dt j¼1jai j¼1jai

2.0

519

½2

1.5

1.0

0.5

0.0

Late Cr

70

Eocene

Paleoc

60

50

Oligocene Cenozoic

40

30

Miocene

20

10

Pl

0

Age (Ma)

Figure 7 Average accumulation of phosphorus in sediments through the Cenozoic (Fo¨llmi, 1995). Timescale as Figure 1.

depending on the availability of nutrients, among which phosphorus is thought to play a key role. The global phosphorus accumulation into sediments increased by a factor of about 4 over the last 10 million years (Figure 7), interpreted as the record of a global continental weathering enhanced by the onset of large ice sheets and glaciers at the end of Cenozoic. However, the question whether this increase is related to an increase in continental chemical weathering, or in mechanical weathering alone, is not clear.

Modelling: An Attempt to Integrate the Records into a Unified Framework The Concept of Box Models

Biogeochemical cycles are usually described with box models. Such models provide a simple mathematical framework appropriate for calculating the geochemical evolution of the Earth through geological times. The Earth system is split into a relatively small number of components or reservoirs assumed to be homogeneous, such as the atmosphere, the ocean, the biosphere, the continental or oceanic crust, and the (upper) mantle. These reservoirs are connected by a series of ‘arrows’ representing the flows of material between them. The biogeochemical cycle of each element is thus represented as a set of interconnected reservoirs and, at any time, its state is characterized by the reservoir sizes or contents qi (amount of the element in reservoir ‘I’, units: mol or kg) and the fluxes Fij (amount of the element transferred per unit time from reservoir ‘I’ to reservoir ‘j’, units: mol y1 or kg y1). The temporal evolution of the system can be calculated by making a budget of input and output fluxes for each reservoir (eqn [2]).

To solve this system of differential equations, the values of the fluxes must be provided at each time step. Kinetic rate laws describing the dependence of the fluxes Fij on the reservoir contents qi, time t, or some external forcing are thus needed. Defining such kinetic rate laws is the most critical task of modeling. The reliability of the solution and hence the usefulness of the results depend strongly on the adopted rate laws. The challenge is clearly to get at least a first-order estimate of the fluxes from a very broad knowledge of the system, i.e., from the values of its state variables q1,y, qN. A useful concept in box modeling is that of turnover time. The turnover time of an element in a given reservoir is defined as the ratio between its reservoir content and its total output flux (eqn [3]). t i ¼ PN

qi

j¼1jai

Fij

½3

The turnover time can be seen as the time needed to empty the reservoir if the input happened to stop suddenly and the current output flux were held constant. It provides a first-order idea of the evolution timescale of a reservoir. At steady state (i.e., when input and output fluxes balance each other), the turnover time is equal to the residence time, which is the average time spent by individual atoms of the element in the reservoir. Finally, the response time of a reservoir characterizes the time needed for the reservoir to adjust to a new equilibrium after a perturbation. Models of the Carbon Cycle

Figure 8 illustrates the present state of the long-term carbon cycle from a recent (unpublished) box model simulation of the authors. The reservoirs and fluxes that have been included in this figure are those that are important to describe the evolution of atmospheric CO2 at the geological timescale. The values of reservoir sizes and fluxes are consistent with current knowledge of the system. Crustal reservoirs include continental (5000 1018 mol C) and pelagic (150 1018 mol C) carbonates, as well organic carbon (1250 1018 mol C) from the sedimentary cover. The atmosphere and ocean have been lumped into one single reservoir containing 3.2 1018 mol C, since the time necessary for the atmosphere to reach equilibrium with the ocean is much shorter than B1 My, the timescale of geological processes. Indeed, with a modern atmosphere-ocean exchange flux of


520


Weathering_ Oxidation 4.0

Weathering 12.3

CONTINENTAL CARBONATES 5000

Metamorphism 2.0

Deposition 4.7 Deposition 13.4

ATMOSPHERE OCEAN 3.2

Metamorphism 0.5

ORGANIC CARBON 1250

Deposition 4.2

Metamorphism 1.5

PELAGIC CARBONATES 150

MOR Degassing 2.0

Reservoirs: 1018 mol C Fluxes: 1012 mol C y

_1

Return to the mantle 1.5

MANTLE

Figure 8 Present-day state of the long-term carbon cycle. Numbers represent 1018 mol C for reservoirs (boxes) and 1012 mol C y1 for fluxes (arrows).

7.5 1015 mol C y1 (i.e., 90 Gt C y1) and an atmospheric content of 62.5 1015 mol C (i.e., 750 Gt C), the turnover time of carbon in the atmosphere can be calculated to be only 8.33 years. Similarly, the terrestrial biosphere has not been included, since its size is small compared to other reservoirs and it can be assumed in equilibrium with the atmosphere– ocean system. The fluxes involved in the cycle are MOR/metamorphic CO2 release (‘volcanism’), weathering fluxes, and deposition of carbonates or organic carbon on the seafloor. The reported values of these fluxes are long-term averages, that is, they should be thought of as averages over several glacial– interglacial oscillations of the Pleistocene, although such averages cannot always be estimated from presently available data. The turnover time of carbon in the atmosphere–ocean reservoir in this ‘geological’ system can be calculated to be 143 000 y. Owing to this relatively short turnover time with respect to the timescale of long-term geological changes, the atmosphere–ocean system is essentially at equilibrium. By contrast, crustal reservoirs that exhibit much larger turnover times are not at equilibrium. This is clearly the case of continental (ti ¼ 350 My) and

pelagic (ti ¼ 50 My) carbonate reservoirs, as a result of the Cenozoic deepening of the ocean lysocline and the associated transfer of carbonate deposition from the shelf to the pelagic environment. To distribute the carbon content of the atmosphere–ocean system among its two components, and hence to derive the atmospheric PCO2 value (and its effect on the climate), it is necessary to know the alkalinity content of the ocean. For this reason, the evolution of the ocean alkalinity (AT) and ocean– atmosphere carbon (CT) content are always calculated in parallel. Writing eqn [2] for these two variables yields eqns [4a] and [4b]. dAT ¼ 2FSW þ 2FCW 2FCD dt

½4a

dCT ¼ FVOL þ FCW FCD þ FOW FOD dt

½4b

FVOL represents the total CO2 release flux from volcanic origin (i.e., the sum of all metamorphic and MOR fluxes in Figure 8), FCW and FSW are the weathering fluxes from respectively carbonate and



silicate rocks expressed in moles of divalent ions (Ca2þ or Mg2þ) per unit of time (i.e., the rates of reactions [I] and [II]), FCD is the carbonate deposition flux, FOW is the carbon input flux from weathering-oxidation of crustal organic carbon, and FOD is the organic carbon deposition flux. Note that the silicate weathering flux does not appear in the carbon budget, eqn [4b], since silicate weathering (reaction [II]) transfers carbon from the atmosphere to the ocean but does not remove it from the atmosphere–ocean system. The factor of 2 in eqn [4a] results from the fact that two equivalents of alkalinity are transferred to the ocean when one Ca2þ or Mg2þ ion is delivered to the ocean by rivers (reactions [I] and [II]). The same factor of 2 holds for carbonate deposition, which is the reverse of reaction [I]. As already mentioned, the atmosphere– ocean system must be close to equilibrium, so that the derivatives on the left-hand side of eqns [4a] and [4b] can be set to zero. This assumption transforms the differential equation system into a set of two algebraic equations, which can be solved to yield eqn [5]. FVOL FSW ¼ FOD FOW

½5

This equation leads to eqn [1] if the effect of the organic subcycle is neglected (i.e., when this subcycle is set to equilibrium). Hence, eqn [5] is a generalization of the Walker, Hays, and Kasting budget. It states that the disequilibrium of the inorganic part of the carbon cycle must be compensated for by a disequilibrium of opposite sign in the organic subcycle. Use of Isotopic Data (Inverse Modeling)

To solve eqns (4) or [5], some kinetic laws must be provided for the fluxes, that is, the relations between these fluxes, time t, and the reservoir contents, or atmospheric PCO2, must be known. Such kinetic laws are, however, poorly known, so it may be preferable, at least for some fluxes, to use forcing functions in the calculation of these fluxes. For example, volcanic fluxes are often made proportional to the seafloor spreading rate and weathering fluxes to land area, for which past reconstructions are available. Ocean isotopic records, such as those presented earlier, can also be used to force the model. Budget equations similar to [2] are then written for the relevant isotopes and transformed into equations containing isotopic ratios r (or d, the relative departure of the isotopic ratio from a standard). The sea water 87 Sr/86Sr ratio has been used in this way to estimate the silicate weathering flux FSW, but as discussed earlier the results are strongly dependent on the hypothesis of constancy for the isotopic ratios of

521

weathered products. The 13C isotopic history of the ocean has been used in many models, since the beginning of the 1980s, to constrain the organic carbon subcycle. The 13C isotopic budget for the ocean can be written as eqn [6]. CT

ddOC ¼ ðdVOL dOC ÞFVOL þ ðdCW dOC ÞFCW dt þ ðdOW dOC ÞFOW ðdOD dOC ÞFOD

½6

doc here is the d13C of the ocean (more precisely, this should be the d13C of the atmosphere–ocean system) ; dVOL, dCW, and dOW are the d13C of the carbon inputs from respectively volcanic, carbonate weathering, and crustal organic carbon weathering–oxidation fluxes. It is assumed that no fractionation occurs with respect to average oceanic carbon during carbonate precipitation, so that this flux does not appear in the equation. dOD ¼ doc D is the d13C of the organic carbon deposited on the seafloor, with y being the average fractionation of photosynthesis with respect to oceanic carbon (this includes both terrestrial and marine photosynthesis). The past values of doc are known from the 13C isotopic history of sea water (Figure 5). Equation [6] can then be solved with respect to FOD and the resulting expression for FOD is then used in eqn [4b] or [5]. The isotopic composition of the input fluxes must, however, be known or derived from similar isotopic budgets for the crustal reservoirs. This procedure is actually an inverse method, since it derives model parameters (fluxes) from an observed signal (ocean isotopic composition) linked to the model parameters through a mathematical operator (the isotopic budget equation). Y. Godde´ris and L.M. François in 1996, and L.R. Kump and M.A. Arthur in 1997, published two separate models inverting the oceanic d13C signal over the Cenozoic, making use of an isotopic fractionation D variable with age and derived from paleodata. The Cenozoic histories of silicate weathering from these models are compared in Figure 9. The predicted trend of the carbonate deposition flux is broadly consistent with an available reconstruction based on carbonate accumulation data (Figure 10). A classical example of a box model using 13C isotopic data to constrain the organic carbon subcycle is the BLAG model of Lasaga, Berner, and Garrels published in 1985. R.A. Berner in 1990 also used such an isotopic budget in GEOCARB to calculate the history of atmospheric CO2 over the Phanerozoic. The results show a decreasing trend of atmospheric CO2 over the Cenozoic. The trend is consistent with the overall trend reconstructed with other models (e.g., François and Walker, 1992) or


_

Continental silicate weathering (1012 mol y 1)

522


8

7

6

5

4

Late Cr

70

Paleoc

60

Eocene

50

Miocene

Oligocene Cenozoic

40 30 Age (Ma)

20

Pl

10

0

_

Carbonate deposition flux (1012 mol y 1)

Figure 9 Model silicate weathering flux (solid line, Godde´ris and François (1996); dashed line, Kump and Arthur (1997)). Timescale as Figure 1.

25

See also Calcium Carbonates. Carbon Cycle. Cenozoic Climate – Oxygen Isotope Evidence. Ocean Carbon System, Modeling of. Paleoceanography, Climate Models in. Past Climate from Corals. River Inputs. Sedimentary Record, Reconstruction of Productivity from the. Stable Carbon Isotope Variations in the Ocean.

20

15

10 5

the overall deepening of the lysocline from late Eocene time to the present, and after the mid-Miocene a marked decrease of ocean d13C together with an increase in total carbonate accumulation and possibly phosphorus deposition. Are these environmental changes related? The role of models is to synthesize and provide a coherent explanation of such records, and then reconstruct the history of other key variables not directly accessible from paleodata. Today, we are still far from this goal. It is fundamental that models use multiple proxy data both as forcings and for validation, implying that other biogeochemical cycles for which proxies are available are modeled together with the carbon cycle. A coupling to other major biogeochemical cycles is also essential because of the interactions with the carbon cycle and the feedbacks involved.

Late Cr

70

Oligocene Eocene Cenozoic

Paleo

60

50

40

30

Miocene

20

10

Pl

0

Age (Ma) Figure 10 Model carbonate accumulation flux compared to the reconstruction of Figure 4 normalized to the present day value of 20 1012 mol y1 (solid line: Godde´ris and François (1996); dashed line, Kump and Arthur (1997); shaded area, reconstruction from Figure 4.)

from various paleoindicators (Figure 1). This does not mean, however, that we understand the carbon cycle (and climate) trends of the Cenozoic, since different models can produce similar trends from completely different underlying mechanisms. To be reliable, models should not rest only on a limited set of data but should be able to explain a wide range of geochemical records.

Conclusions Proxy records indicate that the Earth’s climate cooled gradually over the Cenozoic. This cooling trend was accompanied by a decrease of atmospheric PCO2. Other striking features of the Cenozoic are the sharp increase of the 87Sr/86Sr ratio of sea water and

Further Reading Berner RA (1990) Atmospheric carbon dioxide levels over Phanerozoic time. Science 249: 1382--1386. Berner RA (1998) The carbon cycle and CO2 over Phanerozoic time: the role of land plants. Philosophical Transactions of the Royal Society of London B 353: 75--82. Broecker WS and Peng TH (1982) Tracers in the Sea. Palisades: Eldigio Press. Butcher SS, Charlson RJ, Orians GH and Wolfe GV (eds.) (1992) Global Biogeochemical Cycles. London: Academic Press. Chameides WL and Perdue EM (1997) Biogeochemical Cycles: A Computer-Interactive Study of Earth System Science and Global Change. Oxford: Oxford University Press. Fo¨llmi KB (1995) 160 My record of marine sedimentary phosphorus burial: coupling of climate and continental weathering under greenhouse and icehouse conditions. Geology 23: 859--862. France-Lanord C and Derry LA (1997) Organic carbon burial forcing of the carbon cycle from Himalayan erosion. Nature 390: 65--67. François LM and Walker JCG (1992) Modelling the Phanerozoic carbon cycle and climate: constraints from the 87Sr/86Sr isotopic ratio of sea water. American Journal of Science 292: 81--135.



Freeman KH and Hayes JM (1992) Fractionation of carbon isotopes by phytoplankton and estimates of ancient CO2 levels. Global Biogeochemical Cycles 6: 185--198. Godde´ris Y and François LM (1996) Balancing the Cenozoic carbon and alkalinity cycles: constraints from isotopic records. Geophysical Research Letters 23: 3743--3746. Harland WB, Armstrong RL, Cox AV, et al. (1990) A Geologic Time Scale 1989. Cambridge: Cambridge University Press. Hayes JM, Strauss H, and Kaufman AJ (1999) The abundance of 13C in marine organic matter and isotopic fractionation in the global biogeochemical cycle of carbon during the past 800 Ma. Chemical Geology 161: 103--125. Kump LR and Arthur MA (1997) Global chemical erosion during the Cenozoic: weatherability balances the budgets. In: Ruddiman WF (ed.) Tectonic Uplift and Climate Change. New York: Plenum Press. Kump LR, Kasting JF, and Crane RG (1999) The Earth System. New Jersey: Prentice Hall. Lasaga AC, Berner RA, and Garrels RM (1985) An improved geochemical model of atmospheric CO2 fluctuations over the past 100 million years. In: Sundquist E and Broecker WS (eds.) The Carbon

523

Cycle and Atmospheric CO2: Natural Variations Archean to Present Geophysical Monograph, vol. 32, pp. 397--411. Washington, DC: American Geophysical Union. Opdyke BN and Wilkinson BH (1988) Sea surface area control of shallow cratonic to deep marine carbonate accumulation. Paleoceanography 3: 685--703. Pearson PN and Palmer MR (2000) Atmospheric carbon dioxide concentrations over the past 60 million years. Nature 406: 695--699. Raymo ME (1991) Geochemical evidence supporting T.C. Chamberlin’s theory of glaciation. Geology 19: 344--347. Ruddiman WF (ed.) (1997) Tectonic Uplift and Climate Change. New York: Plenum Press. Van Andel TH (1975) Mesozoic-Cenozoic calcite compensation depth and the global distribution of calcareous sediments. Earth and Planetary Science Letters 26: 187--194. Van Andel TH (1994) New Views on an Old Planet: a History of Global Change, 2nd edn. Cambridge: Cambridge Universitys Press. Walker JCG, Hays PB, and Kasting JF (1981) A negative feedback mechanism for the long-term stabilization of Earth’s surface temperature. Journal of Geophysical Research 86: 9776--9782.


CEPHALOPODS P. Boyle, University of Aberdeen, Aberdeen, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 436–442, & 2001, Elsevier Ltd.

Introduction The Cephalopoda is the class of the Mollusca comprising the octopuses, cuttlefish, squid, and their allies. Exclusively marine and present in all of the world’s oceans and seas, their lineage can be traced from the Ordovician to the present due to fossilization of their large, heavy, chambered shells. The Pearly Nautilus (Nautilus spp.) of the Indo-Pacific region is the only surviving relative of this ancient ancestry (10–12 000 extinct species) Modern living cephalopods (subclass Coleoidea), having reduced or lost the ancestral shell, are represented by only about 650–700 species. These are characteristically large, active, soft-bodied predators, with complex behavioral and physiological capabilities. Occupying a wide range of benthic and pelagic habitats they are abundant in productive shelf regions, where genera such as Octopus and the common cuttlefish Sepia are each credited with over 100 species. The greatest diversity of form and biomass of cephalopods is oceanic and mesopelagic in distribution, but the biology of these offshore species is little understood and generalizations are based mostly on coastal forms. Now, and throughout their evolutionary history, representatives of the cephalopods reach the largest of all invertebrate body sizes. With the exception of Nautilus (and some of the deep-sea forms), cephalopods generally share common life cycle features. The large eggs hatch directly to free-swimming juvenile forms resembling the adult (paralarvae). Growth is very rapid (exponential and logarithmic phases) and adult size is reached in about a year (6–24 months). The sexes are separate and there are complex arrangements formating and fertilization. After spawning of the fertilized egg masses, either attached to the bottom or freely into the water column, most individuals of both sexes die within a short period of time afterwards. Although there are some variations between species in the timing of breeding and the duration of spawning, uniseasonal breeding appears to be more or less universal. The consequences of this life cycle at the population level are that there is little overlap of generations, the species biomass present at any

524

time tends to build and crash seasonally. Distribution and abundance of the shelf species at least is thus highly dependent on inter-annual conditions for recruitment and growth. Cephalopods have a very significant role in the trophic relations of marine ecosystems. Universally predatory, they consumea wide variety of fish, crustacea and other invertebrates. Cephalopods themselves are also preyed upon by many other large marine organisms such as fish, marine mammals of all sorts, and many oceanic birds. Conservative estimates of consumption of cephalopods by these predators considerably exceed 100 million tonnes annually. Human fisheries for cephalopods have increased steadily and are reaching about 3 million tonnes annually. Critical assessment of the role of cephalopods in the world’s oceans is compromised by the relative lack of information on the oceanic and deep-water forms and the consequent extrapolation of knowledge from the better-known coastal species.

Diagnosis of the Cephalopoda (Table 1) As a class of the Mollusca cephalopods share fundamental features of their body layout and development with the other classes (gastropods, bivalves, chitons, etc.) including absolutely characteristic molluskan features such as the radula (feeding organ). Other typically molluskan features such as the calcareous shell are reduced or absent (it remains as the ‘cuttlebone’ in Sepia and the gladius or ‘pen’ in squid) and there are no specialized larval forms (no molluscan trochophore or veliger). Cephalopods have also developed quite unique systems of locomotion and mobility (jet propulsion, suckers), brain development, color change (chromatophores), and light production (photophores). Although the molluskan relationships of cephalopods are without doubt, the scale and dynamics of their extant populations are better understood in terms of comparison with the teleost fishes – coevolution and competition for the most productive marine environments.

Biology Buoyancy and Jet Propulsion

The reduction and loss of a calcareous shell in the modern cephalopods (Coleoidea) has allowed the evolution of their highly mobile, active lifestyles,


CEPHALOPODS

525

Table 1 Diagnosis of the Cephalopoda: Classification is not entirely consistent between different authors and only the principal categories are given here, some common genera are listed and the common names of classification categories are shown in bold. (Abbreviated with permission from Boyle, 1983 (Cephalopod Life Cycles, Vol 1, 1–8, Academic Press, London) with additional information on the numbers of living species from Nesis, 1987.)

quite distinct from the other molluscan classes. Cuttle-fish (Sepia) and Spirula retain an internal remnant of the shell which functions as a buoyancy organ. The distribution of gas and fluid space within

the chambers is controlled osmotically and allows neutral buoyancy to be achieved. The physiological mechanism in these Sepiodea appears to be similar to that used by Nautilus for controlled vertical


526

CEPHALOPODS

movements through 1000 m of the water column and is thought to be the ancestral buoyancy mechanism common to the extinct nautiloids and ammonoids. The squids and octopuses have lost this mechanism entirely, most of them are negatively buoyant, but mesopelagic forms commonly reduce their density by chemical means such as retention of ammonium ions and loss of protein. Active locomotion of the pelagic species is by jet propulsion – regular spasmodic forcing of water from the muscular mantle through the ventral funnel which can be directed forwards, backwards or side-to-side to allow great maneuverability. Paired fins contribute to directional control of swimming and their undulations are used for ‘hovering’ and slow swimming. The common coastal octopuses (Incirrata) are mainly benthic, using the suckered arms for relaxed scrambling over the bottom, and jet propulsion only for rapid attacking or escape movements. Brain and Senses

The nervous system of cephalopods is still arranged in the basic molluscan layout as ganglionic masses grouped around the esophagus. It is centralized and developed to a much greater degree than that of other Mollusca. Coupled with large and complex sense organs, especially the eyes and statocysts (gravity and movement senses), the central nervous system supports an extensive and flexible repertoire of behavior unequalled by other invertebrate taxa. Especially in Octopus, the capability of the animal to discriminate between environmental cues and to make appropriate behavioral responses has been intensively studied and the detailed neuroanatomy has shown how the motor, sensory, and integrative functions of the brain are spatially located in its many subdivisions. Learning the significance of environmental cues and adapting its behavior accordingly is highly developed in Octopus. Color and Pattern

The most remarkable, and immediately visible manifestation of the behavior and responses of cephalopods, is their ability to control and change the colors, pattern, and texture of body surface. The skin of cephalopods is a delicate epithelial surface beneath which are layers of connective tissue, active colored cells (chromatophores), passive reflecting bodies (iridophores, leucophores), and a complex system of muscle fibers for moving the skin over the underlying somatic muscle surface. This capability for altering the appearance of the animal is present throughout the Cephalopoda but is expressed to the greatest degree among the coastal octopuses, cuttlefish, and loliginid squid.

The unique functional components of color change in the cephalopod skin are the chromatophores. Each one consists of a single cell within which is an elastic sac of pigment (yellow-redbrown-black). Inserted onto the pigment sac is a series of muscle fibers (25–30) radiating out into the surrounding connective tissue. In the relaxed state, the pigment sac is passively retracted to a microscopic dark point, the muscle fibers are extended, and the skin surface appears white due to reflection from the underlying somatic muscle. In the active state, the chromatophore muscle fibers contract, extending the pigment sac and spreading the area covered by its contained pigment. The pigment now screens the underlying muscle, the incident light is selectively absorbed by the pigment and the reflected wavelengths give color to the surface. The chromatophore muscles responsible for these pigment movements are innervated by fine nerve fibers ramifying throughout the skin. Contraction of the individual muscles may take only 200–300 ms, and the animal may change its complete appearance in a few seconds. In addition to these active chromatophores there may be several classes of passively ‘reflecting cells’ responsible for colors in the bluegreen range (iridophores) or white by scattering of all wavelengths (leucophores). The arrangement of these layers, overlaying each other throughout the depth of the skin, allows almost infinite combinations of effects. Since the chromatophores are innervated directly from the brain, their activity can be controlled to express a great variety of pattern and contrast. The use of color, contrast, and textural change in the intra- and inter-specific behavior patterns of cephalopods has been described for many species. These capabilities are mostly involved in crypsis (camouflage), mating activities, prey and predator responses, and are generally assumed to be of great survival significance and selective value. Surprisingly, there is no evidence that cephalopods themselves can discriminate colors, they respond mostly to contrast and pattern information. The scientific literature on cephalopod behavior is dominated by their visual capabilities. It is certain that they have also developed senses for tactile, vibration, and chemical stimuli, but little is known about the significance of these senses to behavior. Escape and Luminescence

Squid have evolved a rapid ‘escape response’ behavior which is mediated by three sets of nerve cells with exceptionally large fibers o1 mm in diameter (‘giant fibers’) and specialized connections (synapses). This ‘giant fiber system’ distributes the motor commands from the brain simultaneously to all parts of the


CEPHALOPODS

mantle musculature, and synchronously to each side, ensuring the maximum power of mantle contraction and speed of escape. In common with fish and other invertebrate life of the deep sea, most of the mesopelagic squid show various forms of luminescent display. Most commonly present as a pattern of light-emitting organs distributed on the surface, symbiotic luminescent bacteria are also present in some of the internal organs or may be released into water as a luminescent cloud (see Bioluminescence). The functions of these systems are not fully understood, but presumably they are involved in counter-shading of the animal against surface illumination, sexual signalling, or predator–prey encounters.

The Life Cycle Feeding and Growth

527

Among the better-studied coastal species, sexual maturation occurs rapidly, often at wide range of body sizes, and is usually associated with slowing or cessation of growth. In females, maturation is primarily a process of egg growth due to the accumulation of large amounts of lipoprotein yolk. Males mature spermatozoa, package them into complex spermatophores, and store them in the spermatophoric (Needham’s) sac. Individual matings occur in which the male transfers the spermatophores directly to a receptive female. In many species there may be complex reproductive behavior, allowing the possibility of mate selection by either sex, In addition, females may mate with several males and usually have the capacity for storage of the transferred sperm and the possibility of multiple paternity of offspring. Spawning and Death

The life cycles of coastal cephalopods share many features. All are predators, feeding on a wide range of species especially crustacea and fish, many are also cannibalistic on smaller members of their own species. They ingest food at high rates, ranging between 1.5 and 15% body weight per day in different species and at a range of temperatures. Individual growth rates have been estimated to range from 1% to over 10% body weight per day, with estimates for gross growth efficiency (growth increment as a percentage of food ingested) between 10 and 70% per day for animals in captivity. Feeding and growth rates decline at large body sizes and at lower temperatures, and gross growth efficiency is generally lower in active squid species than the more sedentary octopuses and cuttlefish. Feeding generally entails visual orientation and forward strike at the prey, gripping and pulling it in towards the mouth with the tentacles and arms. Squid and cuttlefish bite immediately into the tissues with the powerful chitinous mandibles (beaks), ingesting the most accessible parts and often releasing the dead remains partially eaten. Octopuses, in contrast, have evolved elaborate methods of prey handling – particularly effective on crustacea – involving external toxins and enzymes, before cleanly extricating the flesh from the carapace. After capture, a minute penetration of the carapace is made (o1 mm long) and a cocktail of compounds, including protease and chitinase enzymes together with paralyzing toxins, is injected. As well as subduing the prey, the enzymes have the effect of releasing the attachments of the crustacean tissues, allowing them to be selectively eaten. Reproduction

All cephalopods are diecious, the sexes are separate and no hermaphroditism or sex change is described.

Coastal squid (family Loliginidae), octopuses (family Octopodidae), and all of the cuttlefish (order Sepiidae) encapsulate their eggs, often in tough secreted coatings, and attach them to the bottom or other hard surfaces in clusters or strings. Some octopuses (e.g., Octopus vulgaris), subsequently stay with the egg mass, protecting it from epigrowths and defending it against predators. Most of the oceanic squid families (suborder Oegopsida) apparently spawn their eggs in fragile mid-water masses, but very little is known about the details of their mating and spawning habits. Compared with other molluscs, cephalopod eggs are large (1–25 mm long) and yolky. Fecundity is estimated to be as low as 10–25 eggs/female for some sepiolids and octopus; 10 000–100 000 for most coastal squid and octopus; and 100 000 to over 1 million for oceanic squid (e.g., family Ommastrephidae). The eggs hatch directly, without any specialized larval forms, to an active swimming miniature of the adult form. Because the hatchling may have different habits and occupy a different ecological zone from the adult, they are usually referred to as ‘paralarvae’. Reproduction in many cephalopods occurs with a seasonal peak which is often rather inconsistent in duration and timing. In most populations some breeding individuals may be found throughout the year. Evidence from captive individuals and field populations consistently shows that modern cephalopods (with the exception of Nautilus) have only one breeding season. Taking account of some variations, such as ‘batch spawning’ (the release of the reproductive output in several episodes over a short period of time), and the apparently ‘continuous’ release of single eggs by deep-water octopods (suborder Cirrata); there is no indication that, after spawning, there is regeneration of the ‘spent’ gonads for a subsequent breeding season.


528

CEPHALOPODS

Instead, there is every indication that spawning in both sexes marks the end of their lives and that death soon follows (called semelparous reproduction).

Ecology Population Biology

The cephalopod life cycle paradigm of fast growth, single breeding, and short life has profound consequences for the population biology of any species.

In its most extreme expression, if breeding is more or less synchronous and strongly seasonal throughout the population, it means that there will be only one adult size mode; biomass will strongly build and crash; and there will be little overlap of generations to buffer the influence on recruitment of environmental variables. In fact, tendencies to asynchronous breeding and lack of seasonality coupled with plasticity of feeding and growth characters, operate to spread the risks to the population of this life cycle and are shown schematically in Figure 1.

Direct _ targeted survey, stratified sampling, areal expansion

Methods of estimation

Indirect _ total fishing catch, catch per unit effort, consumption by predators, egg/larval surveys, environmental correlates

spawning aggregation

pre-spawning mortality

post-spawning mortality

maturation adult growth sub-adult growth juvenile growth hatching success spawning

A

Strongly seasonal synchronous breeding

Factors affecting biomass production

Factors affecting biomass location

B

C

Extended seasonal multiple cohorts

Almost aseasonal continuous recruitment

Numbers _ conditions for spawning and hatching, predation at every phase, post-spawning mortality

Size _ growth rate affected by temperature, food supplies at each phase, reproductive maturation

Migrations _ largescale, meso-scale, spawning aggregations

Hydrography _ seasonal and interannual current shifts

Figure 1 Diagrammatic representation of the periodic fluctuations of biomass of an annual semelparous species. Variations in the pattern of breeding and recruitment are suggested in (A), (B) and (C). Methods of estimation and the factors affecting biomass production are summarized. (Reproduced with permission from Boyle PR and Boletzky SV (1996) Cephalopod populations: definition and dynamics. Philosophical Transactions of the Royal Society of London: Series B 351: 985–1002.)


CEPHALOPODS

Trophic Relations

Numerous studies have shown that cephalopods are of great significance in the diets of many marine top predators. This evidence arises from the presence of the indigestible mandibles or ‘beaks’ in the gut contents of predators such as large fish, many seals, whales, and oceanic birds. As many as 30 000 have been found in the stomach of a sperm whale. The shape, size, and features of the upper beaks can be used to identify the cephalopod family or species, and estimate the size of the individual from which it came. Despite inherent errors in the procedure due to the unknown residence time of beaks in the gut, these beaks can be used to estimate the species composition and relative biomass of different cephalopods in the predator diet. No cumulative estimate of consumption by these large vertebrate predators is available, but taking into account the estimated predator population size and consumption rates it has been variously estimated that sperm whales alone could consume 213– 320 106 t of cephalopods from open ocean areas. Fisheries

Human fisheries for cephalopods have been recorded at least since classical times. Coast dwellers throughout the world still catch octopuses with traditional traps of pots or baskets, while cuttlefish and squid are taken on simple hand lures. The quantities taken by these hand capture fisheries are usually unrecorded. Cephalopods are also valuable fishery products traded on a global market. Using large bottom trawls specially tuned for cephalopods, oceanic drift nets (now banned in many areas), and highly efficient mechanized jigging vessels, the annual commercial harvest has risen steadily from about 1 million tonnes in 1970 to around 3 million tonnes by the mid-1990s. This was achieved largely by the extension of fishing to previously unexploited areas (North Pacific, South Atlantic) and also from increased catches in areas where the teleost fishes have been heavily exploited (Saharan Bank). Role in the Oceans

Combining data from commercial harvest fisheries with crude estimates of total consumption by predators, the global biomass (standing stock of adults and sub-adults) of cephalopods in the oceans has been variously estimated to lie between 193 and 375 106 t. These figures have been derived by tentatively accumulating the estimate ranges for mesopelagic squid (150–300 106 t); oceanic epipelagic squid (30– 50 106 t); slope/shelf-edge squid (8–15 106 t); and shelf sepioids and octopuses (5–10 106 t).

529

Whether or not these apparently massive estimates of biomass are realistic, their compatibility with other global estimates of marine productivity and consistency with the productive potential of cephalopods is uncertain. Overestimation of cephalopod frequency in predator diets could arise because cephalopod remains persist longer than those of other prey. Another source of uncertainty is the scaling up of limited predator and fishery data to ocean basin scales and the low carbon content (watery tissues) of many mesopelagic cephalopods. Some studies suggest the cephalopod biomass in the open sea (nektonic) to be about half that of fish and the mismatch between direct sampling with nets and indirect sampling from higher predators in this environment is well known.

Conclusions Cephalopods are undoubtedly one of the most charismatic groups of marine animals. Sharing a basic body with the other molluscs they have evolved very distinctive biological characters, advanced behavior, and life cycle patterns. Cephalopods comprise a major sector of marine biomass, having central significance to higher tropic levels and global fisheries. Little is understood about the biology of the oceanic and mesopelagic species and, consequently major uncertainties remain surrounding the quantitative role of cephalopods in the world’s oceans.

See also Bioluminescence. Molluskan Fisheries.

Further Reading Boucaud-Camou E (ed.) (1991) La Seiche. Caen: Universite´ de Caen. Boyle PR (ed.) (1983) Cephalopod Life Cycles, vol. 1. Species Accounts. London: Academic Press. Boyle PR (1986) Neural control of cephalopod behaviour. In: Willows AOD (ed.) The Mollusca, vol. 9(2), pp. 1–99. New York: Academic Press. Boyle PR (ed.) (1987) Cephalopod Life Cycles, vol. 2. Comparative Reviews. London: Academic Press. Clarke MR (ed.) (1996) The role of cephalopods in the world’s oceans. Philosophical Transactions of the Royal Society of London, Series B 351: 977–1112. Hanlon R and Messenger J (1996) Cephalopod Behaviour. Cambridge: Cambridge University Press. Nesis KN (1987) Cephalopods of the World: Squids, Cuttlefishes, Octopuses and Allies. TFH Publications (English translation of the original Russian editon, 1982 VAAP Copyright Agency of the USSR for Light and Food Industry Publishing House, Moscow).


530

CEPHALOPODS

Nixon M and Messenger JB (eds.) The Biology of Cephalopods. Symposium of the Zoological Society, London, no. 38. London: Academic Press. Rodhouse PG, Dawe EG, and O’Dor RK (eds.) (1998) Squid Recruitment Dynamics. The Genus Illex as a Model, the Commercial Illex species and Influences on Variability. FAO Fisheries Technical Paper, no. 376. Roper CFE, Sweeney MJ, and Nauen CE (1984) FAO Species Catalogue, vol. 3. Cephalopods of the World.

An Annotated and Illustrated Catalogue of Species of Interest to Fisheries. FAO Fisheries Synopsis. Saunders WB and Landman NH (eds.) (1987) NAUTILUS: The Biology and Paleobiology of a Living Fossil. New York: Plenum Publishing. Wells MJ (1978) Octopus: Physiology and Behaviour of an Advanced Invertebrate. London: Chapman and Hall.


CFCS IN THE OCEAN R. A. Fine, University of Miami, Miami, FL, USA

Atmospheric Source


The chlorofluorocarbons, CFCs, are synthetic halogenated methanes. Their chemical structures are as follows: CFC-11 is CCl3F, CFC-12 is CCl2F2, and CFC-113 is CCl2FCClF2. For completeness the compound carbon tetrachloride, CCl4, is also included in this article as its atmospheric source, measurement, and oceanic distribution are similar to those of the CFCs. The CFCs have received considerable attention because they are a double-edged environmental sword. They are a threat to the ozone layer, and a greenhouse gas. The CFCs are used as coolants in refrigerators and air conditioners, as propellants in aerosol spray cans, and as foaming agents. These chemicals were developed over 50 years ago when no one realized that they might cause environmental problems. When released CFCs are gases that have two sinks, the predominant one being the atmosphere, and to a lesser extent the oceans. Most of the CFCs go up into the troposphere, where they remain for decades. In the oceans and in the troposphere the CFCs pose no problem. However, some escape into the stratosphere where they are a threat to the ozone layer. Due to their role in UV absorption they have been correlated with the increased incidence of skin cancers. Since the recognition of the CFCs as an environmental problem in the 1970s and the signing of the Montreal Protocol in 1987, the use of CFCs has been phased out. The atmospheric concentrations have just started to decrease. This is an important international step toward correcting the dangerous trend of stratospheric ozone depletion. The atmospheric CFC concentrations became significant after the 1940s. The concentrations increased exponentially until the mid-1970s, and then increased linearly until the 1990s at a rate of about 5% per year. The production and release data for CFCs tabulated by the Chemical Manufacturers Association (CMA) were used (Figure 1) to reconstruct the atmospheric time histories for the Northern and Southern Hemispheres. Since 1979 the atmospheric concentrations have been based on actual measurements at various sampling stations around the globe, and these are checked against the CMA production and release estimates. The curves in Figure 1 show all CFCs including CCl4 increasing with time, with CFC-11 leveling off and actually decreasing in the late 1990s. The atmospheric increase of all the CFCs slowed markedly after the Montreal Protocol. The


Introduction The oceans, atmosphere, continents, and cryosphere are part of the tightly connected climate system. The ocean’s role in the climate system involves the transport, sequestration, and exchange of heat, fresh water, and carbon dioxide (CO2) between the other components of the climate system. When waters descend below the ocean surface they carry with them atmospheric constituents. Some of these are gases such as carbon dioxide and chlorofluorocarbons (CFCs). The CFCs can serve as a physical analog for CO2 because they are biologically and chemically inert in oceans. In the oceans the distribution of CFCs provides information on which waters have been in contact with the atmosphere in the past few decades. The CFCs also give information on the ocean’s circulation and its variability on timescales of months to decades. The timescale information is needed to understand and to assess the ocean’s role in climate change, and its capacity to take up anthropogenic constituents from the atmosphere. Thus, the advantage of using tracers like CFCs for ocean circulation studies is the added dimension of time; their time history is fairly well known, they are an integrating quantity and an analog for oceanic anthropogenic CO2 uptake, and they provide an independent test for time integration of models. Tracers serve as a ‘dye’ with which to follow the circulation of ocean waters. There are conventional ocean tracers such as temperature, salinity, oxygen, and nutrients. There are stable isotope tracers such as oxygen-18, carbon-13, and there are radioactive tracers both naturally occurring (such as the uranium/thorium series, and radium), and those produced both naturally and by the bomb tests (such as tritium and carbon-14). The bomb contributions from the latter two are called transient tracers, as are the CFCs, because they have been in the atmosphere for a short time. This implies an anthropogenic source and a nonsteady input function.


531

532

CFCS IN THE OCEAN

CFC-12 500

(ppt)

400

300 CFC-11 200

100

0 1900

CCI4 CFC-113

1920

1940

1960

1980

2000

Year Figure 1 Northern Hemisphere atmospheric time histories. (Atmospheric data from Walker et al., 2000.)

uncertainties in the reconstructed pre-1979 atmospheric histories depend on the atmospheric lifetimes of the compounds. The range of lifetimes for the compounds are 29–76 years for CFC-11, 77–185 years for CFC-12, 54–143 years for CFC-113, and 21–43 years for CC14. The uncertainties are a few percent, and they are highest for the early period. The continuous direct atmospheric measurements, which began in the late 1970s, are uncertain to within 1–2%. It is important to evaluate the uncertainties in the atmospheric source function, because they translate into uncertainties when used to put timescales on ocean processes. Because of their long atmospheric residence times, the CFCs are homogeneously distributed in the Northern and Southern Hemisphere, with the Northern Hemisphere about 8% higher than the Southern Hemisphere.

CFCs in the Oceans Analytical Techniques

Water samples collected from the ocean are measured for CFCs and CCl4 using an electron capture detection gas chromatography system. Analysis of water samples is done onboard ship, usually within hours of collection. The unit of measure is pmol kg 1 or 10 12 moles kg 1. These are extremely low level concentrations that are easily susceptible to

contamination from shipboard refrigerants, solvents, lubricants, etc. Chemical Stability

Under oxygenated oceanic conditions both CFC-11 and CFC-12 are believed to be chemically stable. CFC-11 has been shown to be unstable in anoxic marine waters; both CFC-11 and CFC-12 have been shown to be unstable in anoxic sediments. The compound CCl4 undergoes temperature-dependent hydrolysis, which limits its usefulness in the ocean when sea surface temperatures exceed B181C. CFC113 also has some stability problems at higher temperatures. Gas Flux and Solubility

The CFCs are gases, and like other gases they get into the ocean via air–sea exchange. There is a direct correlation between gas exchange rate and wind speed, and the direction of the gas flux between the air and ocean is from high to low concentration. For CFCs the atmospheric concentrations generally exceed those in the ocean. The concentration of CFCs dissolved in the surface layer of the oceans is dependent upon the solubility, atmospheric concentration, and other physical factors affecting the gas saturation including upwelling, entrainment due to mixing, ice cover, etc. The solubility of CFCs and CCl4 has been measured in the laboratory. The


CFCS IN THE OCEAN

accuracy of the measurements is about 1.5% and precision is about 0.7%. The solubility increases with decreasing temperature, at a rate of about 4% for 11C. Therefore, the colder the water the higher the CFC concentration. At a constant salinity the temperature effect is about two times greater for CFCs than for oxygen. The solubility is only slightly dependent on the salinity, and it decreases with increasing salinity. Surface Saturation

The approach to equilibrium condition or the saturation state is dependent on the mixed layer depth and air–sea transfer rate. It takes from days up to a few weeks after a change in temperature or salinity for ‘normal’ (not very deep) oceanic surface layers to come to equilibrium with the present atmosphere. While the surface waters of the world’s oceans are close to equilibrium with the present day atmospheric concentration of CFCs, there are exceptions. At times of rapid warming, such as in the spring, the surface waters will tend to be a few percent supersaturated with the gas due to lack of time to equilibrate with the atmosphere. Likewise at times of rapid cooling the surface waters will be a few percent undersaturated with the gas. Typically there are undersaturations within a few degrees of the equator due to upwelling of deeper less saturated waters. In high latitudes, where there are deep convective mixed layers that do not readily equilibrate with the atmosphere, there are likely to be undersaturations of as much as 60%. These have been observed in the Labrador Sea. The undersaturations in the high latitude water mass source regions need to be taken into account when using the CFCs to put timescales on oceanic processes. Oceanic Distribution

The compounds CFC-11 and CFC-12 were first measured in the oceans in the late 1970s. The first systematic and intensive survey was carried out in the tropical North and South Atlantic oceans starting in the early 1980s. Since then CFCs have been part of the measurements made during physical oceanography field work. A global survey was conducted as part of the World Ocean Circulation Experiment during the 1990s. Typical vertical profiles versus pressure for stations in the North Atlantic and North Pacific oceans are presented in Figure 2 along with other properties. Although CFC-12 has higher concentrations in the atmosphere, CFC-11 is more soluble in sea water, so its concentrations are about twice that of CFC-12. Note that there are measurable concentrations of CFCs in the western North

533

Atlantic that reach to the ocean bottom, while they reach to only 1000 m in the North Pacific. The difference between the CFC concentrations of the North Atlantic as compared with the North Pacific, reflects the formation of deep waters in the North Atlantic and the absence in the North Pacific. Concentrations generally decrease as the ocean depth increases. However, there may be subsurface concentration maxima due to the lateral intrusion of water that has been in more recent contact with the atmosphere (see applications below). The concentrations of CFCs and oxygen should behave similarly except where the biological effects on the oxygen distribution cause the differences, for example, the oxygen minimum at mid-depth. Combining a series of vertical profiles, as in Figure 2, will give a slice or section through the ocean. Sections through the eastern Pacific and Atlantic are shown in Figure 3. The absence of CFCs in the deep waters of the Pacific Ocean shows the relative isolation of the deep Pacific from contact with the atmosphere on timescales of decades. In contrast, the North Atlantic north of 351N has CFCs in deep and bottom waters, because these waters form in the high latitudes of the North Atlantic and easily spread equatorward on timescales of 10–20 years. As part of the density-driven, thermohaline circulation some of these waters will eventually be transported into the Pacific, but it will take hundreds of years. The upper waters of both oceans are in contact with the atmosphere on much shorter timescales. These upper waters are part of the wind-driven circulation.

CFC Ages in the Ocean Age Calculations

One of the main advantages of using CFCs as tracers of ocean circulation is that the time-dependent source function permits the calculation of timescales for these processes. A tracer age is the elapsed time since a water parcel was last exposed to the atmosphere. The tracer-derived age is the elapsed time since a subsurface water mass was last in contact with the atmosphere. Two estimates of ‘age’ can be calculated, one from the CFC-11/CFC-12 ratio and one from the partial pressure of either dissolved CFC. In both cases, the atmospheric value of either the ratio or partial pressure with which the water had equilibrated is compared to the atmospheric source function to determine the corresponding date. To normalize the concentrations for the effects on the solubility of temperature and salinity CFCs are expressed in terms of their partial pressures, pCFC, where the pCFC is the concentration divided by the


534

CFCS IN THE OCEAN

P17C, stn20

P17C, stn20 34

35

36

37

0

3

4

1000

1000

1000

2000 3000

10

0 (A)

20

2000 3000

5000

5000 0

(B)

35

1 2 3 _1 CFC-11 (pmol kg )

0

4 (C)

STACS4, stn7

STACS4, stn7

CFC-12 (pmol kg ) 36

37

0

1

2

3

4 0

1000

1000

1000

2000 3000 4000

Pressure (dB)

0

Pressure (dB)

0

2000 3000

5000 20

0

30


(E)

3000

5000

5000 10

2000

4000

4000

0

50 100 150 200 250 300 _1 Oxygen (μmol kg )

_1

Salinity 34

3000 4000

STACS4, stn7 33

2000

4000

30


Pressure (dB)

0

5000

Pressure (dB)

2

0

4000

(D)

1

0

Pressure (dB)

Pressure (dB)

33

P17C, stn20 _1

CFC-12 (pmol kg )

Salinity

1

2

3

4

_1

CFC-11 (pmol kg )

0 (F)

50 100 150 200 250 300 _1 Oxygen (μmol kg )

Figure 2 Vertical profiles of oceanographic data. (A) North Pacific salinity and potential temperature, (B) North Pacific CFC-11 and CFC-12, (C) North Pacific oxygen, (D) North Atlantic salinity and potential temperature, (E) North Atlantic CFC-11 and CFC-12, (F) North Atlantic oxygen. North Pacific World Ocean Circulation Experiment cruise P17C station 20, 331N, 1351W, June 1991; North Atlantic Subtropical Atlantic Climate Studies cruise station 7, 26.51N, 761W, June 1990. (North Atlantic data from Johns et al. (1997) Journal of Physical Oceanography 27: 2187–2208; Pacific data from Fine et al. (2001) Journal of Geophysical Research.)

solubility of the gas. This value is then adjusted for what the percent surface saturation is thought to be based on the measured temperature and salinity, then matched to the atmospheric time histories, and a corresponding year is assigned to the water mass. This age is an average of the water parcel. The pCFC is used to calculate the age of upper ocean waters, because at low concentrations the effects of dilution will bias the age toward the older components of a mixture. The age can also be calculated using the ratio of two CFCs; instead of using one pCFC the ratio of two pCFCs are used. In this case no assumptions are needed about surface equilibrium saturation at the

time of water mass formation. Since the atmospheric changes in the ratio of CFC-11/CFC-12 have remained unchanged since the mid-1970s, this restricts the application of the ratio age for CFC-11 and CFC12 to waters dating back further than 1975. However, either CFC-11 or CFC-12 can be combined with CFC-113 to extend age estimates to the present. Similarly they can be combined with CCl4 to extend age estimates further into the past. Unlike the pCFC age, the ratio ages are actually the ages of the CFCbearing components. Figure 4 shows sections of CFC ratio ages from the eastern North Atlantic and North Pacific oceans. Note that the intermediate and deep waters of the eastern North Atlantic (between 2000


CFCS IN THE OCEAN

535

Pressure (dB)

0

2000

4000

6000 _5

(A)

5

15

25

35

55

45

65

Pressure (dB)

0

2000

4000

6000 (B)

_ 30 _ 25 _ 20 _ 15 _ 10 _ 5

0

5

10 15 Latitude

20

25

30

35

40

45

50

Figure 3 (A) Sections versus pressure of CFC-11 concentrations (pmol kg 1) in the eastern Atlantic (latitude 651N–51S) along 201W in summer 1988. (B) Sections of CFC-11 concentrations (pmol kg 1) in the eastern Pacific (latitude 541N–321S) mostly along 1351W in summer 1991. (North Atlantic data from Doney SC and Bullister JB (1992) Deep-Sea Research 39: 1857–1883; Pacific data from Fine et al. (2001) Journal of Geophysical Research.)

and 4000 m) have CFCs younger than 30 years north of 451N, because of their proximity to the formation regions, whereas this is not the case in the North Pacific. In the Pacific below 2000 m the water column has been isolated from interaction with the atmosphere on similar timescales (except for the far western South Pacific). Caveats For Using CFC Ages

There are several caveats to the use of CFC ages. Both ages – partial pressure and ratio – may be subject to biases when there is mixing of more than one water mass component. Because of nonlinearities in the source functions and solubilities, neither age mixes linearly in multicomponent systems over the entire concentration range observed in the ocean. The atmospheric source function is nonlinear for much of

the input history; however, it can be approximated as being linear between the late 1960s and 1990. The solubilities are nonlinear functions of temperature, but they are approximately linear over ranges of a few degrees. Thus, for some regions of the ocean, these nonlinearities are not significant. The different types of ages are appropriate for putting timescales on different processes. For thermocline ventilation, where equilibrated water is subducted and mixed isopycnally along extensively outcropping density surfaces, the water subducted within a given year mixes with water subducted in previous years. In this situation, a water parcel is a mixture of water that has left the surface over a period of several years. The average age of this water parcel can be represented by the pCFC age if the change of CFC concentration in the source region is constant with respect to time. This has been


536

CFCS IN THE OCEAN

0

Pressure (dB)

12 2000

20 30

4000

6000 _5

(A)

5

25

15

35

0

65

20

30

Pressure (dB)

55

45

30

2000 > 30 years

4000

6000

_ 30 _ 25 _ 20 _ 15 _ 10

_5

0

5

(B)

10 15 Latitude

20

25

30

35

40

45

50

Figure 4 (A) Sections of CFC-11/CFC-12 ratio ages (years) in the eastern Atlantic (latitude 651N–51S) along 201W in summer 1988. (B) Sections of CFC-11/CFC-12 ratio ages (years) in the eastern Pacific (latitude 541N–321S) mostly along 1351W in summer 1991. (North Atlantic data from Doney SC and Bullister JB (1992) Deep-Sea Research 39: 1857–1883; Pacific data from Fine et al. (2001) Journal of Geophysical Research.)

confirmed for the North Atlantic thermocline in the eastern basin by comparing pCFC ages to tritium/ He-3 ages. In regions where surface waters are converted to deep and bottom waters which then spread into a background of low-tracer water, the high CFC concentrations of the cold surface water are diluted by entrainment and mixing. The resulting pCFC age is much too young for the average age of the mixture and much too old for the CFC-bearing component. However, a tracer ratio is conserved in this situation, and the corresponding ratio age represents that of the youngest component of the mixture, not the average age of the water parcel. Thus, there are different estimates of ages that can be derived from CFC-11 and CFC-12, and the associated timescales can be expanded in regions where CFC-113 and CCl4 data are available.

In most high latitude intermediate and deep-water source regions the age clock is not reset to zero due to lack of time to equilibrate deep mixed layers with the atmosphere. Thus, water masses will start out with an age of a few years (rather than zero), that is, they are not completely renewed during formation. This additional age is called a relic age which can be estimated from observations of the tracers at the water mass formation regions. The relic age can then be subtracted from the tracer ages calculated downstream from the water mass formation regions.

Applications of CFCs to Ocean Processes Examples of the application of CFCs to understanding oceanographic processes are divided into


CFCS IN THE OCEAN

537

four subjects: the thermohaline circulation, upper ocean circulation, model constraints, and biogeochemical processes.

advances that have come about in our understanding of the upper ocean circulation to which observations of CFCs have contributed:

Thermohaline Circulation

•

There is a close coupling of the surface waters in high latitudes to the deep ocean through the density-driven thermohaline circulation. During the process of deep-water formation, atmospheric constituents such as CFCs are introduced into the newly formed water. In recent years, major advances in our knowledge of the thermohaline circulation can be attributed to information derived from transient tracer data, particularly for two reasons. First, the development of analytical techniques so that oceanographers can easily produce large quantities of high quality data. Tracer oceanographers have benefited from multiinvestigator programs like the World Ocean Circulation Experiment. The following highlights some of the advances that have come about in our understanding of the thermohaline circulation to which observations of CFCs have contributed:

•

• • •

• •

Discovery of a new water mass component of North Atlantic Deep Water (NADW), called Upper Labrador Sea Water, location of its formation region and contributing processes, and timescales of eastward spreading along the equator. Identification of Denmark Straits Overflow Water as the primary source of bottom water of the western subpolar basin. Confirmation of the structure and continuity of the Deep Western Boundary Current throughout the western North Atlantic Ocean, and extension into the South Atlantic. Extension of the CFCs well into the interior of the western North Atlantic show the importance of deep recirculation gyres in ventilating the interior basins, and in slowing the equatorward transport to timescales of o30 years with effective spreading rates of 1–2 cm s 1. Contribution to quantifying formation rates and decadal climate variability in the Arctic, Greenland and Labrador Seas. Estimates for the formation rates of Weddell Sea Deep and Bottom Waters, production rate of Antarctic Bottom Water and pathways and timescales for spreading into the North Atlantic.

Upper Ocean Circulation

The use of CFCs for upper ocean processes has involved the application of concentrations to deduce sources and circulation pathways, and application of pCFC ages. The following highlights some of the

• • • •

•

Identification of the Sea of Okhotsk and Alaskan Gyre as important location for the ventilation of North Pacific Intermediate Water, these waters then spread into the subtropics on a timescale of o20 years. Quantification of the flux of water from the mixed layer into thermocline and intermediate layers of the North and South Pacific. Contribution to the description of sources and pathways of water masses transported from the Pacific through the Indonesian Seas into the Indian Ocean. Quantification of the sources of northern and southern water and the processes needed to ventilate the tropical Pacific and Atlantic, including advection, diapycnal and vertical mixing. Observation that pathways of the most recently ventilated Antarctic Intermediate Waters are into the eastern South Indian Ocean, while at that level there appears to be flow of older waters from the South Pacific into the western Indian Ocean. Quantification of subduction and formation rates for subtropical underwaters and in the North Atlantic its interannual variability that is negatively correlated with intermediate waters of the eastern subpolar gyre.

Model Constraints

In general CFC concentrations and inventories have been used in comparison with model simulated concentrations and inventories. The time-dependent nature of the CFCs provides a stringent test of a model’s ability to integrate property distributions over time. The following highlights some of the advances that have come about in our ability to put constraints on models from the use of CFCs in models:

• • • • •

Dilution of CFCs transported by the Deep Western Boundary Current and effect on tracer ages. Testing the sensitivity of a model for correct simulation of formation rates, pathways, and spreading rates. Testing the sensitivity of a model for correct simulation of ocean model velocity fields. Determining the model sensitivity to subgrid scale mixing for purposes of estimating ventilation rates. The importance of considering seasonal variations in the upper oceans as part of the tracer boundary


538

• •

CFCS IN THE OCEAN

conditions when trying to simulate subduction processes. Demonstration in a model simulation that eddy transport is required to transport South Indian subtropical gyre waters across the equator along the western boundary. Use of CFCs to validate model parameterizations of gas fluxes.

Biogeochemical Processes

The tracers provide a method for calculating rates of biogeochemical fluxes that is independent of direct biological measurements. Again the age information from the CFCs is used to calculate rates for these processes. (see Nitrogen Isotopes in the Ocean). The following highlights some of the advances that have come about in our understanding of biogeochemical processes to which observations of CFCs have contributed:

• • •

Apparent oxygen utilization rates from the central Arctic that are so high, they need to be balanced by transport of high production water from over the continental shelves. Quantification of moderate biological consumption and initially low oxygen concentrations in the Arabian Sea are needed to maintain the low oxygen layer. Calculation of denitrification rates for the Arabian Sea and Bay of Bengal.

Conclusions The advantage of oceanic tracers like CFCs is that they can be used to provide timescale information for oceanographic processes. Direct application of the timescale information from the CFCs is used to calculate fluxes of atmospheric constituents, such as CO2. The oceans have taken up a considerable portion of the anthropogenic CO2 released to the

atmosphere. A large part of the uptake involves water mass formation in high latitudes. The rate at which these waters are transported into the interior will have an effect on the rate at which anthropogenic CO2 is taken up.

See also Abyssal Currents. Air–Sea Gas Exchange. Carbon Dioxide (CO2) Cycle. Current Systems in the Atlantic Ocean. Current Systems in the Indian Ocean. Current Systems in the Southern Ocean. Nitrogen Isotopes in the Ocean. Ocean Subduction. Tritium–Helium Dating. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Broecker WS and Peng T-H (1982) Tracers in the Sea. Palisades, NY: Lamont-Doherty Geological Observatory, Columbia University. Fine RA (1995) Tracers, time scales and the thermohaline circulation: the lower limb in the North Atlantic Ocean. Reviews of Geophysics 33: 1353--1365. Rowland FS and Molina MJ (1994) Ozone depletion: 20 years after the alarm. Chemical & Engineering News 72: 8--13. Schlosser P and Smethie WS (1994) Transient Tracers as a Tool to Study Variability of Ocean Circulation. Natural Climate Variability on Decadal-to-Century Time Scales, pp. 274--288. Washington, DC: National Academic Press. Smethie WS, Fine RA, Putzka A, and Jones EP (2000) Reaching the flow of North Atlantic Deep Water using chlorofluorocarbons. Journal of Geophysical Research 105: 14 297--14 323. Walker SJ, Weiss RF, and Salameth PK (2000) Reconstructed histories of the annual mean atmospheric mole fractions for the halocarbons CFC-11, CFC-12, CFC-113 and carbon tetrachloride. Journal of Geophysical Research 105: 14 285--14 296.


CHEMICAL PROCESSES IN ESTUARINE SEDIMENTS W. R. Martin, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The physical and chemical environments of estuarine sediments vary over wide ranges. The sediments may lie beneath nearly fresh water or may have salinities near that of the open ocean. They may be intertidal or always under several meters of water. They may be covered by grasses, by algal or bacterial mats, or may be free of surface biological cover. They may be permeable sands, affected by groundwater flow, or may be impermeable muds. Nonetheless, common features determine their chemical environments. They exist in productive, coastal waters and therefore experience large fluxes of particulate organic matter. They experience large inputs of terrestrial material that supply Fe and Mn oxides. Most estuarine sediments lie underneath oxic water columns, and support abundant macrofauna that both mix sedimentary particles and irrigate the seabed. Common features such as oxic bottom water, abundant supplies of organic matter, Fe and Mn oxides, a supply of sulfur from seawater (except for freshwater end members of estuarine systems), and active macrofauna produce the set of biogeochemical processes that determines the chemical environment and its effects on carbon, nutrient, and contaminant cycling in the coastal ocean. The upper sediment column in estuarine and coastal environments can be regarded as a slowly stirred reactor to which substrates are added at the top and mixed downward. Transport processes, both in the dissolved and solid phases, distribute reactants within the reactor, while at the same time biogeochemical processes alter the particles and the pore waters surrounding them. This combination of transport and reaction drives solute exchanges between the sediments and overlying waters that are important to coastal carbon, nutrient, and contaminant cycles. It transforms the ‘reactive’ component of sedimentary particles so that the accumulating sediments differ in important ways from the particles that fall through the water column to the seafloor. These processes are outlined in Figure 1.

Because the rates of chemical reactions in this system are faster than mixing rates, the reactor is not chemically uniform; rather, roughly speaking, it has a layered structure. In all but a few locations, there is a thin, oxic layer at the sediment surface. As particles pass through this layer and through the underlying anoxic sediments, they are altered by heterotrophic respiration: organic matter is gradually oxidized, and a series of oxidants – O2, N(V) in NO3 , Mn(IV) and Fe(III) in particulate oxides, and S(VI) in SO4 2 – are consumed. Some of the energy originally stored in organic matter remains in the reduced products of these reactions: Fe(II), and S(–II), and, to some extent, Mn(II). This energy is available to chemolithotrophic bacteria, and the result is rapid internal cycling between oxidized and reduced forms of Mn, Fe, and S. Reduced Mn, Fe, and S that escape bacterial oxidation may be oxidized abiotically by O2 in the surface oxic layer. The result of these processes is sediments with an oxic ‘cap’, in which precipitation of newly formed Fe and Mn oxides may limit benthic/ water column exchange of many solutes, and an underlying anoxic layer in which rapid cycling between oxidized and reduced forms limits the rate of burial of products of heterotrophic respiration. The chemical processes defining this system are oxidation–reduction reactions, including ‘heterotrophic respiration’, by which organic matter is oxidized, ‘chemolithotrophic respiration’, most importantly oxidizing Fe(II) and S(–II), and ‘abiotic oxidation’; and ‘authigenic mineral formation’, which converts dissolved products of these reactions to solid phases. The stirring within the sedimentary reactor, which is vital to its operation, is accomplished by solute diffusion and particle mixing, accomplished predominantly by the feeding and burrowing activities of sedimentary fauna. In the following paragraphs, oxidation/reduction reactions in sediments and authigenic mineral formation are discussed. The effects of these reactions on the sedimentary environment are illustrated using profiles of solutes in sedimentary pore waters at a near-shore location. Then, there is a brief discussion of the role that sedimentary chemical processes play in nutrient cycles and the cycling of anthropogenic contaminants. It is important to note that this article is not a comprehensive review, but instead focuses on the range of chemical reactions occurring in coastal


539

540

CHEMICAL PROCESSES IN ESTUARINE SEDIMENTS

Particulate inputs Solute exchanges

Particulate

Dissolved

Interface

Chemical transformations

Mixing

Mixing

Accumulating sediments Figure 1 Biogeochemical cycles in estuarine and coastal sediments. Particles that fall to the sediment–water interface simultaneously undergo chemical reactions and are mixed through the upper few centimeters of the sediment column. Chemical transformations alter the particles and the sedimentary pore waters surrounding them, driving solute exchange between sediments and the water column and determining the composition of accumulating sediments.

sediments and their role in creating sedimentary environments that are important in coastal carbon, nutrient, and contaminant cycles. Therefore, a single type of coastal sediment – muddy, impermeable sediments underlying oxic bottom water – is discussed. Other sediment types are important: for example, permeable, sandy sediments are widespread, and sediments underlying permanently or seasonally anoxic bottom water can be important. In some high-energy areas, the effects of sediment resuspension can dominate sediment–water-column exchange. Although this article emphasizes sediments as sites of respiration, primary production at the surface of sediments in shallow waters can be important. There are important differences between sedimentary environments that determine their involvement in coastal biogeochemical processes. Nonetheless, similar chemical processes occur in all of these environments. The goal of this article is to illustrate these fundamental chemical processes, so that their role in a range of sedimentary environments can be examined.

Oxidation/Reduction Reactions The distinctive chemical environment of estuarine and coastal sediments is determined to a large extent by electron transfer, or ‘redox’ reactions, in which a

chemical species with a greater affinity for electrons accepts these carriers of negative charge from a species with a lesser affinity. In the process, energy is released. Although these reactions are often slow, they are catalyzed by organisms – in particular, microbes – that can harness the released energy. To understand the chemical environment of sediments, it is necessary to know which redox reactions will occur. Oxidants and Reductants

In chemical terms, we can consider some constituents of sediments as oxidants, capable of accepting electrons under conditions extant in the sediments, and reductants, capable of donating them. The primary sources of reductants are the organic matter, formed by production in the water column or on the sediment surface or by anthropogenic activities, that reaches the sediments, and the reduced products of sedimentary reactions. Oxidants diffuse from bottom water (O2, NO3 , and SO4 2 ), arrive to the sediments with the particulate flux (Fe and Mn oxides), or are formed by sedimentary processes (NO3 and authigenic Fe and Mn oxides). The principal electron donors and acceptors are listed in Table 1. pe: A Measure of the Affinity to Accept Electrons

Free electrons do not exist in aqueous solution. Nonetheless, since each electron acceptor (or donor)



Table 1

541

Oxidants and reductants in shallow-water sediments

e acceptor/donor

Formal oxidation state

Chemical form

Main sources

Oxidants O N Mn

0 þV þ IV, þ III

Dissolved O2 NO3 Mn oxyhydroxides

Fe

þ III

Fe oxyhydroxides

S

þ VI

SO4 2

Solute transport from bottom water Sedimentary nitrification Terrigenous material Cycling within sediments Terrigenous material Cycling within sediments Solute transport from bottom water Cycling within sediments

A range of oxidation states

Organic matter

Marine production and anthropogenic

þ II þ II II

Dissolved Mn2 þ Dissolved Fe2 þ FeS Dissolved HS FeS

Product of microbial respiration Product of microbial respiration Product of microbial respiration

Reductants C N S Mn Fe S

can, in principle, react with several different electron donors (or acceptors), it is useful to consider ‘halfreactions’ of the sort 1 4O 2

þ

þH þe

312H2 O

ðH2 OÞ1=2 ðO2 Þ1=4 Hþ ðe Þ

½2

in which parentheses denote activities of the species. Using the definitions, pH ¼ log(H þ ) and pe ¼ log(e ), we can rewrite this equation: ( pe ¼ logfKg log

ðH2 OÞ1=2 ðO2 Þ1=4

DG0 ¼ 12G0f ðH2 OðlÞÞ 14G0f ðO2 ðaqÞÞ

½1

In this reaction, the oxygen atom starts in the 0 oxidation state and ends in the –II state. To undergo this change, each O atom ‘accepts’ two electrons (from an unspecified source). Thus, 1/4 of a mole of O2 molecules accepts 1 mol of electrons. In formal terms, we can write an ‘equilibrium constant’ for the half-reaction: K¼

half-reactions allows the calculation of DG0 for reaction [1] as

) pH

½3

The equilibrium constant, K, can – in principle – be determined from the free energies of formation of the reactants and products through the relationship between DG0 and K. However, since free electrons do not exist in aqueous solution, chemists have devised a way of eliminating the free electron from the reaction. Half-reactions of the sort shown in eqn [1] are combined with with the oxidation of H2(g) to H þ (aq). This reaction releases one electron and is ‘defined’ to have a free energy change of 0. Some algebraic manipulation will show that this combination of

½4

Then, logðKÞ ¼

DG0 pe0 2:303RT

½5

can be calculated from tabulated thermodynamic data. In these equations, DG0 is the Gibbs free energy yield of the reaction, when all reactants and products are in their standard states, R is the gas constant, and T is temperature (in K). Now, eqn [3] can be rewritten as ( pe ¼ pe0 log

ðH2 OÞ1=2 ðO2 Þ1=4

) pH

½6

A similar equation can be derived for any halfreaction, written as a reduction of the reactant involving a single electron. Then, three quantities can be calculated: 1. pe0, with all reactants and products having activity ¼ 1; 2. pe(sw), which is applicable to reactions in seawater, and therefore is calculated with pH ¼ 7.5 and ½HCO3 ¼ 2 mM, but other species with activity ¼ 1; 3. pe, with all species at concentrations that are observed in the environment. pe0 (calculation (1)) is not very useful for environmental calculations, as in situ activities are far from 1.


542


Table 2

Redox couples important to shallow water marine sediments.

Reaction

pe pe0

1 4O2

þ Hþ þ e -12H2 O

1 5NO3

1 þ 65Hþ þ e -10 N2 þ 35H2 O

þ 2Hþ þ e -12Mn2þ þ H2 O

pe(sw)

21.49

13.99

20.81

11.81

21.54

6.54

FeOOH þ 3Hþ þ e -Fe2þ þ 2H2 O

15.98

6.52

þ 98Hþ þ e -18HS þ 12H2 O

4.21

4.23

1.43 14.88

7.95 5.41

1 2MnO2

2 1 8SO4

1 5 þ 1 1 4HCO3 þ 4H þ e -4CH2 O þ 2H2 O þ 1 5 þ 1 3 8NO3 þ 4H þ e -8NH4 þ 8H2 O

pe O2 ¼ 350 mM ¼ 1 mM N2 ¼ 500 mM NO3 ¼ 1 mM ¼ 30 mM Mn2 þ ¼ 1 mM ¼ 50 mM Fe2 þ ¼ 1 mM ¼ 400 mM HS =SO4 2þ ¼ 0:1 mM=28 mM ¼ 1 mM/1 mM HCO3 ¼ 2 mM

13.1 12.5 10.94 11.24 9.54 8.69 0.52 3.12 4.12 4.23 8.62

The reactions are written as one-electron half-reactions. pe values are calculated from thermodynamic data from Stumm and Morgan (1996). Note that the reduced form of carbon is shown as ‘CH2O’. This commonly used shorthand denotes a carbohydrate (not formaldehyde), and its standard state free energy is taken to be one-sixth that of glucose (C6H12O6, or, in this notation, ‘(CH2O)6’).

DGredox ¼ 2:303RT ðpereduction peoxidation Þ

½7

When the species that is oxidized lies below the species that is reduced on the pe scale, the reaction will release energy and can occur either abiotically or via microbial catalysis. The Oxidation of Organic Matter in Estuarine Sediments

Most of the organic matter falling to the seafloor in estuarine and coastal environments is formed by local primary production. A useful (but not quite accurate) representation of marine organic matter is ðCH2 OÞ106 ðNH3 Þ16 ðH3 PO4 Þ

½8

15

O2/H2O −

10

NO3/N2 Mn(IV)/Mn(II) −

+

5

NO3 /NH4 (org)

0

Fe(III)/Fe(II)

pe

pe(sw) is often used because the difference between this calculation (2) and pe (3) is often small. pe0, pe(sw), and pe values are listed in Table 2 for the electron acceptors and donors in Table 1, and they are shown graphically in Figure 2. The order of decreasing pe (Figure 2) is the order of decreasing affinity of the oxidant species for electrons. All reactions in Table 2 and Figure 2 are written as reductions (a species accepts an electron) involving the transfer of one electron. Thus, complete redox reactions, in which one chemical species is reduced while a second is oxidized, can be constructed by combining the reaction as written in Table 2 for the oxidant (which is reduced) with the reverse of a reaction that is below it on the pe scale. The resulting reaction will have DGo0:

−

2−

−5

SO4 /HS −

−10

HCO3 /CH2O (org) Redox couple

Figure 2 A graphical depiction of the pe of several environmentally important reduction reactions. A reduction as depicted in this figure, when combined with the oxidation of the reduced member of a couple listed lower on the figure, will yield a DGo0 under the conditions at which the pe values were calculated.

In this simple model, organic C is assumed to be in the form of a carbohydrate, and organic N in the form of ammonia (or a primary amine). In actuality, organic C has a lower average oxidation state than shown in this representation, but the model is still useful and widely used. We have used this organic matter stoichiometry, and neglected P since it is not oxidized during organic matter breakdown, to calculate the relative free energy yield for the oxidation of the C and N in organic matter by the different



Relative 2G (kJ (mol CH2O)−1)

−600 −500 −400 −300 −200 −100 0 O2

NO3

Mn(IV)

Fe(III)

SO4

Electron acceptor Figure 3 The relative free energy yield of the oxidation of a ‘model’ organic compound by the different electron acceptors. The model compound was (CH2O)(NH3)16/106. DG values were calculated using the data in Table 2. NH3 was assumed to be oxidized to NO3 by O2 and to N2 by NO3 .

electron acceptors. CH2O is oxidized to CO2 (converted to HCO3 in the marine environment) and NH3 to NO3 (oxidation by O2), N2 (oxidation by NO3 ), or NH4 þ (Mn(IV), Fe(III), S(VI)). The result is shown in Figure 3. The vertical bars shown for each electron acceptor represent the range in DG expected over the ranges concentrations of reactants and products observed in sediments. There is a distinct order of DG. In order of free energy yield per mole of CH2O, the order of yield by reaction with the different electron acceptors is O2 > NO3 > MnðivÞ > FeðiiiÞ SO4 2

½9

Observations in a variety of sedimentary environments confirm that this order of electron acceptor use is followed. Apparently, organisms that can obtain a higher energy yield per mole of organic matter oxidized are favored over those obtaining less energy per mole. When an electron acceptor with a higher free energy yield is depleted, the next in the order is used. Under the circumstances depicted in Figure 1, with reactants supplied at the sediment surface and mixed downward, the sequence of reactions by which organic matter is oxidized leads to a predictable sequence of solute concentrations. Starting from the sediment surface and moving downward: first, O2 is removed from the pore waters; over the depth range where O2 is removed, NO3 is added; it is then removed as NO3 is used as an electron acceptor; when NO3 is depleted, Mn2 þ is added to the pore waters, followed by Fe2 þ ; then, SO4 2 is removed and H2S is added.

543

An example of this sequence of concentrations is shown in pore waters from a coastal site in Figure 4. The site from which the pore water profiles shown were taken lies under a 5-m water column with salinity B31 psu and bottom water dissolved O2 concentrations ranging from 240 to 390 mM during the year. The annually averaged organic carbon oxidation rate is high, B800 mmol C cm 2 yr 1. Clearly, under these conditions, O2 is very rapidly consumed in the sediments. The absence of dissolved Fe2 þ , which is very rapidly oxidized by O2, in the upper 2 mm of the sediments indicates that the O2 flux from overlying water is consumed within the upper 2 mm of the sediment column. There is little NO3 in the water overlying the sediments, but, as for O2, its supply to the sediments is consumed in the upper 2 mm. Dissolved Mn2 þ appears in the upper 2 mm, indicating that its production by oxidation of organic matter (and potentially other reduced phases) occurs very near the sediment/water interface. The shape of the dissolved Fe2 þ profile shows that its release begins within 3–7 mm of the sediment– water interface. Finally, the SO4 2 profile shows that sulfate reduction begins by B20 mm below the interface. In these high-carbon-flux environments, the redox zones are compressed into a very small region near the sediment–water interface; nonetheless, the pore water data are consistent with the order of use suggested by DG calculations. This order has been observed at many locations. The site shown falls in the class of ‘sulfide-dominated’ sediments, in which the sulfate reduction rate is rapid enough to lead to a buildup of H2S in the pore waters. Other sediments, either with lower salinity (hence a smaller source of SO4 2 ) or lower organic matter oxidation rates (hence a lower sulfate reduction rate), do not show this sulfide buildup. These sites are often considered ‘iron-dominated’ sediments because the pore water dissolved Fe2 þ concentration never drops to near 0 and H2S never increases to high levels.

Authigenic Mineral Formation Reduced Fe and Mn Phases

The dissolved SO4 2 profile in Figure 4 clearly shows that sulfate reduction, which produces H2S, occurs in the region of the sediments where there are very large concentrations of Fe2 þ in the pore waters. However, no H2S is present in the pore waters in this depth region, and H2S does not appear until the dissolved Fe2 þ concentration begins its rapid decline to values o1 mM. The explanation for both the absence of H2S and the decline of Fe2 þ is the formation of solid-phase FeS. It is believed that the solid phase


544


Mn (µM)

Depth (cm)

0

Fe (µM)

25

0

0

0

5

5

10

10

15

15

20

20

THS (µM)

SO42− (mmol kg−1) 10 20

500

0

1000

0

0 5 10 15 20

5 10 15 20

Depth (cm)

N + N (µM) 0 4 8 0.0 0.5 1.0 1.5 2.0

Depth (cm)

TCO2 (µmol kg−1) 5 20 x 103

Alk (µmol kg−1) 5 20 x 103

NH4 (µM) 0 2500

TPO4 (µM) 0 200 400

0

0

0

5

5

5

0 5

10

10

10

10

15

15

15

15

20

20

20

20

Figure 4 Pore water concentration vs. depth profiles, measured at a site in Hingham Bay in Boston Harbor, Massachusetts, in October 2001. Different Symbols represent replicate cores taken at the site, all at the same time. The depth ranges over which chemical reactions occur in the sediments and the rates of these reactions can be inferred from the solute profiles. The methods used for making these inferences are outlined in a classic paper by Froelich et al. Two important assumptions underlying the interpretation of the profiles are: (1) They are approximately in ‘steady state’. That is, the rates of the transport and reaction processes whose balance sets the shape of the profiles are rapid relative to the rates at which environmental variability causes them to fluctuate over time. (2) The most important transport process in the sediments is vertically oriented molecular (or ionic) diffusion in the pore waters. Possible deviation from these simplifications can be seen in the pore water TCO2, alkalinity (‘Alk’), and NH4 þ profiles, which all have inflections near the depth at which H2S builds up in pore waters. These inflections could be caused either by transient changes in solute concentrations or by removal of the solutes from the sediments by ‘sediment irrigation’, a solute transport process due to the activities of animals living in the sediments. Data from Morford JL, Martin WR, Kalnejais LH, François R, Bothner M, and Karle I-M (2007) Insights on geochemical cycling of U, Re, and Mo from seasonal sampling in Boston Harbor, Massachusetts, USA. Geochimica et Cosmochimica Acta 71: 895–917.

formed is a fine-grained, poorly crystalline form of mackinawite (which we will call FeSm). Studies of the dissolution of mackinawite have shown that, in alkaline solutions (such as seawater), FeSm may be in equilibrium with either dissolved Fe2 þ and HS or with the uncharged species, FeS0aq. In either case, the total dissolved Fe2 þ is expected to be B1 mM. Thus, the relationship between [Fe2 þ ], [SO4 2 ], and total [H2S] is explained by precipitation of FeSm until the supply of dissolved Fe2 þ from dissimilatory Fe reduction is exhausted. Then, dissolved [H2S] begins to increase. In many cases, Fe(II) and S( II) do not remain sequestered in mackinawite. Mackinawite is metastable, and is subject to oxidation, either to Fe(III) and S(VI) or to the stable mineral, pyrite (FeS2). In ‘sulfide-dominated’ systems such as that shown in Figure 4, pyrite (FeS2) can form either through

oxidation of S in FeS by H2S: FeS þ H2 S ) FeS2 þ H2

½10

or by the reaction of aqueous FeS ‘clusters’ with polysulfide: 2 FeSaq þ S2 n ) FeS2 þ Sn1

½11

Thus, mackinawite is best viewed as an unstable intermediate product of iron and sulfate reduction. The fate of Fe and S present in mackinawite may be recycling to Fe(III) and S(VI), to be used again as electron acceptors for organic matter oxidation, or sequestration in pyrite. As noted above, not all sediments – particularly those in low-salinity environments – have elevated



H2S levels in pore waters. In these locations, Fe phosphate minerals may control Fe solubility. The combination of laboratory equilibration studies and pore water solute concentration measurements led Martens et al., for example, to conclude that deep pore waters were in equilibrium with vivianite – Fe3PO4 8H2O – in a coastal sediment. Hyacinthe et al. found that iron was sequestered as an Fe(III) phosphate in low-salinity, estuarine sediments. This ferric phosphate may have been formed in surface sediments (see below) or in the water column. Figure 4 also shows the removal of Mn2 þ from pore waters below the zone where it is added by reductive dissolution of Mn oxides. Equilibrium calculations have been used to infer that the mineral controlling Mn solubility in coastal and estuarine sediments is most likely to be a mixed Mn, Ca carbonate, xMn2þ þ ð1 xÞCa2þ þ CO3 2 ) Mnx Cað1xÞ CO3

½12

with a solubility product between those of calcite (CaCO3) and rhodochrosite (MnCO3): log KMnx Cað1xÞ CO3 ¼ x log KMnCO3 þ ð1 xÞlog KCaCO3

½13

Oxidized Phases

The removal of reduced Fe, Mn, and S to solid phases is readily apparent from the pore water solute profiles in Figure 4. Close inspection shows that dissolved Fe2 þ is also removed from solution above its pore water maximum, as dissolved [Fe2 þ ] drops to near zero before Fe2 þ reaches the sediment–water interface by diffusion. In addition, there is a small maximum in dissolved [SO4 2 ] just below the sediment–water interface, indicating its addition by oxidation of S( II). Finally, although the pore water Mn2 þ profile does not clearly show its removal, O2 (µmol l−1) 250

0

Depth (cm)

Depth (cm)

4

Fe2þ þ 12MnO2 þ H2 O ) FeOOH þ 12Mn2þ þ Hþ kJ DG7:5 ¼ 2:303RTð8:69 þ 0:52Þ ¼ 53 mol Fe ½14

Fe2+ (µmol kg−1) 250

0

0

0 2

direct measurement of the Mn2 þ flux across the sediment–water interface at this site (using in situ benthic flux chambers) showed that the actual flux is significantly smaller than is calculated by diffusion driven by the pore water concentration gradient: Mn2 þ must be removed in a thin layer at the sediment–water interface. These Fe and Mn removal processes are shown more clearly by pore water profiles at a site with a lower organic matter oxidation rate than the Figure 4 site. Figure 5 shows the seasonality in oxygen penetration depth that is common in shallow-water sediments. In March, the depth of removal of Fe2 þ from pore waters coincides closely with the O2 penetration depth, suggesting that there is rapid oxidation of Fe2 þ by O2 to insoluble oxides. In August, O2 penetration is much shallower and Fe2 þ production during organic matter oxidation is much more rapid, but the dissolved Fe2 þ profile still shows evidence of removal (the upward curvature of the profile is indicative of removal from solution). In this case, the removal coincides with both the zone of Mn2 þ release into the pore waters and with O2 penetration; oxidation by both O2 and Mn oxides may be occurring. The figure shows removal of Fe from solution by upward diffusion of dissolved Fe2 þ and its oxidation. As at the Figure 4 site, directly measured fluxes of dissolved Mn2 þ across the sediment–water interface show that removal of dissolved Mn in a thin layer at the sediment–water interface impedes its transport from sedimentary pore waters to bottom water. The removal of Fe(II), S( II), and Mn(II) when they are transported toward the sediment–water interface is predicted by the pe scale shown in Figure 2. The oxidation of Fe2 þ (the reverse of the reaction in Table 2) could be coupled to the reduction of MnO2,

2 4

Mn2+ (µmol kg−1) 20 40

0

Buzz Bay Mar. 03 Aug. 03

Depth (cm)

0

545

2 4

Figure 5 Pore water concentration vs. depth profiles, measured at a site in Buzzards Bay, Massachusetts, in Mar. and Aug., 2003.


546


or to the reduction of O2, Fe2þ þ 14O2 þ 32H2 O ) FeOOH þ 2Hþ kJ DG ¼ 2:303RTð13:1 þ 0:52Þ ¼ 78 mol Fe

½15

These reactions release energy, and can occur both abiotically or by microbial catalysis. Similar calculations using the data in Table 2 and Figure 2 show that S( II) can be oxidized by Mn oxides or O2, and Mn(II) can be oxidized by O2. The rapid oxidation of Fe and S by O2 is well known and occurs abiotically. Oxidation by O2 was demonstrated by Aller using sediments from Long Island Sound. Aller and Rude used incubation experiments to demonstrate the occurrence of microbially catalyzed oxidation of solidphase Fe(II) and S( II) by Mn oxides. Figure 5 illustrates these oxidative removal processes through loss of dissolved components of pore waters. Equally important is the upward mixing of solid phase Fe sulfides to zones of O2 or Mn oxide reduction, followed by their oxidation. Particle mixing by bioturbation is an important mechanism for transport of reduced phases toward the sediment–water interface, particularly in warm-weather (and high productivity) months. Measurement

The above discussion highlights the key role that authigenic solid phases play in sedimentary processes in estuarine sediments. Because rapid particle transport due to the activities of benthic fauna is present in virtually all of these sediments, solid phases are important participants in cycling between reduced and oxidized forms of Fe, S, and Mn. Unfortunately, the quantification of the solid phases in these sediments is difficult. Because the solids are predominantly made up of nonreactive, terrigenous minerals, and because the authigenic minerals tend to be very fine-grained and poorly crystalline, they cannot be measured directly. Instead, their presence is inferred by removal of dissolved constituents from pore waters and calculation of solubilities based on laboratory studies. The only means currently available to quantify their concentrations are selective chemical leaches of sediments. Several studies have evaluated the use of reducing agents to determine reactive Fe and Mn oxides in sediments. Hyacinthe et al. showed recently that ligand-enhanced reductive dissolution using a buffered ascorbate–citrate solution dissolves approximately the same amount of Fe as microbial processes, but the correspondence is still quite uncertain. The quantification of reduced phases is still more difficult. Cornwell and Morse compared

several procedures for quantifying reduced sulfur in sediments, concluding that dissolution in 6 N or 12 N HCl released S associated with FeS (and other relatively reactive reduced S species), but that the procedures also release Fe associated with oxides and some silicates. The procedure achieved reasonable separation of FeS from pyrite, which was dissolved by a more intense chemical leach. Rickard and Morse have emphasized the difficulty in interpreting the results of these procedures. Clearly, given the importance of cycling of solid phases in estuarine and coastal sediments, the quantification of their concentrations remains an important problem.

Elemental Cycling within Estuarine and Coastal Sediments It is clear that chemical processes do not involve just alteration of particles that fall to the sediment–water interface. Rather, these systems are dominated by cycling between oxidized and reduced chemical forms within the sediment column. Heterotrophic respiration leads to the oxidation of organic matter, releasing C, N, and P into solution, and the reduction of O2, Fe, Mn, and S. The latter three elements are then subject to transport, both in dissolved and solid forms, back toward the sediment–water interface, where they may be reoxidized either abiotically or by lithotrophic bacteria. The burial of reduced Fe, Mn, and S is a slow leak from these rapid internal cycles (Figure 6). The upward transport and oxidation of reduced phases introduces complications into the determination of the relative importance of different organic matter oxidation pathways. For instance, O2 is consumed not only for direct organic matter oxidation, but also for oxidation of reduced products of organic matter oxidation by Mn and Fe oxides and SO4 2 (Mn(II), Fe(II), S( II)). Similarly, Mn oxides are consumed by oxidation of Fe(II) and S( II) as well as by oxidation of organic matter. Thus, in considering the role played by each electron acceptor in organic matter decomposition, allowance must be made for oxidation of chemical constituents other than organic matter. Jorgensen used measurements of dissolved fluxes across the sediment–water interface to infer that oxidation of organic matter by O2 and SO4 2 were of roughly equal importance in shallowwater marine sediments. More recently, Thamdrup used sediment incubation techniques to show that those results overestimated the role of O2 but underestimated that of Fe. The uncertainties in earlier estimates were both due to the failure to account for oxidation of species other than organic matter and due to the failure to account for the



547

Particles

O2

Sediment /water interface Oxic layer

Oxic/anoxic boundary

Mn(IV), Fe(III), SO42−

Heterotrophic respiration

Mn(IV), NO3−, O2

Anoxic layer Chemolithotrophy, abiotic oxidation

Mn(II), Fe(II), S(−II)

CO2, inorganic nutrients

Mn(II), N2, or NH4+

Accumulating sediments Figure 6 The cycling of Fe, Mn, and S within the sediment column. Transport of reduced Fe, Mn, and S occurs both in the dissolved and solid phases.

formation of solid-phase FeS. O2 is still likely to be of great importance, however. For instance, while the Fe(III)/Fe(II) pathway may be used directly for organic matter oxidation, it is the internal cycling between Fe(III) and Fe(II) that allows this pathway to be important: the ultimate sink for electrons is likely to be O2. This linking of oxidation/reduction reactions in series is illustrated in Figure 7 using a diagram of the sort introduced by Aller.

The Effect of Sedimentary Redox Cycling on Nutrient Cycles Sedimentary respiration results in the decomposition of the majority of the organic matter that falls from

the water column to the sediments. However, sedimentary processes can result in less efficient cycling of P and N than C, with potentially significant implications for coastal productivity. Phosphate has a strong tendency to adsorb to Fe oxides. Therefore, the precipitation of Fe oxides near the interface of coastal and estuarine sediments can impede the return of phosphate, released into solution by the decomposition of organic matter, to the water column. Figure 4 shows that dissolved reactive phosphate (‘TRP’) can reach very high levels in sediments to which there is a large flux of organic matter. In fact, the TRP level in Figure 4 is significantly greater than would be predicted by the decomposition of organic matter alone. This phenomenon is further illustrated in Figure 8.


548


Figure 8(a) shows that P is highly enriched in the solid phase in the upper centimeter of the sediment column. This enrichment cannot be explained simply by larger organic matter concentrations in surface seciments. Figure 8(b) illustrates that internal P

Corg

Fe2+

FeOOH

CO2 Fe2+

Corg

O2 H2O

FeOOH

CO2

Mn2+

Fe2+

O2

MnO2

FeOOH

Corg

CO2

Figure 7 The oxidation of organic matter by FeOOH. In each panel, the oxidation of organic carbon (‘Corg’) to CO2 is coupled to the reduction of FeOOH to Fe2 þ . In the top panel, the source of the FeOOH is the particle flux to the sediments. The lower two panels show that internal cycling within the sediments can also supply Fe(III) for organic matter oxidation. In the middle panel, Fe(II) that was produced by organic matter oxidation is oxidized by O2 to Fe(III), to be reused ot oxidize organic matter. In the lower panel, Fe(II) is oxidized to Fe(III) by MnO2. The MnO2 for this reaction, in turn, may be supplied either by the particle flux to the sediments or by oxidation of Mn(II) to Mn(IV) by O2. The figure illustrates that the coupling of organic matter oxidation to Fe(III) reduction can be quantitatively important even if the supply of reactive Fe(III) through the fall or particles to the seafloor is small. When this occurs, the ultimate sink for the electrons for organic matter oxidation by the Fe(III)/Fe(II) pathway is O2.

(a)

1600

5NH4 þ þ 3NO3 ) 4N2 þ 9H2 O þ 2Hþ

½16

Quasi-equilibrium

(b)

Solid P (ppm) 1200

cycling with Fe establishes the enrichment. The dashed line in the figure shows the slope of a pore water TPO4 versus TCO2 plot if TPO4 is simply released by organic matter oxidation. If the only process affecting dissolved TPO4 were decomposition of organic matter, then the pore water concentration data would plot along a line of that slope from the origin. The pore water data deviate from that simple relationship in two ways: TPO4 is essentially completely removed from pore waters near the sediment– water interface by upward diffusion and removal with freshly formed Fe oxides. When Fe(III) is reduced to oxidize organic matter, the rate of P release is much greater than can be accounted for by the organic matter alone: the Fe oxide-associated P is released. The result is a strong internal cycle of P. A fraction of the P that arrives at the sediment surface with organic matter is sequestered within the sediments by precipitation with Fe oxides. Figure 8 illustrates this process at a marine sediment site. Hyacinthe et al. showed that P can also be sequestered in estuarine sediments when P that is removed with Fe oxides remains in the solid phase after incomplete reduction of the oxide phase. The sediments are also important sites for ther removal of fixed nitrogen from coastal waters. Because there is relatively little NO3 in the shallow waters overlying coastal and estuarine sediments, diffusion of NO3 from bottom water is not a major source of N for denitrification. However, rapid organic matter oxidation results in the release of NH4 þ to the pore waters. NH4 þ can be converted to N2 by a nitrification/denitrification cycle or by NH4 þ oxidation coupled to the reduction of NO3 :

400

2000

Release during Fe reduction 0 TPO4

300 Depth (cm)

5 10

25

200

Hull Bay Jan. 2002 Sep. 2002

100

15 20

Release with o.m. oxidation ΔP/ΔTC = 0.014 Red ==> 0.012

Hull Bay Jan. 2002

Removal in Jan.

0 5

10

15 TCO2

20

25x103

Figure 8 (a) Solid-phase P, measured at a site in Hingham Bay, Boston Harbor. (b) The relationship between the pore water concentrations of total reactive phosphate (TPO4) and TCO2 in pore waters at that site. Concentrations of both species are in mmol l 1.



Anthropogenic materials Input from water column

Terrigenous input

549

Marine production Solute exchange

Sediment resuspension

Oxic layer Sediment cycling layer

Fe reduction SO4 reduction H2S in pore waters

Rapid solidphase mixing

Anthropogenic materials deposited in the past

Figure 9 The structure of the chemical environment in shallow-water marine sediments.

Thamdrup and Dalsgaard illustrated the occurrence of this N2 formation mechanism. The result of these processes is that 15–75% of the flux of N to the surface of coastal and estuarine sediments may be removed to N2.

Contaminant Cycling: Anthropogenic Metals The ‘layered’ structure of the chemical environment in estuarine and coastal sediments has important implications for the cycling of contaminants. Anthropogenic materials can both be present in the sediments from deposition in the past and can arrive with current deposits. They encounter an environment similar to that depicted in Figure 9. The upper centimeters of the sediment can be seen as an active, ‘sediment-cycling layer’, in which particles are rapidly mixed by bioturbation. This layer consists of an oxic ‘cap’ just below the sediment–water interface, where Fe and Mn oxides precipitate. Below that is a region of Fe and S reduction, but without buildup of dissolved H2S. Below this layer – if the supply of organic matter supports extensive sulfate reduction – is a layer of elevated dissolved H2S. An extensive discussion of contaminant cycling in estuarine and coastal sediments is beyond the scope of this article. Two examples are cited to show the importance of the redox structure of these environments to contaminant cycling. Many contaminant metals – Cu, Ag, and Pb are examples – both form insoluble sulfides and tend to sorb onto precipitating Fe oxides. Thus, they are released to pore waters when Fe oxide reduction occurs, but tend to be immobilized by precipitation as

sulfides. Dissolved fluxes to the overlying water are determined by a balance between their rates of diffusion from the site of Fe reduction toward the sediment/water interface and of removal by precipitation with Fe oxides. Because both the thickness of the oxic layer and the rate of Fe reduction vary seasonally and spatially, these metals have seasonally and spatially variable rates of release from sediments to the water column. Their precipitation with Fe oxides can lead to enrichments in fine-grained solids at the sediment– water interface. Thus, sediment resuspension and subsequent reaction and transport in the water column can be an important mechanism for redistribution of contaminant metals. Because the FeS that is formed by precipitation of the products of Fe(III) and SO4 2 reduction is metastable, the formation of pyrite may be an important step in the immobilization of these metals. One example of a study of these processes in coastal systems is that of Kalnejais. A different type of metal is Hg. The toxic, bioavailable form of Hg (methylmercury) may be formed in sediments when sulfate-reducing bacteria incorporate HgS0 and methylate the Hg. This process appears to occur where sulfate reduction is important, but dissolved sulfide levels are less than B10 mM. Under these conditions, sediments can be an important source of methylmercury to the coastal water column.

See also Nitrogen Isotopes in the Ocean. Phosphorus Cycle. Sediment Chronologies. Sedimentary Record, Reconstruction of Productivity from the


550


Further Reading Aller RC (1980) Diagenetic processes near the sediment– water interface of Long Island Sound. Part II: Fe and Mn. In: Saltzman B (ed.) Estuarine Physics and Chemistry: Studies in Long Island Sound, vol. 22, pp. 351--415. New York: Academic Press. Aller RC and Rude PD (1988) Complete oxidation of solid phase sulfides by manganese and bacteria in anoxic marine sediments. Geochimica et Cosmochimica Acta 52: 751--765. Benoit JM, Gilmour CC, and Mason RP (2001) The influence of sulfide on solid-phase mercury bioavailability for methylation by pure cultures of Desulfobulbus propionicus. Environmental Science and Technology 35: 127--132. Cornwell JC and Morse JW (1987) The characterization of iron sulfide minerals in anoxic marine sediments. Marine Chemistry 22: 193--206. Elderfield H, Luedtke, McCaffrey RJ, and Bender M (1981) Benthic studies in Narragansett Bay. American Journal of Science 281: 768--787. Froelich PN, Klinkhammer GP, Luedtke NA, et al. (1979) Early oxidation of organic matter in pelagic sediments of the eastern equatorial Atlantic: Suboxic diagenesis. Geochimica et Cosmochimica Acta 43: 1075--1090. Hammerschmidt CR, Fitzgerald WC, Lamborg CH, Balcom PH, and Visscher PT (2004) Biogeochemistry of methylmercury in sediments of Long Island Sound. Marine Chemistry 90: 31--52. Huerta-Diaz MA and Morse JW (1992) Pyritization of trace metals in anoxic marine sediments. Geochimica et Cosmochimica Acta 56: 2681--2702. Hyacinthe C, Bonneville S, and Van Cappellen P (2006) Reactive iron(III) in sediments: Chemical versus microbial extractions. Geochimica et Cosmochimica Acta 70: 4166--4180. Hyacinthe C and Van Cappellen P (2004) An authigenic iron phosphate phase in estuarine sediments:

Composition, formation and chemical reactivity. Marine Chemistry 91: 227--251. Jorgensen BB (1982) Mineralization of organic matter in the sea bed – the role of sulphate reduction. Nature 296: 643--645. Kalnejais LH (2005) Mechanisms of Metal Release from Contaminated Coastal Sediments, p. 238. Woods Hole, MA: Massachusetts Institute of Technology, Woods Hole Oceanographic Institution. Kostka JE and Luther GW (1994) Partitioning and speciation of solid phase iron in saltmarsh sediments. Geochimica et Cosmochimica Acta 58: 1701--1710. Martens CS, Berner RA, and Rosenfeld JK (1978) Interstitial water chemistry of anoxic Long Island Sound sediments. Part 2: Nutrient regeneration and phosphate removal. Limnology and Oceanography 23(4): 605--617. Morford JL, Martin WR, Kalnejais LH, François R, Bothner M, and Karle I-M (2007) Insights on geochemical cycling of U, Re, and Mo from seasonal sampling in Boston Harbor, Massachusetts, USA. Geochimica et Cosmochimica Acta 71: 895--917. Rickard D (2006) The solubility of FeS. Geochimica et Cosmochimica Acta 70: 5779--5789. Rickard D and Morse JW (2005) Acid volatile sulfide (AVS). Marine Chemistry 97: 141--197. Seitzinger SP (1988) Denitrification in freshwater and coastal marine ecosystems: Ecological and geochemical significance. Limnology and Oceanography 33: 702--724. Stumm W and Morgan JJ (1996) Aquatic Chemistry: Chemical Equilibria and Rates in Natural Waters, 3rd edn. New York: Wiley. Thamdrup B (2000) Bacterial manganese and iron reduction in aquatic sediments. Advances in Microbial Ecology 16: 41--84. Thamdrup B and Dalsgaard T (2002) Production of N2 through anaerobic ammonium oxidation coouled to nitrate reduction in marine sediments. Applied and Environmental Microbiology 68: 1312--1318.


CHLORINATED HYDROCARBONS J. W. Farrington, Woods Hole Oceanographic Institution, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 450–462, & 2001, Elsevier Ltd.

Introduction These chemicals are considered in a pollution category because both deliberate and accidental release to the environment of several of these types of compounds, for example the industrial chemicals such as PCBs (polychlorinated biphenyls) and the chlorinated pesticides p,p0 DDT (dichlorodiphenyltrichloroethane; formal chemical name 1,10 -(2,2,2trichloroethylidene)-bis (4-chlorobenzene)), have had unintended adverse environmental effects on diverse plants and animals and on people. Initially, chemicals such as PCBs and DDT were beneficial to human civilization: PCBs as industrial chemicals allowing economical, safe delivery of electricity, and DDT as a pesticide eradicating vector pests of human health concern and agricultural crop pests. Only after these chemicals had entered widespread use did it become apparent that there were environmental problems, although in hindsight there was evidence of potential problems early in the history of their manufacture and use. Chlorinated hydrocarbons are chemicals made up of the elements carbon (C) and hydrogen (H) at the combine for the ‘hydrocarbon’ part of the molecule, and chlorine atoms (Cl) substituted for hydrogen where a hydrogen atom was normally bonded to a carbon atom. Examples of structures of chlorinated hydrocarbons are given in Figure 1. Chlorinated hydrocarbons have a wide range of molecular weights (related to size), and complexity, i.e., there are various distinct configurations or arrangements of constituent atoms. For example, there are 209 individual chlorobiphenyls (known collectively as congeners) making up the family of chemicals known as PCBs. Not all of these are present in the commercial chemical mixtures of PCBs, but there are usually 20–50 chlorobiphenyl congeners in a given commercial mixture. Smaller molecules among the class of chlorinated hydrocarbons, such as tetrachloroethylene, trichloroethylene, and carbon tetrachloride are used for activities such as degreasing of machinery and dry cleaning. Presently, these compounds are of

environmental concern for the oceans only in the near-shore coastal areas where their presence in sewage effluents and contaminated or polluted ground water interfacing with coastal sea water results in elevated concentrations in coastal waters near sources of input. Environmental and human health concerns associated with chlorinated pesticides and PCBs have evolved over the past several decades into wider concerns with organochlorine compounds of all types, ranging from those found in plastic trash bags to the chemicals of Agent Orange defoliant used by the United States during the war in Vietnam. Among the chemicals of greatest concern on a per unit amount basis are tetrachlorodibenzodioxins; often the name is shortened in general public use to ‘dioxins’. Assessments of risks to human health and wildlife for the various chlorinated hydrocarbon pesticides and industrial chemicals are often expressed relative to tetrachlorodibenzodioxin risks. There are an estimated 10 000 to 11 000 organochlorines in commercial production and many thousands more may be present, but are as yet unidentified, as by-products of the production. In addition, processes such as chlorination of sewage effluent to kill bacteria, result in active forms of chlorine which react with natural organic chemicals in the sewage to produce a myriad of organochlorines; perhaps hundreds to thousands depending on the effluent and the chlorination conditions. Small amounts of organochlorines are also reported to result from various combustion processes, both natural fires and volcanic eruptions, and human-controlled processes such as incineration of wastes. Analytical chemical, biochemical and molecular biological methods can detect very low concentrations of these compounds in environmental samples, including marine organisms (ng g1, or about one unit mass of chlorinated hydrocarbon molecule per billion unit masses of tissue molecules), sea water (ng per 1000 liters or kg, ng kg1), and marine sediments (nanograms per gram of sediment, ng g1). Given the widespread occurrence of these compounds, and the known or suspected adverse environmental and human health effects at elevated concentrations, there is a challenge, as with most chemicals of environmental concern, in establishing a ‘safe’ concentration in environmental samples. This has stimulated intense debate among environmental activists, the chemical industry, researchers, government officials and the public about the adverse


551

552

CHLORINATED HYDROCARBONS

H Biphenyl molecule

H

H

H C

C

C

Cl

C

C

C

H

H

H

H

H

H Cl

Cl

C

C

Pentachlorobiphenyls

H C

C C

H

One example of a PCB

C

C

C

C

C

H

Cl

C

C

C

C

C

C

H

Cl

Cl

Cl Cl

H

H

Cl

Cl

Cl

Cl Cl

Cl

Cl

Cl

C

Cl

Cl Cl

Cl

Cl

Cl

Cl Cl Cl

Cl

Cl Simpler chemical representation of the same PCB

Cl

H

Cl C

Hexachlorobiphenyls

Cl

Cl

Cl Cl

Cl

Cl

Cl

Cl Cl

Cl Cl

Cl Cl

Cl

Cl Cl

(A)

Cl Cl

C

Cl

Cl

Cl

Hept-[7], octa-[8], nona-[9] chlorobiphenyls and the single decachlorobiphenyl are not as common in use or in the environment

Cl Cl

Cl

H

Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl

H C

Cl

Cl

C

Cyclohexane

CCl3

CHCl2

CCl2

o, p′ – DDT

o, p′ – DDD

o, p′ – DDE

H Cl

Alpha

Cl

Cl

CCl3

C

Cl

Cl

C

CHCl2

Cl

Cl

Cl Cl

o designates ortho or next to the central carbon atoms

H

H

Cl

Cl Cl

H

Cl

Cl

Cl

Alpha Chlordane

Trichlorobiphenyls Cl

Tetrachlorobiphenyls Cl Heptachlor

Cl

Cl

H

Gamma

Examples of PCBs

Cl

Cl

Cl

designates the right hexagonal or phenyl molecule

Monochlorobiphenyls

Cl

Cl

Cl

Cl

Cl

p designates para or across from the central carbon atoms ′

Cl Cl

Cl

CCl2 p, p′– DDE

p, p′ – DDD

p, p′ – DDT

Cl

Cl Cl

Cl Cl

Cl Cl Cl

Cl

Cl

Cl

Cl Cl

H

C

Gamma

Beta Cl

Cl

Cl

Cl Cl Cl Cl

Cl

Cl

Cl

Cl

Endrin

Hexachlorobenzene

Cl

Cl

Cl

Cl Cl

Cl Cl

Cl

Cl Cl

Cl

Cl

Cl

Cl

Cl

Cl

Cl Cl

Cl

Cl

2,4,5-trichlorobiphenyl

Dichlorobiphenyls Cl

Cl

Cl Cl

Cl

Cl Cl

Cl

Cl Cl

Cl

O Cl

(B)

(C)

Cl

Cl Aldrin

Dieldrin

Figure 1 Chemical structures of (A) polychlorinated biphenyls (PCBs); (B) dichlorodiphenyltrichloroethane (DDT) and the metabolites dichlorophenyldichloroethane (DDD) and DDE; (C) other chlorinated pesticides. C, carbon atoms; H, hydrogen atom; Cl, chlorine atom.

environmental effects and human health risks associated with low amounts of organochlorine compounds, including chlorinated hydrocarbons. Lessons learned about the environmental behavior and adverse effects of chlorinated hydrocarbons provide guidance about what to expect for the more

general class of organochlorine compounds. Much information is available about chlorinated pesticides such as the DDTs (in this article DDT includes p,p0 and o,p-DDT and the immediate biodegradation and metabolism products DDE and DDD, see Figure 1) and PCBs. Thus, DDT and PCBs are used herein as



examples for the larger class of chlorinated hydrocarbons. However, the approach of using examples is pursued with the caveat that movement of each of the chlorinated hydrocarbons through the environment, and biological effects associated with each chemical, are specific in important details for each chemical. Although not discussed further here, it is important to note that these chemicals can and have been used as tracers of general processes acting on naturally occurring organic chemicals in marine ecosystems.

History DDT

DDT (dichlorodiphenyltrichloroethane) is not a naturally occurring compound. It was first synthesized in 1874 and its insecticidal properties were discovered in 1939. Initial large use of DDT as an insecticide began in 1944 and continued into the late 1960s. It was used with success against vectors of human diseases such as malaria and with dramatic effect in agriculture around the world in controlling insect pests. The environmental problems associated with DDT in terms of adverse effects on nontarget organisms such as birds were brought to popular attention in the highly influential book ‘Silent Spring’ by Rachael Carson in 1962. Further studies of DDT in the mid 1960s to early 1970s documented the presence of DDT and several other chlorinated pesticides in marine organisms at all major sectors in the marine food web. Analyses of samples from organisms dwelling in the deep part of the oceans, for example 4000–5000 m depth in the North Atlantic, and from Arctic and Antarctic marine ecosystems contained measurable concentrations of DDT. Evidence of adverse effects on nontarget terrestrial, freshwater and marine organisms, especially birds, resulted in curtailed use or bans on the use of DDT in several developed countries in the early 1970s. The legacy of past releases to the environment is present in marine ecosystems in the form of measurable concentrations of DDT compounds. In addition, the use of DDT continues in a few countries. PCBs

PCBs (polychlorinated biphenyls) have been used industrially since 1929. Industrial mixtures of PCBs are known by commercial names, e.g., Aroclors (United States), Kaneclor (Japan), Chlophen (Germany), Sovol (former USSR), Fenchlor (France). PCBs were widely used in insulating fluids in transformers and capacitors, as well as hydraulic systems, surface coatings, flame retardants, inks and other minor uses. Concerns about human health effects associated with halogenated aromatic compounds such as PCBs

553

and halowax (polychlorinated naphthalenes) date to the 1930s and 1940 with the reports of rashes and liver abnormalities for workers in manufacturing plants and electricians. The identification of polychlorinated biphenyls (PCBs) as chemicals of environmental concern dates to the late 1960s when they were reported nearly simultaneously by three different research groups to be present in seabirds and seabird eggs in three different coastal ecosystems. Subsequent research in the late 1960s and early 1970s confirmed the widespread presence of PCBs in numerous ecosystems, their relative persistence in the environment, and several instances of known or suspected adverse effects associated with various organisms exposed to and incorporating PCBs into their tissues, e.g., mink and chickens fed on fish or fish meal. In 1968, contamination of rice oil used in food preparation at a location in western Japan by PCBs from a leaking transformer caused human health effects for people who consumed the polluted food; this was designated the Yusho incident. A similar incident occurred in 1979–81 in Taiwan: the Yuncheng incident. Detailed studies of the PCB oil involved in these incidents suggested that some or all of the observed adverse effects may have resulted from the presence of small amounts of chlorinated bibenzonfurans or chlorinated dibenzodioxins. By 1971, the concerns about human health and environmental impacts led Monsanto, the producer of PCBs in the United States, to a voluntary ban on sales of PCBs except for closed systems use. Monsanto ceased all production in 1977 and there was no largescale increase in imports. PCBs were banned from production and further use in the United States in 1978. Equipment that already contained PCBs, e.g., transformers, were allowed to remain in use but restrictions were placed on the disposal of PCBs when the equipment was decommissioned. Delegates from 122 countries completed a draft treaty on persistent organic pollutants (POPs) in December 2000. The POPs that were initially addressed and banned from further use include chlordane, DDT, dieldrin, endrin, heptachlor, mirex, toxaphene, PCBs, hexachlorbenzene, chlorinated dibenzofurans, and chlorinated dibenzodioxins. Limited selective use of DDT for human disease vector control is allowed in some countries.

Distribution in the Marine Environment Early 1970s

Analyses during the early 1970s of various species of marine biota for DDT family chemicals and PCBs


554


established widespread distribution of these chemicals in many areas of the world’s oceans from equatorial to polar regions and to depths of 4000– 5000 m in the Atlantic Ocean. Biota with significant lipid (fat) in their body tissues tended to accumulate DDT family compounds and PCBs much more than biota with lower lipid content. Marine mammals and birds accumulated higher concentrations of DDT compounds and PCBs, presumably due to being near the top of the food web (biomagnification of contaminants), and having a high body lipid content (especially for marine mammals and birds). Analyses of surface seawater samples (both dissolved and particulate) and samples of the atmosphere over the oceans established the presence of low concentrations (ranges of 0.01–1 ng kg1 water or 0.001 ng m3 of air) of DDT family compounds and PCBs. Very low concentrations and difficulties in avoiding contamination from the sampling ship and the sampling gear made measurement of deep seawater samples problematic. The few deep-water samples analyzed documented that the concentrations of DDT and PCBs were not higher than about 0.001 ng kg1 of water. Confirmatory measurements were not made for years thereafter because of intense debate among chemical oceanographers about how to make reliable measurements for these compounds at very low concentrations in sea water. In contrast, the underlying sediments had accumulated sufficient concentrations, because of sorption on particles and deposition, to allow undisputed measurements of both DDT and PCB in deep ocean surface sediments. Several surveys and research programs documented much higher concentrations of DDT and PCBs in coastal waters and coastal ecosystems compared to open ocean ecosystems; especially near urban areas for DDT and PCBs, and in coastal regions near agricultural drainage areas for DDT, as might be expected given patterns of use for these compounds and probable release to the environment. 1980s to the Present Day

Open Ocean Recent measurements of PCBs and DDT family compounds in surface sea water and air samples over the ocean on a regional oceanic scale are few and are exemplified in Figure 2(A–D). Despite the paucity of data, some important findings are evident. There are higher concentrations of DDT in surface sea waters and air overlying the oceans near south-east Asia in comparison to the other areas sampled. This is consistent with environmental concerns associated with continued use of DDT in the Asian continent, South Asia subcontinent, and

Oceania areas beyond the years when DDT use was curtailed or eliminated in the developed countries of North America, Europe, Japan, and Australia. The PCB concentrations in the surface sea water are in the range of 1–60 pg kg1 and in the overlying atmosphere 3–600 pg m3. The distribution of concentrations is more even across the areas sampled compared to DDT (Figure 2). This is consistent with the continued presence of PCBs cycling in the environment as a result of past releases, leakage from landfills and products containing PCBs still in use, and perhaps continued new uses even though PCB manufacture has been eliminated or severely reduced in many countries. Recent progress with measuring low concentrations of PCBs in deep ocean waters has enabled a few measurements of deep ocean waters. A depth profile of the sum of concentrations of several chlorobiphenyl congeners for a station in the eastern North Atlantic (Figure 3) documents higher concentrations in surface waters with decreasing concentrations with depth as expected due to the greater contact of the surface waters with the atmosphere and contemporary environment. However, all the concentrations are very low in comparison to concentrations found in near-shore waters, lakes, and rivers. Concentrations in the deepest waters are below or at detection limits of the analytical methods used. PCBs in mid-depth and deeper waters are most likely a result of the flux of particles from the surface water carrying sorbed PCBs in and on particles to deeper waters. There has been considerable progress during the 1980s and 1990s in understanding the role and details of particles as conveyers of chemicals from the surface ocean to deep waters and sediments. As the particles sink through the deep waters, desorption and disaggregation of particles and subsequent desorption releases PCBs. Therefore, it is likely, though not proved due to lack of a series of data over time, that current deep water concentrations of PCBs reflect inputs from particle fluxes over the total time of PCB use and release to the environment; for example inputs from peak use in North America and Europe during the 1950s to 1960s. Deep-water sediments contain low concentrations of DDT and PCBs in the range of 1–100 ng g1 dry weight or parts per billion. Relatively few deepocean benthic (bottom dwelling) animals have been analyzed, but those that have been analyzed contain detectable concentrations of DDT and PCBs in the range of 0.001–1 mg g1 dry weight. Analyses of a few samples of mid-water fish in the deep ocean document the presence of PCBs in a pattern that reflects metabolism of the PCBs after uptake. Although small in numbers of samples analyzed,



_3

Sum of 15 chlorobiphenyl congeners (pg dm ) 0

1

2

3

4

5

6

0

500 Suspension 1000

Solution

Depth (m)

1500

Table 1 General types of responses to PCB contamination for marine organisms. (PCB concentrations in tissues or habitat types eliciting a given intensity of response varies with species and ecosystem) Level of biological organization

Types of responsea

Biochemical-cellular

Toxication Metabolic impairment Cellular damage Detoxication

Organismal

Physiological change Behavioral change Susceptibility to disease Reproductive effort Larval viability Immune responses

Population

Age, size structure Recruitment Mortality Biomass Adjustments in reproductivity and other demographic characteristics

Community

Species abundance Species distribution Biomass Trophic interactions

2000

2500

3000

3500

4000

4500 Figure 3 Depth profile of PCB concentrations in sea water, May, 1992. at 471N, 201W. (Adapted from Petrick et al., 1996.)

the organism’s food and tissues eliciting a given effect can range over many orders of magnitude from parts per million to parts per trillion. The near-shore and estuarine waters of the coastal ocean contain elevated concentrations of DDT and PCBs in comparison to the open ocean. Therefore, attention has been focused on obtaining more data for the coastal ocean. The data sets are more numerous and provide better geographic and temporal coverage for coastal areas of developed countries but much less so for most of the developing countries. Sufficient data have been collected in several areas and sufficient laboratory experiments have been completed to provide a reasonable general understanding of the inputs, fates and effects of DDT and PCBs in coastal ecosystems. Figure 4 shows a general depiction of the cycling of PCBs in a coastal ecosystem. One key aspect of this biogeochemical cycle is the uptake by animals of DDT and PCBs both from food sources and from water across membrane surfaces such as gills. Exceptions are air-breathing organisms such as birds and marine mammals for

557

a Responses are mostly adverse effects, but some are beneficial in offering protection against adverse effects. Adapted with permission from Farrington JW and McDowell JE (1994) Toxic chemicals in Buzzards Bay: Sources, fates, and effects. In: Costa JE, Gibson V and Pedersen JM (eds) A Synthesis of Pollutant Inputs to Buzzards Bay. Buzzards Bay Project Technical Report Series BBP94-30, 18 October 1994. Marion, MA, USA.

which the predominant source is food. Another key aspect of the biogeochemical cycle is sorption of DDT and PCBs onto particles and deposition to sediments. During inadvertent or deliberate discharges or releases to the environment, a portion of these compounds move through coastal ecosystems with portions lost to the atmosphere and transported elsewhere and to be deposited by dust or aerosols, and by rain and snow. Even though the chlorinated hydrocarbons are among the chemicals more resistant to chemical or biological alteration in the environment, there are physical–chemical (e.g., sorption–desorption, transfer from water to air), microbial transformation and degradation, and animal enzyme modifications or transformations, that change the mixture of compounds as the chemicals move through the environment. For example, the mixture of chlorobiphenyl congeners found in a lobster were dramatically different from the original mixture discharged in a


558


Land runoff

Vapor exchange

Input from air (vapor and particles) Particles

Effluents Dissolved

Microbial transformation and degradation

Pelagic fauna Vertical-horizontal advection / mixing

Fecal matter

Fecal matter

Particle Pelagic fauna

Dissolved

Metabolism, excretion, incorporation into fecal matter

Colloids

Resuspension/ sedimentation

Colloids

Benthic fauna

Diffusion Bioturbation

Pore water Colloids

Burial

Figure 4 Biogeochemical cycle for chlorinated hydrocarbon pesticides and PCBs in coastal ecosystems.

2000 1990 1980

Year of sediment deposition

nearby effluent from a capacitor manufacturing facility. Those chlorobiphenyl congener mixtures were also different in composition compared to the PCB congener mixture in flounder caught in the same area. The sediment congener mixture for the habitat of both the lobster and flounder had yet another composition. Biological effects of chlorinated hydrocarbons can be specific to each individual chemical in terms of mode of action and potency of action. Thus, the presence of complex and diverse mixtures of these chemicals in various compartments of an ecosystem introduces significant complications to the task of providing an assessment of ecological and human health risks associated with a given site of chlorinated hydrocarbon contamination. In addition, the presence of other chemicals of environmental concern in many of the same areas, means that the present knowledge base is insufficient to provide a high degree of accuracy to a quantitative risk assessment for ecological and human health concerns. In several coastal areas near urban harbors or major industrial production or use areas, sediments accumulated high concentrations and substantive amounts of DDT or PCBs. Once discharges were reduced or eliminated with curtailment of production and use of DDT and PCBs, the accumulations of these compounds in sediments continued to be of concern as a source of contamination for coastal ecosystems. The DDT and PCB contaminated sediments can leak REB and DDT or PCBs to the

1970 1960 1950 1940 1930 1920 1910 1900 0

20

40

60

80

100

120

Sum of DDT compounds _ (ng g 1 dry weight) Figure 5 Depth profile of DDT concentrations in San Pedro Shelf sediments, Southern California, USA, documenting the historical input of DDT to the area. (Adapted from Eganhouse and Kaplan, 1988.)



overlying water column or DDJ can be taken up from sediments and water in the spaces or pores between sediment particles (pore waters) by animals living in or on the sediments (Figure 4). Much of the present scientific effort related to DDT and PCBs is focused towards three broad issues: (1) in support of remediation and clean-up of areas of high levels of concentrations as a result of past practices; (2) preventing or limiting mistakes made in developed countries from occurring in developing countries; and (3) tracking the spatial and temporal trends of concentrations of these compounds in marine ecosystems, especially coastal ecosystems.

559

There are two principal approaches available to track the trend in concentrations over time. One approach is to find areas where coastal sediments are accumulating at a steady and sufficient rate, and are reasonably undisturbed by activities such as mixing of the upper layers by organisms or storm turbulence. Sediments deposited in waters with very low or no oxygen content have limited or no mixing by organisms and usually meet the criteria. Cores of sediments can be carefully sliced or sectioned at fine intervals and analyzed to provide a historical record, layer by layer, of DDT and PCB concentrations. This has been accomplished with success in several coastal

Figure 6 Concentrations of (A) DDT and (B) PCBs in Mussel Watch Stations US NOAA Status and Trends Program (Mussels and Oysters) 1986. (Adapted from A Summary of Data on Tissue Contamination from the First Three Years (1986–1988) of the Mussel Watch Project. NOAA Technical Memorandum NOS OMA 49. NOAA Office of Oceanography and Marine Assessment, Ocean Assessments Division, Rockville, MD US, 1989.)


560


areas, one example being the San Pedro Shelf, Southern California, US coastal area (Figure 5). The other approach that incorporates geographic or spatial assessments with the time series measurements has been incorporated into a monitoring strategy for assessing and monitoring concentrations of several chemicals of environmental concern. This involves the use of bivalves, mainly mussels and oysters, as sentinels of biologically available contaminants such as DDT and PCBs: the ‘Mussel Watch’ approach. Prototypes of such a program were evaluated in the 1970s in the US, Canada, Europe, and Australia and there are currently several operational programs such as the Mussel Watch component of the US National Oceanographic and

Atmospheric Administration (NOAA) National Status and Trends Program. DDT and PCB concentrations from samples of mussels and oysters obtained between 1986 and 1988 at over 150 stations located around the US coast are summarized in Figure 6. The higher concentrations of both DDT and PCBs correspond to known or suspected sources of inputs from industrial facilities making or using these chemicals, or are near urban areas. Generally it is accepted that concentrations of DDT began to decrease dramatically in portions of ecosystems for some areas of the US coast as a result of curtailed input. Concentrations of PCBs also decreased in the few areas measured during the 1970s and early 1980s as a result of curtailed

Figure 6 Continued



Table 2 Trends in concentrations of selected chlorinated hydrocarbons 1986–95 in bivalves (mussels and oysters), US coastal areasa Chemical

Chlordane DDT Dieldrin PCBs

Number of sampling locations Increased

Decreased

No trend

1 1 1 1

81 38 32 37

104 147 153 148

PCB residue concentration _ (μg g 1 fat wt.)

a Sites of several elevated concentrations are indicated in Figure 6. Data obtained from the US Department of Commerce, National Oceanic and Atmospheric Administration World Wide Web Site, October, 2000 http://state-of-coast.noaa.gov/bulletins. Data compiled by Dr Thomas P. O’Connor, US NOAA.

approach have been carried out in the 1990s for developing countries of Central and South America and South-east Asia under the auspices of UNESCOIOC and UNEP. Other time trends of DDT and PCB concentrations have been assessed such as concentrations in cod liver oil collected from samples in the southern Baltic Sea from 1971 to 1989 (Figure 7). Consistent with the preceding discussion, DDT concentrations decrease by a factor of three to four comparing 1971–1974 with 1987–1989 and PCB concentrations decrease at a slower rate. The following summary of one aspect of the PCB and chlorinated pesticide saga illustrates the importance of understanding the global, regional, and local biogeochemical cycles of these compounds and their relationship to environmental and human health risks. PCBs and several chlorinated pesticides released to the environment in developed countries of the Northern Hemisphere enter the atmosphere from land and from surface ocean waters in the tropics, subtropics, and temperate zones. Subsequently these compounds are transported by atmospheric circulation patterns to Arctic regions, and enter Arctic ecosystems by precipitation and dry deposition. There may be several cycles of precipitation and volatilization back to the atmosphere before these compounds reach the Arctic. Contamination of the Arctic aquatic ecosystems results in the transfer of these compounds through the food web and biomagnification in marine mammals. Inuits, a native Arctic region or Northern peoples, hunt several of these marine mammals and eat their tissues. The resulting contamination of mother’s breast milk transfers these chemicals to infants. There are good reasons to be concerned that subsequent normal development of the children is impaired or slowed. This is the net result of actions of human civilization and complex environmental processes operating over decades and distances of thousands of kilometers.

20 25 15

20

PCB

15

10

10 5

DDT 5

0

DDT residue concentration _1 (μg g fat wt.)

manufacture and use. The exceptions where concentrations remained elevated were generally in ecosystems with a significant burden of DDT or PCBs in surface sediments as a result of past inputs. The trends in DDT and PCB concentrations, and two other chlorinated hydrocarbon pesticides, chlordane and dieldrin, in bivalve tissues at locations in the US coastal area for 1986 to 1995 are summarized in Table 2. The decrease noted from limited sampling for a few areas in the 1970s and early 1980s continues for some locations. For many other locations, examination of the data indicates that the concentrations are so low that general global and regional biogeochemical cycles are causing only a slow further decrease. A few stations continue to maintain elevated concentrations and for most this can be attributed to continuing contamination of the bivalves from nearby surface sediments containing high concentrations of the compounds. Similar types of ‘Mussel or Oyster Watch’ data have been collected in some European countries (e.g., France) with similar results. Prototypes of this

561

0 1971 1973 1975 1977 1979 1981

1983 1985

1987 1989

Figure 7 Time trends of DDT (J) and PCB () concentrations in cod liver oil from the southern Baltic, 1971–89. (Adapted from Kannan et al., 1992.)


562


Conclusions Humanity was fortunate to learn valuable lessons from early results that indicated widespread regional and global transport of chlorinated pesticides and PCBs accompanied by environmental and human health problems. Much knowledge has been gained about the biogeochemical cycles and about ecological and biological effects of these chemicals in the oceans. This knowledge has been used to guide policy and management actions in many instances. Otherwise, the deplorable situation faced by Inuits might be much more severe and widespread and natural resource populations, including oceanic species, might have been more severely impacted. Despite policy and management actions in many developed countries limiting or eliminating production and release of many of these compounds, there are still concerns about the legacy of past releases to the environment present in coastal ocean surface sediments in several locations. There are serious coastal environmental and human health concerns associated with continued uses of several of these chlorinated hydrocarbons in developing countries.

See also Crustacean Fisheries. Large Marine Ecosystems. Molluskan Fisheries. Network Analysis of Food Webs.

Further Reading Dawe CJ and Stegeman JJ (eds.) (1991) Symposium on Chemically Contaminated Aquatic Food Resources and Human Cancer Risk. Environmental Health Perspectives 90: 3–149. Eganhouse RP and Kaplan IR (1988) Depositional history of recent sediments from San Pedro Shelf, California: reconstruction using elemental, isotopic composition, and molecular markers. Marine Chemistry 24: 163--191. Erickson MD (1997) Analytical Chemistry of PCBs, 2nd edn. New York: Lewis Publishers.

Farrington JW (1991) Biogeochemical processes governing exposure and uptake of organic pollutant compounds in aquatic organisms. Environmental Health Perspectives 90: 75--84. Fowler S (1990) Critical review of selected heavy metal and chlorinated hydrocarbon concentrations in the marine environment. Marine Environmental Research 29: 1--64. Giesy J and Kannan K (1998) Dioxin-like and non dioxinlike toxic effects of polychlorinated biphenyls (PCBs): implications for risk assessment. Critical Reviews in Toxicology 28(6): 511--569. Goldberg ED (1991) Halogenated hydrocarbons, past, present and near-future problems. Science of the Total Environment 100: 17--28. Gustafsson O, Gschwend PM, and Buesseler KO (1997) Settling removal rates of PCBs into the northwestern Atlantic derived from 238U-234Th disequilibria. Environmental Science and Technology 31: 3544--3550. Iwata H, Tanabe S, Sakai N, and Tatsukawa R (1993) Distribution of persistent organochlorines in the oceanic air and surface seawater and the role of the ocean on their global transport and fate. Environmental Science and Technology 27: 1080--1098. Kannan K, Falandysz J, Yamashita N, Tanabe S, and Tatsakawa R (1992) Temporal trends of organochlorine concentrations in cod-liver oil from the southern Baltic proper, 1971–1989. Marine Pollution Bulletin 24: 358--363. O’Connor TP (1991) Concentrations of organic contaminants in mollusks and sediments at NOAA National Status and Trends sites in the coastal and estuarine United States. Environmental Health Perspectives 90: 69--73. Petrick G, Schulz-Bull DE, Martens V, Scholz K, and Duinker JC (1996) An in-situ filtration/extraction system for the recovery of trace organics in solution and on particles tested in deep ocean water. Marine Chemistry 54: 97--103. Schwarzenbach RP, Gschwend PM, and Imboden DM (1993) Environmental Organic Chemistry. New York: John Wiley. Thornton J (2000) Pandora’s Poison. Chlorine, Health and a New Environmental Strategy. Cambridge, MA: MIT Press. Waid JS (ed.) (1986) PCBs and the Environment, vol. 1–3. Boca Raton: CRC Press.


CLAY MINERALOGY H. Chamley, Universite´ de Lille 1, Villeneuve d’Ascq, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 462–471, & 2001, Elsevier Ltd.

Introduction Clay constitutes the most abundant and ubiquitous component of the main types of marine sediments deposited from outer shelf to deep sea environments. The clay minerals are conventionally comprised of the o2 mm fraction, are sheet- or fiber-shaped, and adsorb various proportions of water. This determines a high buoyancy and the ability for clay to be widely dispersed by marine currents, despite its propensity for forming aggregates and flocs. Clay minerals in the marine environments are dominated by illite, smectite, and kaolinite, three families whose chemical composition and crystalline status are highly variable. The marine clay associations may include various amounts and types of other species, namely chlorite and random mixed layers, but also vermiculite, palygorskite, sepiolite, talc, pyrophyllite, etc. The clay mineralogy of marine sediments is therefore very diverse according to depositional environments, from both qualitative and quantitative points of view. As clay minerals are considered to be dependent on chemically concentrated environments, and as they commonly form in surficial conditions on land especially through weathering and soil-forming processes, their detrital versus authigenic origin in marine sediments has been widely debated. The transition from continental fresh to marine saline water, marked by a rapid increase of dissolved chemical elements, was the central point of discussion and arose from both American and European examples. In fact the mineralogical changes recorded at the land-to-sea transition are either important or insignificant, are characterized in estuarine sediments by various, sometimes opposite trends impeding consistent geochemical explanations, and often vanish in open marine sediments. The changes observed at the fresh-to-saline water transition in the clay mineral composition essentially proceed from differential settling processes or from mixing between different sources, and not from chemical exchanges affecting the crystalline network. Such a historical debate underlines the interest in investigating the

sensitive clay mineral associations for understanding and reconstructing environmental conditions. This article will consider the general distribution and significance of clay minerals in recent sediments, some depositional and genetic environments, and a few examples of the use of clay assemblages to reconstruct paleoclimatic and other paleoenvironmental changes.

General Distribution and Significance As a result of extensive reviews made by both American and Russian research teams the general characters of the clay mineral distribution in deep sea sediments have been known since the late 1970s. The maps published by various authors demonstrate the dominant control of terrigenous sources, which comprise either soils and paleosoils or rocks. The impact of soils on the marine clay sedimentation is largely dependent on weathering intensity developing on land, and therefore on the climate. For instance, kaolinite mostly forms under intense warm, humid conditions characterizing the intertropical regions, and prevails in the clay fraction of corresponding marine sediments. By contrast chlorite and illite chiefly derive from physical weathering of crystalline and diagenetic sedimentary rocks outcropping widely in cold regions, and therefore occur abundantly in high latitude oceans. The kaolinite/chlorite ratio in marine sediments constitutes a reliable indicator of chemical hydrolysis versus physical processes in continental weathering profiles and therefore of climatic variations occurring on the land masses. Other clay minerals are also able to bear a clear climatic message, as for instance the amount of random mixed layers and altered smectite in temperate regions, the crystalline status of illite in temperate to warm regions, and the abundance of soilforming Al-Fe smectite in subarid regions. Detailed measurements on X-ray diffraction diagrams, electron microscope observations and geochemical analyses allow precise characterization of the different continental climatic environments from data obtained on detrital sedimentary clays. Some terrigenous clay minerals in recent sediments reflect both climatic and non-climatic influences. For instance, the distribution of illite (Figure 1), a mineral that primarily derives from the erosion of mica-bearing rocks, shows increased percentages in high latitude oceans due to predominant physical


563

564

CLAY MINERALOGY

80

?

?

?

60 40

20 ?

40°

80°

120°

160°E

160°W

120°

80°

0°

40°

0 20

40 0

60

2000 km 80

< 20

20 _ 30

30_40

40 _ 50

>50%

Figure 1 Worldwide distribution of illite in the clay fraction of surface sediments in the ocean. (After Windom, 1976. Reproduced with permission from Chamley, 1989.)

weathering, but also in a few low latitude regions depending on active erosion of tectonically rejuvenated, high altitude domains (e.g., supply by Indus and Ganges river drainage systems of Himalayan material to the northern Indian Ocean). The abundance of illite in the Atlantic Ocean, especially in its high latitude and northern parts, is due to several converging causes: cold to temperate climate, extensive outcrops of crystalline and metamorphic rocks, active erosion and river input, relative narrowness of the ocean favoring the ubiquitous transportation of the mineral particles, etc. Abundant illite percentages centered on the 301 parallel of latitude in the North Pacific result from aeolian supply by high altitude jet streams blowing from eastern Asia, and subsequent rainfall above the ocean. The general distribution of illite in marine sediments therefore proceeds from direct and indirect climatic control, meteorological conditions, petrographic and tectonic characteristics, physiography, river influx, etc. All clay minerals may potentially be reworked from continental outcrops and transported over long distances until they settle on the ocean bottom. This is the case for nearly all geochemical types of smectite minerals (except perhaps for some very unstable ferriferous varieties formed in dense saline brines of the Red Sea), and also of palygorskite and sepiolite, two fibrous species wrongly suspected to not undergo significant transport. For instance, palygorskite and sepiolite are widely transported by wind and or water and deposited as detrital aggregates around the Tertiary basins bordering Africa and Arabia, where they initially formed under arid and evaporative conditions.

The clay mineral family whose distribution is the most complex and dependent on various detrital and autochthonous processes is the smectite group. Moderately crystalline smectites of diverse chemical types form pedogenically by chemical weathering under temperate conditions (essentially by degradation of illite and chlorite), and are supplied by erosion to sediments of mid-latitude regions where they are associated with various types and amounts of random mixed layers. Climate is also the dominant factor in warm, subarid regions where Al-Fe smectite forms in vertisolic soils and is reworked towards the ocean. Fairly high percentages of Fesmectite characterize the low latitude eastern Pacific basins, where clay minerals in the clay-sized fraction are accessory relative to Fe and Mn oxides, and result from in situ hydrogenous genesis. In addition, smectites of Fe, Mg, and even Al types may form by alteration of volcanic rocks, a process which is more intense in well drained, subaerial conditions (hydrolysis) than in submarine environments (halmyrolysis). The diversity of the factors controlling the distribution of clay minerals in modern deep sea sediments is widely used to trace the influence of continental climate, geological and petrographic sources, tectonics, morphological barriers, etc., and also to identify the nature, direction and intensity of transportation agents. As an example, the distribution of smectite and illite in the western Indian Ocean depends on different source provinces as well as on land geology, climate, volcanism, aeolian and marine currents (Figure 2). The terrigenous sources and climatic conditions relieved by north-to-south or southto-north surface to deep currents are responsible for


CLAY MINERALOGY

Zambesi

50 _ 40 30 _ 20

20 _ 10

10%

10 20

30

50 10

30

30 20

Antarctic

30 Province

50

60 _ 50 40 _ 30

0°

20

70 Province

60

70 _ 60

70

50 40

Ganges prov

pr

50

70

20°N

40

30

50 Central African Province

20

Indus prov

rov

n ia ab Ar

-p can Dec

20

30 40

565

In situ Province Smectite 20°

Illite 40°

60°

80°

100°E

20°

60°S 40°

60°

80°

100°E

Figure 2 Distribution of smectite and illite in the western Indian Ocean, and related source provinces. (After Kolla et al., 1976. Reproduced with permission from Chamley, 1989.)

long-distance transportation of Antarctic-derived smectite in the Crozet and Madagascar basins, of abundant volcanogenic smectite derived from Deccan traps erosion off the Indian coasts, of Himalayan illite in the Indus and Ganges deep sea fans, of illite associated with up to 30% palygorskite off Arabian and especially on submarine ridges (i.e., aeolian supply), and of illite associated with soil-derived kaolinite off Southeastern Africa. Both illite and smectite are dominantly inherited from various terrestrial rocks and soils, including Antarctic outcrops responsible for illite dominance to the west of the Indian Ocean (351C) and for smectite dominance to the east (45–751E). An in situ smectite-rich province located in the southern ocean around 551S and 701E is attributed to the submarine alteration of volcanic rocks. Volcanic contributions are also suspected in the Central Indian basin and in the vicinity of Indonesia. Of course such investigations constitute very useful guidelines for reconstructing past climatic, oceanographic, and physiographic conditions.

Marine Autochthonous Processes From Volcanic to Hydrothermal and Hydrogenous Environments

Until the 1970s, the submarine weathering of volcanic material (basalt, glass, ash) was often considered to be responsible for important in situ

formation of clay minerals, especially of smectite, in deep sea sediments. Effectively basalt altered by surficial oxidation and hydration may give way to Mg-smectite, sometimes Fe-smectite, frequently associated with celadonite (a glauconite-like Fe-Al micaceous species), phillipsite (a Na-rich zeolite), calcium carbonates, Fe-Mn oxyhydroxides, etc. The more amorphous, the smaller sized and the more porous the volcanic material (e.g., pumiceous ashes), the more intense the submarine formation of clay. In fact the clay minerals resulting from halmyrolysis of volcanic material are quantitatively limited and essentially located at close vicinity to this material (e.g., altered volcaniclastites or basalts); they are unable to participate in a large way in the formation of the huge amounts of clay incorporated in deep-sea sediments. Additional arguments contradicting the importance of volcanic contribution to deep-sea clay consist of the frequent absence of correlation between the presence of volcanic remains and that of smectite, and in the non-volcanogenic chemistry of most marine smectites (e.g., aluminum content, rare earth elements, strontium isotopes). The shape of smectite particles observed by electron microscopy is typical of volcanic influence only in restricted regions marked by high volcanic activity, especially explosive activity. Notice that local overgrowths of lath systems oriented at 601 from each other may characterize marine clay particles and especially smectites, but they are neither specifically related to volcanic


566

CLAY MINERALOGY

environments nor associated with noticeable increase of smectite proportion or specific change in the clay chemical or isotopic composition. The intrusion of basalt sills in soft marine sediments may determine some metamorphic effects and the very local formation of ordered mixed layers (corrensite), chlorite, and associated non-clay minerals. The hydrothermal impact on deep-sea sedimentation is fundamentally characterized by in situ precipitation of Fe-Mn oxyhydroxides relatively depleted in accessory transition elements (Co, Cu, Ni), and locally by the deposition of massive sulfides near the vents where hot and chemically concentrated water merges. The autochthonous clay minerals in such environments are marked by various species depending on fluid temperature, oxidation-reduction processes, and fluid/rock ratio. For instance, drilling holes in Pacific hydrothermal systems show different mineral evolutions. In the hydrothermal mounds of the Galapagos spreading center, the fluids are rich in silicon and iron and of a low temperature (201–301C) throughout the 30 m-thick sedimentary column; this gives way in oxidized conditions to the precipitation of Fe-smectite as greenish layers interbedded in

biogenic oozes that at depth evolve into glauconite by addition of potassium (Figure 3A). By contrast the detrital to authigenic deposits of the Middle Valley of Juan de Fuca ridge show on a 40 m-thick series the in situ formation from high temperature Mg-rich fluids (2001C) of a downwards sequence characterized by saponite (a Mg-smectite), corrensite (a regular chlorite-smectite mixed layer), swelling chlorite, and chlorite (Figure 3B). At this site geochemical and isotope investigations reflect a noticeable downhole increase of temperature and strong changes in the fluid composition. A more widespread process consists of the hydrogenous formation of clay at the sediment– seawater interface, in deep-sea environments characterized by water depths 44000 m, insignificant terrigenous supply, and very low sedimentation rate (o1mm/1000 years). This is particularly the case for some Central and South Pacific basins. The sediments mostly consist of reddish-brown oozes rich in Fe and Mn oxides (i.e., ‘deep sea red clay’). There iron-rich smectites of the nontronite group may form in significant proportions, probably due to long-term low temperature interactions between (1) metal Juan de Fuca Middle Valley

Galapagos spreading center DSDP 509B

ODP 509B Saponite

Fe-smectite Corrensite Fe-smectite, glauconite

Fe-smectite + K Glauconite +Fe oxides

Fe-smectite, glauconite

Corrensite, swelling chlorite

Glauconite

40 m

30 m

Seawater influx

Swelling chlorite

Chlorite

20 _ 30°C

(A)

Hydrothermal sediments

200°C

(B)

Figure 3 Schematic vertical distribution of typically hydrothermal clay minerals in the sedimentary systems of (A) the Galapagos spreading center, and (B) the Juan de Fuca Middle Valley. (Reproduced with permission from Buatier and Karpoff, 1995.)


CLAY MINERALOGY

Table 1

567

Examples of chemical composition of hydrothermal to hydrogenous smectites in Central and South Pacific sediments

Type of smectite

Tetrahedra

Pure hydrothermal (Galapagos mounds field) Hydrothermal and hydrogenous (Bauer Deep) Hydrogenous>hydrothermal (Galapagos spreading centre) Pure hydrogenous (?) (North Marquises fracture zone)

Octahedra

Interlayers

Si

Al

Fe

Al

Fe

Mg

Ca

NH4

K

3.94 3.97 3.97 3.37

0.06 0.03 0.03 0.63

– – – –

0.03 0.44 1.12 0.39

1.59 1.07 0.48 1.12

0.38 0.54 0.37 0.46

0.03 0.05 0.09 0.46

– – – –

0.36 0.06 0.11 0.17

(Reproduced with permission from Chamley, 1989.)

oxyhydroxides supplying the iron, (2) sea-water supplying the magnesium and other minor to trace elements, (3) biogenic silica supplying the silicon, and (4) allochthonous accessory particles (e.g., aeolian clay) supplying the other chemical elements (e.g., Al). Notice that the distinction between pure hydrothermal and pure hydrogenous clay minerals forming on the deep-sea floor necessitates detailed chemical analyses (Table 1) and often additional microprobe and isotope investigations. To summarize, the distribution of clay minerals in deep sea deposits marked by active volcanic-hydrothermal activity and by very low sedimentation rates depends on various and complex in situ influences among which the hydrogenous processes quantitatively prevail. The distinction of these autochthonous influences is complicated both in the vicinity of land masses where terrigenous supply becomes active, and in shallower areas where biogenic influences may intervene more intensely (e.g., Nazca plate, southeast Pacific). Ferriferous Clay Granules

Iron-rich clay granules are traditionally called glauconite, which is somewhat incorrect as glauconite is a specific clay mineral, whereas clay granules may include various iron-bearing clay species. Ferriferous clay granules form on continental margins at water depths not exceeding 1000 m, and comprise two major types characterized by specific colors, clay minerals, and habits. Glaucony, the most widespread type, constitutes dark green to brown clayey aggregates, and may comprise different varieties of ironrich illite- and smectite-like minerals such as glauconite (Fe- and K-rich illitic clay), Fe-smectite, and Fe illite-smectite mixed layers. Glaucony may form at latitudes as high as 501 and in water depths as great as 1000 m, but usually occurs in 150–300 m water depths at the shelf-slope transition of temperate-warm to equatorial regions. Verdine, which is less ubiquitous and has been identified more recently, constitutes light green to light brown granules

characterized by phyllite V or odinite, a ferriferous clay mineral of the kaolinite family (described by G.S. Odin, who has developed outstanding investigations on clay granules). Verdine forms in rather shallow water sediments (maximum 50–80 m) of intertropical regions, and depends on the supply of abundant dissolved iron by low latitude rivers. Ferriferous clay granules form at the sediment– water interface and evolve at burial depths rarely exceeding a few decimeters. They develop in semiconfined environments at the expense of various substrates submitted to ‘greening’: chiefly fecal pellets and microfossil chambers (e.g., foraminifera), calcareous or siliceous bioclasts, minerals (especially micas), and rock debris. The formation of glaucony (which somewhat leads to diffuse habits), occurs in successive stages marked by a rapid and strong enrichment of iron and then potassium, a volume increase causing external cracks, and the obliteration of the initial shape (Figure 4). The formation of verdine still has to be documented, but both clay granule types correspond to true authigenic formation rather than to transformation of pre-existing clay minerals. The chemical evolution of ferriferous clay granules vanishes either after a long exposure at the sediment–water interface (105–106 years for glaucony), or after significant burying.

Organic Environments

The influence of living organisms on clay-rich sediments is mainly marked by physical processes referred to as bioturbation, and concerns various marine environments, especially on continental shelves. Chemical modifications of clay associations are only occasionally reported and seem to affect the crystalline status of chlorite and associated random mixed layer clays locally through ingestion and digestion processes of shallow water crustaceans, annelids or copepods. The chemical interactions developing in digestive tracts between clay minerals and organic acids appear to have small quantitative


568

CLAY MINERALOGY

Seawater

Nascent

Slightly evolved

3

10

10

3

4

Authigenic globules

Detrital fraction

Sediment

Authigenic flakes, rosettes

Evolved

4

Highly evolved

5

Years

10

5

7

6

8

9

K2O%

Figure 4 Successive stages of glaucony formation from a pre-existing substrate. (Reproduced with permission from Odin, 1998.)

effects, as the marine clay associations are roughly the same as the terrestrial associations. The chemical impact on clay mineral stability of the organic matter incorporated in deep marine sediments is variable. Most sedimentary series containing significant amounts of dispersed organic matter (i.e., 1–3%) do not display any specific clay mineral composition. For example, this is the case for black shales deposited during the Cretaceous period in the Atlantic, where clay mineral

associations may comprise vulnerable species such as smectite and palygorskite, the abundance and crystalline status of which vary independently of the content and distribution of the organic matter. In contrast, the sapropels developing in the eastern Mediterranean during the late Cenozoic era, especially in Quaternary high sea level stages, show some in situ degradation processes of the detrital clay minerals (Figure 5). Submarine alteration affects the mineral species in successive stages depending on

Increasing submarine alteration

Increasing resistance

Palygorskite

Mixed layers Degradation

Smectite Degradation Complexation?

Chlorite Degradation

Illite

Kaolinite High topographic zones Periphery of sapropelic areas Top of sapropels

Low topographic zones Center of sapropelic areas Base of sapropels

Figure 5 Characters of the clay mineral degradation in Quaternary sapropels of the eastern Mediterranean basins.


CLAY MINERALOGY

their ability to resist acid conditions. Palygorskite is the more vulnerable species and kaolinite the more resistant. The degradation of clay assemblages tends to increase toward the central and deepest parts of marine basins, in depressed morphological zones, and at the base of the decimeter-to-meter thick sapropels. The degradation of clay minerals under organic conditions has occurred close to the sediment–water interface and appears to depend on the chemical nature and evolution stage of the terrestrial and marine organic matter.

Paleoenvironmental Expression Clay mineral assemblages of sediments successively deposited in marine basins express various environmental messages related to the geological history. A few examples from recent Quaternary to late Cenozoic series will be considered here. Similar messages may be preserved in much older series of Mesozoic and even Paleozoic ages, provided that the diagenetic imprint due to lithostatic overburden, geothermal gradient, and fluid circulation has remained moderate. Clay-rich, low permeability sedimentary formations 2–3 km thick and submitted to normal heat flow (c. 301C/km) are usually prone to preserve such paleoenvironmental characteristics. Climate

As clay minerals at the surface of the Earth are dominantly formed through pedogenic processes depending on climate and are particularly subjected to surficial erosion and reworking, their assemblages successively deposited in a given sedimentary basin are a priori able to reflect successive climatic conditions that prevailed on adjacent land masses. This implies that very little post-depositional, i.e., diagenetic changes have affected the clay assemblages after their storage in sediments. This is observed to be the case in many series drilled or cored in the oceans. The climatic message borne by clay has been documented by numerous investigations, and corroborated by the comparable range of variations recorded in the nature and proportions of clay minerals in both present-day soils outcropping at various latitudes and marine sedimentary columns. Marine clay mineral assemblages basically express the type and intensity of continental weathering, which depend predominantly on the ion leaching through the action of humidity and temperature, and secondarily on seasonal rainfall and drainage conditions. Quaternary glacial–interglacial alternations caused terrestrial alternation of physical and chemical weathering processes, and this was reflected in

569

the clay assemblages successively brought to marine sediments through soil erosion and river or wind transport. Sedimentary levels contemporary with cold periods are usually characterized by more abundant rock-derived minerals such as richly crystalline illite, chlorite, smectite and associated feldspars reworked from active physical weathering. Warm, humid periods generally correspond to increased supply of soil-derived kaolinite and metal oxides, poorly crystalline smectite and various random mixed layer clay minerals. For instance, the terrigenous fraction of hemipelagic sediments deposited from 500 000 to 100 000 years ago in the Northwestern Atlantic off New Jersey and dominantly derived from the erosion of Appalachian highlands shows increased proportions of chlorite in glacial isotopic stages, and of kaolinite in interglacial stages. This is clearly expressed by the kaolinite/chlorite ratio (Figure 6). Paleoclimatic reconstructions from clay mineral data are available for various geological periods, as for instance the passage since about 40 Ma from a non-glacial world dominated by chemical weathering (smectite, kaolinite) to a glacial world in which physical weathering was greater (chlorite, illite). The comparison of climatic curves provided by clay minerals and other indicators (oxygen isotopes, micro-faunas or floras, magnetic susceptibility, etc.) allows a better understanding of the nature, intensity, and effect of the different factors characterizing the terrestrial and marine climate in given regions during given geological intervals. High resolution studies show that clay assemblages may express terrestrial climatic variations at a centennial scale or even less, and that the influence of Earth’s orbital parameters varies to different extents according to the latitude. For example, the clay minerals data of Quaternary North Atlantic deep sea sediments were submitted to cross-correlation spectral analyses on 5.5–14 m-long cores encompassing the last 300 000 years. The mineral composition displays a general 100 000-year cyclic signal (eccentricity) in the whole 451–601N range, a 41 000-year signal (obliquity) at highest latitudes related to dominant aeolian supply, and a 23 000-year signal (precession) at mid-latitudes related to dominant transport by marine currents (Table 2). The paleoclimatic expression by clay mineral successions is direct or indirect, i.e., it either indicates the climate that actually prevailed at a given period, or reflects other events depending on climate: migration of lithospheric plates across successive climatic zones, varying extension of ice caps controlling the surficial erosion, variations in the marine circulation regime due to changing latitudinal and


570

CLAY MINERALOGY

Oxygen isotope curve 18 δ O Temperate Cold 2

0

_2

Kaolinite %

Chlorite %

2.5 5.0 7.5 10.0 12.5

2.5 5.0 7.5 10.0 12.5

100

Kaolinite / chlorite ratio Increasing weathering 0.5

1.0

1.5

0

Stage 5 10 150

Stage 6

20 30

200

40

250

50 Stage 8 60

300 70

Stage 9

80

Stage 10

350

Depth (mbsf)

Age before present (ky)

Stage 7

90 400

Stage 11 100

450

110

Stage 12

120 500

130

Figure 6 Comparison and climatic significance of clay mineral and oxygen isotope data from stages 12 to 5 at ODP Site 902, New Jersey continental margin. (Reproduced with permission from Vanderaveroet et al., 1999.)

vertical heat transfers. The direct paleoclimatic reconstructions from clay mineral data are all the more reliable since the marine basins investigated are preserved from important erosion of paleosoils, changes in detrital sources, differential settling processes, longitudinal oceanic currents, and major geomorphological changes. Marine Currents

The different marine water masses may carry the small and light clay mineral particles over long distances, and therefore leave an imprint within the sediments at the depth range they are moving. This has been demonstrated for late Quaternary sediments of the southwestern Atlantic, where the southward-flowing North Atlantic deep water is enriched in kaolinite supplied from rivers draining the intertropical South American continent, and the northward flowing Antarctic Atlantic bottom water supplies chlorite and smectite issuing from southernmost Argentina and Antarctica. Paleocurrent reconstructions from clay data exist mainly about Atlantic and Southern Oceans, which are marked by numerous and distinct terrigenous sources, vertical mixing and longitudinal heat transfers, and Tertiary to Quaternary changing

conditions of the superimposed water masses volume and celerity. Tectonic Activity

The tectonic instability determines some changes in the composition of clay mineral assemblages which are usually much more important than those due to climate or circulation. First, the subpermanent rejuvenation by neotectonics of continental relief increases the erosion potential and therefore impedes the development of pedogenic blankets where clay minerals tend to be in equilibrium with current climatic conditions. Such a chronic tectonic activity explains the abundance of rock-derived illite and chlorite in equatorial Indian Ocean basins depending on Himalayan output. Second, a continental tectonic uplift determines changes in the nature of clay minerals eroded from rocky substrates, while submarine uplift may determine morphological barriers to the clay transfer. This was the case for the Hellenic Trench in the eastern Mediterranean during late Pliocene to early Pleistocene periods, when the combined uplift of Peloponnese and of Mediterranean ridge both increased the terrigenous input of European illite and chlorite and blocked the supply of African palygorskite. Due to their sensitivity to


CLAY MINERALOGY

571

Table 2 General relationships between the clay mineral distribution and the three main Earth’s orbital frequency bands according to latitude, from cross-correlation spectral analysis of X-ray diffraction data on North Atlantic cores Core Latitude Orbital parameters

SU 90-08 441N E O

Illite Chlorite Kaolinite Illite-vermiculite random mixed layer

H V H –

– – – –

P

SU 90-12 511N E O

V V V –

H V H –

– – – –

P

SU 90-38 541N E O

V V – V

V V V –

H V V –

P

SU 90-33 601N E O

P

– – – –

H V H –

– – – –

V V V –

E, eccentricity band, 100000 year; O, obliquity band, 41000 year; P, precession band, 23000 year; H, high variance power; V, very high variance power. Maximum correlations in bold characters. (Reproduced with permission from Bout-Roumazeilles et al., 1997.)

geomorphological changes and their aptitude for long distance transportation, clay minerals are able to express slight and progressive epeirogenic changes as well as very remote tectonic events.

See also Aeolian Inputs. Cenozoic Climate – Oxygen Isotope Evidence. Deep-Sea Drilling Results. Hydrothermal Vent Deposits. Rare Earth Elements and their Isotopes in the Ocean. River Inputs. Water Types and Water Masses.

Further Reading Bout-Roumazeilles V, Debrabant P, Labeyrie L, Chamley H, and Cortijo E (1997) Latitudinal control of astronomical forcing parameters on the high-resolution clay mineral distribution in the 451–601 N range in the North Atlantic Ocean during the past 300,000 years. Paleoceanography 12: 671--686. Buatier MD and Karpoff AM (1995) Authigene´se et e´volution d’argiles hydrothermales oceániques:

exemples des monts des Galapagos et des se´diments de la ride de Juan de Fuca. Bulletin de la Socie´te´ Geólogique de France 166: 123--136. Chamley H (1989) Clay Sedimentology. Berlin: SpringerVerlag. Hoffert M (1980) Les ‘argiles rouges des grands fonds’ dans le Pacifique centre-est. Sciences geólogique. Strasbourg, Mem 61: 257. Millot G (1970) Geology of Clays. Berlin: Springer-Verlag. Odin GS (ed.) (1988) Green Marine Clays. Developments in Sedimentology, 45, Amsterdam: Elsevier. Robert C and Chamley H (1992) Late Eocene-early Oligocene evolution of climate and marine circulation: deep-sea clay mineral evidence. American Geophysical Union. Antarctic Research Series 56: 97--117. Vanderaveroet P, Averbuch O, Deconinck JF, and Chamley H (1999) A record of glacial/interglacial alternations in Pleistocene sediments off New Jersey expressed by clay mineral, grain-size and magnetic susceptibility data. Marine Geology 159: 79--92. Weaver CE (1999) Clays, Muds, and Shales. Developments in Sedimentology, 44. Amsterdam: Elsevier. Windom HL (1976) Lithogenous material in marine sediments. Chemical Oceanography vol. 5, pp. 103--135. New York: Academic Press.


COASTAL CIRCULATION MODELS F. E. Werner and B. O. Blanton, The University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 472–480, & 2001, Elsevier Ltd.

Introduction Coastal environments are among the most complex regions of the world’s oceans. They are the transition zone between the open ocean and terrestrial watersheds with important and disparate spatial and temporal scales occurring in the physical as well as biogeochemical processes. Coastal oceans have three major components, the estuarine and nearshore areas, the continental shelf, and the continental slope. The water column depth ranges from areas where flooding and drying of topography occurs over a tidal cycle at the landward boundary, to depths of thousands of meters seaward of the shelf break. The offshore extent of coastal environments can range from a few kilometers (off the Peru/Chile coast), to hundreds of kilometers (over the European or Patagonian shelf). The topography in coastal oceans can be relatively featureless, or it can be complex and include river deltas, canyons, submerged banks, and sand ridges. Coastlines and coastal oceans span the globe, from near the North Pole to the Antarctic, and thus are subject to a full range of climatic conditions. Circulation in coastal regions is forced locally (for example by winds, freshwater discharges, formation of ice) or remotely (through interactions with the neighboring deep ocean, terrestrial watersheds, or large-scale atmospheric disturbances). Resulting motions include tides, waves, mean currents, jets, plumes, eddies, fronts, instabilities, and mixing events. Vertical and horizontal spatial scales of motion range from centimeters to hundreds of kilometers, and timescales range from seconds to interannual and longer. Coastal regions can be very productive biologically and they support the world’s largest fisheries. These regions are also preferred as recreational and dwelling sites for our increasing human population. There is evidence suggesting that changes in the coastal environment, such as degradation in habitat, water quality, as well as changes in the structure and abundance of fisheries have resulted from increases in commercial and residential development,

572

agriculture, livestock, soil, and sediment loss. Therefore, although natural phenomena shaped the coastal environment in the past, in the future it will be defined jointly by natural and anthropogenic processes. To understand the coastal oceans, predict their future states, and reduce the human impact on the region through management strategies, it is necessary to develop a quantitative understanding of the processes that define the state of the coastal ocean. Coastal circulation models are tools rooted in mathematical and computational science formalism that allow the integration of measurements, theory and computational capability in our attempt to quantify the above processes.

Governing Equations and State Variables The starting point for coastal circulation models is modified versions of the Navier–Stokes equations derived for the study of classical fluid mechanics. The fundamental differences are the inclusion of the Coriolis force associated with the Earth’s rotation, and the inclusion of hydrostatic and Boussinesq approximations appropriate for a thin layer of stratified fluid on a sphere for circulation features of hundreds of meters and larger. Certain smaller-scale motions, such as convection and mixing, are not admitted by these approximations as they may be possibly nonhydrostatic. Additional departures from descriptions of other fluid motions are the consideration of temperature and salinity as thermodynamic variables, and a nonlinear equation of state. Coastal ocean domains are subsets of the global ocean basins and are typically defined by a solid wall (landward) boundary and open (wet) boundaries which connect the region of interest to neighboring bodies of water. Islands inside the model domains are also considered as solid wall boundaries. The open boundaries are generally of two types: offshore boundaries along which the coastal domain exchanges information with the neighboring deep ocean, and cross-shelf boundaries along which the coastal domain receives and/or radiates information to regions up- or downstream of the study site. The sea surface and the bottom complete the definition of the model domain. With the model domain defined, solution of the governing equations is sought subject to specified initial and boundary conditions.


COASTAL CIRCULATION MODELS

The simplest initial conditions specify the fluid to be at rest (no motion), and the sea level, temperature and salinity fields to be flat (no horizontal gradients). Boundary conditions are more problematic and must be specified on all model boundaries. They include stress conditions at the free surface where atmospheric winds are imposed and input energy into the coastal ocean, and at the bottom where frictional forces extract energy from the overlying motions. Heating and cooling are imposed through prescribed flux conditions at the surface (where the coastal ocean is in contact with the atmosphere), or along the model’s lateral boundaries (where it is in contact with offshore regions). Additional buoyancy fluxes due to variations in salinity are imposed through prescribed evaporation or precipitation fluxes at the surface, along the open boundaries as a result of exchanges with the offshore and up- and downstream regions, and from either point- or line-sources representing riverine or larger watershed (terrestrial) inputs. The proper specification of boundary conditions is one of the more difficult aspects of modeling coastal circulation. Although for most cases the mathematical approaches are well established, the data required to quantitatively specify the mass and momentum fluxes across the model boundaries are lacking. As a result, boundary conditions are generally idealized. In practice boundary conditions are a mixture of imposed observed quantities, derived values from larger-domain models, and conditions that minimize the uncertainties associated with the artificial nature of open (wet) boundaries. The solution of the governing equations consists of the time-history in three-dimensions of the velocity field, the temperature, salinity, and density, and additional derived quantities describing mixing rates of mass, momentum, and other tracers. Analytic solutions, also known as closed form solutions, are not possible except for highly idealized cases. For example, topography and forcing must be simplified and certain nonlinear processes must be ignored. Nevertheless, even in these limiting cases, analytic approaches are desirable as they include in a single statement the solutions’ dependence on a wide range of parameters over the entire model domain, allowing for a comprehensive understanding of the interaction between the fundamental processes. In the early 1980s, analytic solutions developed for coastal-trapped waves presented a breakthrough in the study of remotely forced currents in coastal regions. Numerical approaches offer the possibility of retaining full dynamic and topographic complexity in the study of coastal circulation. In these approaches,

573

the governing equations are discretized in space and time and the resulting algebraic discrete equations are solved using methods of numerical analysis. Spatial discretization in the horizontal is accomplished using the finite difference method with either structured regular grids, or, in cases where the shape of the coastline is highly irregular, curvilinear grids (Figure 1). The latter allow some degree of resolution and geometric flexibility. The finite element method uses unstructured grids and allows for greatest flexibility in capturing spatial heterogeneity and geometric complexity (Figure 2). Horizontal spatial discretization is important as the convergence of the models’ solution is dependent on proper refinement of topography and flow structures associated with the presence of stratification, among others. The relative merits of structured versus unstructured meshes has not been fully addressed by the research community. In the vertical, three approaches are commonly used in computing the depth-dependent structure of the circulation. The ‘z’-coordinate computes the vertical structure along constant geopotential levels, the ‘sigma’ or s-coordinate is bottom- or terrainfollowing and the solution is computed at the same number of points in the vertical regardless of the water column depth, and the isopycnal or densitycoordinates in which the vertical structure is computed along the time-dependent location of the density surfaces. As in the case of horizontal discretization approaches, there is no optimum choice of vertical coordinate systems. There are many algorithmic questions and mathematical formulations that are still not fully answered. Assuming smoothness in forcing functions, initial data, topography, etc., the choice of discretization method to the solution should not matter in the continuum limit of the equations. However, errors arising from solving the approximate forms of the governing equations display different behaviors due to discretization methods, and in some cases these solutions are spurious. Higher-order discretization schemes that reduce the truncation error although not significantly increasing the computational effort continue to be investigated. Similarly, many physical processes are not well understood, such as vertical mixing near the free surface, flow instabilities and horizontal mixing rates. The scales of these processes are frequently too small to be resolved in models and it is necessary to represent them by what is often ‘ad hoc’ parametrization. Thus, the development of coastal circulation models continues to be a specialized undertaking, with several approaches being developed by teams of investigators worldwide.


574

COASTAL CIRCULATION MODELS 50 NOAA COASTAL OCEAN FORECAST SYSTEM COFS: Horizontal grid and Bathymetry

48 46 44 42 40 38 36 34 32 30 28

Bathymetry Contours (m) 100 m intervals: 0 m to 1000 m 1000 m intervals: 1000 m to 6000 m

26 24

22 _ 84 _ 82 _ 80 _ 78 _ 76 _ 74 _ 72 _ 70 _ 68 _ 66 _ 64 _ 62 _ 60 _ 58 _ 56 _ 54 _ 52 _ 50 _ 48 _ 46 NWS/NCEP/EMC/OMB & NOS/CSDL/OP & Princeton Univ. Figure 1 National Oceanic and Atmospheric Administration’s Coastal Ocean Forecast System (COFS) curvilinear finite difference grid. Provided courtesy of National Weather Service’s Environmental Modeling Center.

54°N 52°N

Gulf of St Lawrence

50°N

Latitude

46°N

nd

undla

Newfo

48°N

Gulf of Maine

44°N

ia

cot

S va No

Flemish Cap

42°N Grand Banks

Scotian Shelf

40°N 38°N 70°W

65°W

60°W

55°W

50°W

45°W

Longitude Figure 2 Northwest Atlantic shelf finite element domain. Flexibility of method increases resolution near sharp bathymetric gradients at the shelf break, deep channels, submerged banks, and along the coast.


575


Applications Applications of coastal circulation models can be broadly classified as process studies or regional studies. Process studies seek to identify the fundamental physical mechanisms responsible for observed features of the coastal ocean by idealizing complications of irregular shoreline geometry, timedependent stochastic boundary forcing, and possibly simplifying the governing equations. Typically, these models retain effects such as the earth’s rotation, idealized stratification and topography, idealized boundary conditions of heat flux and wind stress, and simplified turbulence closure. Early studies in the 1970s and 1980s focused on understanding largescale wind-forced response of coastal regions including upwelling, and the nonlinear propagation of tides. Recently, with the increase in computing capabilities and improved mathematical formulations, the spatial and temporal resolution of process studies has also increased. The result has been a greater understanding of the detailed structure of phenomena such as the interaction of coastally trapped waves with bathymetric features and irregular coastlines (e.g., canyons, ridges, and capes), the generation of instabilities in the currents and formation of upwelling filaments, the formation of temperature and/or salinity fronts along the continental shelf break, and of river plume dynamics in the near-shore coastal and estuarine regions. Regional coastal circulation studies attempt to include as much realism as possible into the numerical simulation of a specific region, including geometry and boundary conditions. These studies include the estimation of climatological circulation and tracer (e.g., temperature and salinity) distribution, fine resolution tidal simulations, storm surge analysis and prediction, transport of dissolved and particulate matter, coastal ocean prediction and forecasting, and coupled effects between estuaries, tidal inlets and the coastal ocean. Sea Level

Many of the world’s coastal regions are affected by large variations in sea level. The ability of coastal

circulation models to accurately simulate coastal sea level has enabled the quantitative study of the impact of large tidal amplitudes and storm surges on lowlying coastal areas. Regional coastal sea-level models have become de facto components of emergency management systems in areas sensitive to sea-level variations. Robust and very good predictions of sea level can be obtained with horizontal two-dimensional models provided that accurate predictions of the surface wind field are available. The simplification from fully three-dimensional approaches is accomplished by averaging the governing equations along the vertical coordinate. Usually these models will also ignore effects of stratification, but include very high-resolution bottom topography and coastline features. The simplifications allow for significant speed-up of computations, which is necessary when issuing real-time forecasts. The rise of sea level associated with the passage of storm systems is known as storm surge. Accurate prediction of sea level during a storm surge (Figure 3) and its timing relative to the time of high tide are essential for the protection of property and life in low-lying coastal areas, such as the Dutch coast, the Gulf of Mexico, the east coast of the United States, and the southern Asian continent. Engineering

Coastal circulation models are also used in engineering applications, such as in the design of ports, offshore platforms, and in the dredging of shipping lanes. These applications usually deal with the impact of the circulation on the structure being built. However, there are instances where the structure 1.5

Modeled Observed

1.0

Water level (m)

However, the advent of significant computational capabilities (readily accessible on present-day desktop and laptop computers) has enabled coastal ocean models to become increasingly complex, and are now based on the fully stratified, nonlinear equations of motion. These advances coupled with the importance and interest in understanding coastal ocean processes has resulted in expansive growth and applications in some of the areas discussed next.

0.5 0 _ 0.5 _ 1.0 12

12.5

13

13.5

Day in July 1996

Figure 3 Observed and modeled water levels of New Bern, North Carolina, during the passage of Hurricane Bertha in July 1996. The modeled water level was computed using a twodimensional circulation model, and accurately captures the magnitude and timing of the sea level response to the storm.


576


itself can have a significant effect on the coastal ocean, and circulation models are used in the quantitative assessment of its impact. The Bay of Fundy in the Gulf of Maine is known for its extreme (over 6 meter) tidal amplitudes, a result of the region’s near-resonance with the principal lunar M2 tide with period 12.4 h. The natural resonant period of the gulf-bay system is about 13.3 h. The large amplitude tide offers the opportunity to harness the tidal elevation and resulting potential energy to generate hydroelectric power by constructing dams. In 1987, a two-dimensional circulation model of the Gulf of Maine was used to investigate the tides, the effects of building tidal power plants, and their potential impact on the natural resonant period of the gulf-bay system. These studies showed that the tidal amplitude near the proposed barriers would decrease by about 25 cm due to the shortening of the bay’s natural length and consequent decrease in the bay’s resonant period. Furthermore, and perhaps somewhat unexpectedly the results also indicated that increases in tidal amplitude of 15–20 cm would occur in remote coastal areas, some of which are potentially sensitive to sea-level fluctuations and flooding. Predicted changes in circulation also suggested changes in sedimentation rates.

Coastal Ecosystems

The study of marine ecosystems requires that models of different systems be coupled to properly capture biological, geochemical, and hydrodynamic interactions across a wide range of temporal and spatial scales. Important biological processes are affected by transport mechanisms that can occur over hundreds of kilometers as well as turbulent mixing events that can occur on scales of several meters or less. Coastal circulation models have now achieved a level of sophistication and realism where new and significant opportunities for scientific progress in studying coupled physical–biological simulations are within reach. We are close to being able to construct spatially and temporally explicit models of the coastal physical environment, including the specification of velocity, hydrography, and turbulent fields, on scales relevant to biological processes. The investigation of ecosystem-level questions involving the role of hydrodynamics in determining the variability and regulation of planktonic and fish populations (Figures 4 and 5) are now being attempted. There are now many case studies that have coupled the growth and feeding environment of planktonic and larval fish species with coastal circulation models. Examples include: the study of retention, survival, and

200 200

North

42°

100 41° 60 60 _1

25 cm s

40° 200 70°

69°

68°

67°

66°

West Figure 4 Modeled depth and time-averaged circulation on Georges Bank forced by tides, winds and upstream inflow. Adapted from Werner et al (1996).



Day 0 _ Spawn

Day 20_ Hatch

Day 40

Day 60

577

North

42˚

41˚

40˚

North

42˚

41˚

40˚

70˚

69˚

68˚ West

67˚

66˚

70˚

69˚

68˚ West

67˚

66˚

Figure 5 Simulated larval fish trajectories over Georges Bank at 20, 40, and 60 days post-spawn using flow fields from Figure 4. These trajectories are used to evaluate the on-bank retention versus off-bank loss. Adapted from Werner et al. (1996).

dispersal of larval cod, haddock and their prey in the Northwest Atlantic and North Sea; the transport of estuarine dependent fish from offshore coastal spawning regions to estuarine nursery habitats on the eastern US coast; the interannual recruitment variability of pollock in the Gulf of Alaska; and the dispersal of coral reef species. The development of management strategies used in the definition of marine sanctuaries or marine reserves can now look to circulation models for guidance in estimating population exchanges within and among neighboring coastal regions. Operational Forecast Systems

A recent and evolving application of coastal circulation models is in operational coastal ocean prediction and forecasting systems. These systems are used to estimate in real-time the state of a particular region of the coastal ocean for the purposes of

navigation, naval operations, search-and-rescue, oil spill impact assessment, or commercial and recreational fishing. The coastal circulation model is driven in part by forecasts of heat,moisture, and momentum from weather, tidal or large-scale ocean circulation models. However, to partially correct for erroneous (or imperfectly known) boundary and initial conditions and the resulting continuous accumulations of these errors, algorithms have been developed that allow assimilation of observed currents, water level, and/or hydrography within the domain and thereby improve model forecasts in real-time. The US National Oceanic and Atmospheric Administration’s Coastal Ocean Forecast System (COFS) provides real-time forecasts of the coastal and open ocean state for the eastern US coast (Figure 6) by taking advantage of recent advances in coastal circulation models and observational systems. The coastal circulation model is forced at the surface by forecast surface flux fields of momentum, heat, and


578


50

NOAA COASTAL OCEAN FORECAST SYSTEM (Ver. 3.4) 24_forecasy of top model layer water temperature valid: 0000 UTC November 30, 2000

45

40

35

30 Temperature (C) _1 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 25 COFS 3.4 : Forecast cycle :Sfc forcing from Eta_29 forecasts _ 85

_ 80 _ 75 _ 70 Assimilation of SST + TOREX + RS2 (agi 100)

_ 65

_ 60

_ 55

_ 50

_ 45

NWS/NGEP/EMC/OME & NOS/CSDL/OF & Prinston Univ

20

Figure 6 ECOFS 24-hour sea surface temperature forecast for 30 November 2000, computed on the grid in Figure 1. Note that a large portion of the deep water adjacent to the east coast shelf needs to be included in such a forecasting system. Provided courtesy of National Weather Service’s Environmental Modeling Center.

moisture from a high-resolution weather forecast model. An assimilation system that incorporates both in situ and remotely sensed observations of surface and sysubsurface temperatures and seasurface heights enables ECOFS to make relatively accurate 24-hour forecasts of Gulf Stream frontal position, water levels, three-dimensional currents, temperature, and salinity on a daily basis. Sampling Design

Intense observational efforts focus on sampling physical and biogeochemical fields in the coastal ocean. The design of field sampling programs through Observational System Simulation Experiments (OSSEs) is a challenging modeling opportunity with extremely valuable results. Sampling strategies at sea can be difficult due to the evolving nature of the circulation and OSSEs provide a realistic sitespecific simulation that can affect field protocols. Additionally, real-time limited-area forecast systems have been implemented on board research vessels to predict the transport of physical or biological tracers

at sea. Thus, using a coastal circulation forecasting system the likely path of the tracer of interest can be predicted and help researchers in the field develop appropriate sampling schemes.

Conclusions and Future Directions The ocean science community is currently presented with unprecedented opportunities and advanced technologies for understanding and managing coastal ecosystems. Rapid advancement of computer resources, observational systems and instruments, and numerical techniques are converging to enable real-time coastal observation systems for coastal monitoring and marine forecasting. In the past two decades, a variety of models for simulating the coastal ocean have emerged as significant tools for investigating processes and mechanisms as well as regional coastal ecosystems and environmental questions and issues. The application of these models to almost any regionin the world represents a remarkable scientific achievement. The state of the art of coastal oceancirculation models and computer



technology is such that a comprehensive and quantitative description of the hydrodynamics in a specific region can be obtained relatively easily by coastal oceanographers in general. Their application no longer requires expertise in numerical techniques and mathematics. However, many fundamental research questions still remain. There are several areas of active investigation for coastal circulation models that include formal development issues as well as applications. As computational power increases, larger-scaleproblems requiring more memory and faster computer speed will enable higher resolution regional studies as well as faster longer-term integrations for the purposes of climate studies. Advanced numerical methods for discretizing the model domain in both thehorizontal and vertical are being developed, particularly regarding mass conservation and the algorithms that transport scalar properties of the fluid volume like salt and heat, as well as nonconservative tracers like oxygen and nutrients. Advances in observational systems that include satellite and radar remote sensing, fixed instrumented platforms, remotely operated vehicles, and moored instruments, are currently being harnessed to provide as much near real-time information as possible to use in data assimilation schemes for oceanic numerical models. The modeling community in general is striving to provide forecasted global ocean circulation fields in nereal-time that resolve basinscale to coastal-scale features. Coastal circulation models will play an important role in: (1) communicating the open-ocean information to coastal and near-shore regions; (2) providing extensions to the basin-scale models to regions that are typically underresolved by the larger-scale models; and (3) providing realistic cross-shelf fluxes of mass and momentum to the bain-scale models. Operational forecasting systems are being developed for site-specific, limited-area predictions of the coastal ocean. In situ and remotesensed data are being assimilated by these systems, driven by forecasting results frommeteorological and basin-scale ocean models. Mesoscale weather models are being used to provide spatially dense estimates of surface flux parameters to the coastal circulation models but this coupling is largely one-way; the forecasted weather parameters affect the coastal hydrodynamic evolution. However, it is well known that the ocean affects the atmosphere. Observations have shown, for example, that the surface heat flux between the ocean and atmosphere over Gulf Stream waters significantly affects the development and evolution of extratropical cyclones that routinely pass along the eastern United States seaboard. Effective two-way

579

coupling that communicates surface fluxes from the coastal ocean to overlying atmosphere in coupled coastal ocean and regional weather forecasting models is currently being developed and will provide a significant enhancement to regional meteorological forecasting skill. The open-water boundaries of coastal circulation models require specification of either the velocity or the water level. For realistic regional simulations, there exists uncertainty in these boundary conditions related to the sparsity of observations on which to directly deduce them. The further development of schemes for assimilating observations into model integrations to provide optimal boundary conditions for forecasting is critical. Obtaining the open water boundary conditions from a larger basin-scale prediction model is another method for specifying open boundary conditions in operational limited-area coastal prediction models. This, in effect, generates the smaller domain boundary conditions from the larger model. Recognizing that the entire ocean functions as a single unit from global to estuarine scales, the coupling of coastal- and basin-scale ocean models will represent a significant advance toward global ocean forecasting. Since the formulations of the coastal and basin models are usually quite different, this poses an unsolved question of the communication between the two models. The ability of coastal circulation models to integrate the governing equations and boundary conditions for the coastal environment is a powerful tool for exploring both questions of process and mechanism and for addressing realistic regional problems that include forecasting of the coastal ocean analogous to atmospheric weather prediction. As the need to understand and address the growing list of environmental concerns accelerates, broad interdisciplinary efforts that couple models of different physical, biological, chemical, and geological systems will be critical in addressing these issues.

See also Coastal Trapped Waves. Data Assimilation in Models. Fishery Management. General Circulation Models. Heat and Momentum Fluxes at the Sea Surface. Moorings. Oil Pollution. Satellite Remote Sensing SAR. Storm Surges. Tidal Energy. Tides. Wind Driven Circulation.

Further Reading Brink KH and Robinson AR (eds.) (1999) The Sea, vols 10 and 11. New York: John Wiley.


580


Crowder LB and Werner FE (eds.) (1999) Fisheries oceanography of the estuarine-dependent fishes in the South Atlantic Bight. Fisheries Oceanography 8: 242. Gill AE (1982) Atmosphere–Ocean Dynamics. New York: Academic Press. Greenberg DA (1987) Modeling Tidal Power. Scientific American November: 128--131. Haidvogel DB and Beckmann A (1998) Numerical models of the coastal ocean. In: Brink KH and Robinson AR (eds.) The Sea, vol 10, pp. 457--482. New York: John Wiley. Heaps NS (ed.) (1987) Three-dimensional Coastal Ocean Models. Washington, DC: American Geophysical Union. Lynch DR and Davies AM (eds.) (1995) Quantitative Skill Assessment for Coastal Ocean Models. Washington, DC: American Geophysical Union. Malanotte-Rizzoli P (ed.) (1996) Modern Approaches to Data Assimilation in Ocean Modeling. New York: Elsevier.

Mooers CNK (ed.) (1998) Coastal Ocean Prediction. Washington, DC: Amercian Geophysical Union. Werner FE, Perry RI, Lough G, and Naimie CE (1996) Trophodynamic amd advective influences on Georges Bank larval cod and haddock. Deep-Sea Research II 43: 1793--1822. Werner FE, Quinlan JA, Blanton BO, and Luettich RA Jr (1997) The role of hydrodynamics in explaining variability in fish populations. Journal of Sea Research 37: 195--212. Westerink JJ, Luettich RA Jr, Baptista AM, Scheffner NW, and Farrar P (1992) Tide and storm surge predictions using a finite element model. ASCE Journal of Hydraulic Engineering 118: 1373--1390.


COASTAL TOPOGRAPHY, HUMAN IMPACT ON D. M. Bush, State University of West Georgia, Carrollton, GA, USA O. H. Pilkey, Duke University, Durham, NC, USA W. J. Neal, Grand Valley State University, Allendale, MI, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 480–489, & 2001, Elsevier Ltd.

Introduction The trademark of humans throughout time is the modification of the natural landscape. Topography has been modified from the earliest farming to the modern modifications of nature for transportation and commerce (e.g., roads, utilities, mining), and often for recreation, pleasure, and esthetics. While human modifications of the environment have affected vast areas of the continents, and small portions of the ocean floor, nowhere have human intentions met headlong with nature’s forces as in the coastal zone. A most significant change in human behavior since the 1950s has been the dramatic, rapid increase in population and nonessential development in the coastal zone (Figure 1). The associated density of development is in an area that is far more vulnerable and likely to be impacted by natural processes (e.g., wind, waves, storm-surge flooding, and coastal erosion) than most inland areas. Not only are more

Figure 1 The coastal population explosion has resulted in too many people and buildings crowded too close to the shoreline. As sea level rises, the shoreline naturally moves back and encounters the immovable structures of human development. In this example from San Juan, Puerto Rico, erosion in front of buildings has necessitated engineering of the shoreline.

people and development in harm’s way, but the human modifications of the coastal zone (e.g., dune removal) have increased the frequency and severity of the hazards. Finally, coastal engineering as a means to combat coastal erosion and management of waterways, ports, and harbors has had profound and often deleterious effects on coastal environments. The endproduct is a total interruption of sediment interchange between land and sea, and a heavily modified topography. Natural hazard mitigation is now moving with a more positive, albeit small, approach by restoring natural features, such as beaches and dunes, and their associated interchangeable sediment supply.

The Scope of Human Impact on the Coast The natural coastal zone is highly dynamic, with geomorphic changes occurring over several time scales. Equally significant changes are made by humans. On Ocean Isle, NC, USA, an interior dune ridge, the only one on the island, was removed to make way for development. The lowered elevation put the entire development in a higher hazard zone, with a corresponding greater risk for property damage from flooding and other storm processes. Another example of change, impacting on property damage risk, can be seen in Kitty Hawk, North Carolina. A large shorefront dune once extended in front of the entire community. The dune was constructed in the 1930s by the Civilian Conservation Corps to halt shoreline erosion, and provide a ‘protected’ area along which to build a road. The modification was done before barrier island migration was understood. Erosion was assumed to be permanent land loss. The artificial dune actually increased erosion here by acting as a seawall in a longterm sense, blocking overwash sand which would have raised island elevation and brought sand to the backside of the island, although the dune did afford some protection for development. As a consequence, buildings by the hundreds were built in the lee of the dune. Fifty years on, however, the price is being paid. During the 1980s, the dune began to deteriorate due to storm penetration, and the 1991 Halloween northeaster finished the job by creating large gaps in the dune, resulting in flooding of portions of the community. The dune cannot be rebuilt in place because the old dune location is now occupied by the


581

582

COASTAL TOPOGRAPHY, HUMAN IMPACT ON

beach, backed up against the frontal road. Between the time of dune construction and 1991, the community had only experienced major flooding once, in the great 1962 Ash Wednesday storm. Between 1991 and 2000, the community was flooded four times. The effect of shoreline engineering on a wholeisland system is starkly portrayed by the contrast between Ocean City, MD and the next island to the south, Assateague Island, MD. It has taken several decades to be fully realized, but the impact of the jetties is now apparent. Assateague Island has moved back one entire island width due to sand trapping by an updrift jetty. Similar stories abound along the coast. The Charleston lighthouse, once on the

backside of Morris Island, SC, now stands some 650 m at sea; a sentinel that watched Morris Island rapidly migrate away after the Charleston Harbor jetties, built in 1898, halted the supply of sand to the island (Figure 2). Human alterations of the natural environment have direct and indirect effects. Some types of human modifications to the coastal environment include: (1) construction site modification, (2) building and infrastructure construction, (3) hard shoreline stabilization, (4) soft shoreline stabilization, and (5) major coastal engineering construction projects for waterway, port, and harbor management and inlet channel alteration. Each of the modification types impacts the coastal environment in a variety of ways and also

1854 high water line 1979 high water line

N

Sullivans Island

Breach Inlet

Charleston Harbor jetties, completed 1896

Morris Island

Folly Island

Lighthouse 1890's: 2,700 feet (800 meters) back from shore 1940's: at the shoreline 1991's: 2,000 feet (600 meters) offshore and slightly tilted

Figure 2 The jetties that stabilize Charleston Harbor, South Carolina, were completed in 1896. The jetties block the southward transport sediment along the beach. As a result, the islands to the north of the jetties have grown seaward slightly, but the islands to the south of the inlet have eroded back more than 1400 m. The Morris Island Lighthouse, once on the back side of the island, is now 650 m offshore.



has several direct and indirect effects. Some of the effects are obvious and intuitive, but many are surprising in that there can be a domino effect as one simple modification creates potential for damage and destruction by increasing the frequency and intensity of natural hazards at individual sites. Construction Site Modification

Building sites are often flattened and vegetation is removed for ease of construction. Activities such as grading of the natural coastal topography include dune and forest removal. Furthermore, paving of large areas is common, as roads, parking lots, and driveways are constructed. Direct effects of building site modification, in addition to changes in the natural landform configuration, include demobilization of sediment in some places by paving and building footpaths, but also sediment mobilization by removal of vegetation. In either case, rates of onshore– offshore sediment transport and storm-recovery capabilities are changed, which can increase or decrease erosion rates as sediment supply changes. Other common site modifications include excavating through dunes (dune notching) to improve beach access or sea views. This is particularly common at the ends of streets running toward the beach. After Hurricane Hugo in South Carolina in 1989, shore-perpendicular streets where dunes were notched at their ocean termini were seen to have acted as storm-surge ebb conduits, funneling water back to the sea and increasing scour and property damage. The same effect was noted after Hurricane Gilbert along the northern coast of the Yucatań Peninsula of Mexico in 1988. Building and Infrastructure Construction

A variety of buildings are constructed in the coastal zone, ranging from single-family homes to high-rise hotels and commercial structures. Some of the common direct effects of building construction are alteration of wind patterns as the buildings themselves interact with natural wind flow, obstruction of sediment movement, marking the landward limit of the beach or dune, channelizing storm surge and storm-surge ebb flow, and reflection of wave energy. Indirect effects result from the simple fact that once there is construction in an area, people tend to want to add more construction, and to increase and improve infrastructure and services. As buildings become threatened by shoreline erosion, coastal engineering endeavors begin. Roads, streets, water lines, and other utilities are often laid out in the standard grid pattern used

583

inland, cutting through interior and frontal dunes instead of over and around coastal topography (Figure 3). Buildings block natural sediment flow (e.g., overwash) while the ends of streets and gaps between rigid buildings funnel and concentrate flow, accentuating the erosive power of flood waters. As noted above, during Hurricane Hugo, water, sand, and debris were carried inland along shore-perpendicular roads in several South Carolina communities. Storm-surge ebb along the shoreperpendicular roads caused scour channels, which undermined roadways and damaged adjacent houses and property. Even something as seemingly harmless as buried utilities may cause a problem as the excavation disrupts the substrate, resulting in a less stable topography after post-construction restorations. Plugging dune gaps can be a part of nourishment and sand conservation projects. Because dunes are critical coastal geomorphic features with respect to property damage mitigation, they are now often protected, right down to vegetation types that are critical to dune growth. Prior to strict coastal-zone management regulations, however, frontal dunes were often excavated for ocean views or building sites, or notched at road termini for beach access. These artificially created dune gaps are exploited by waves and storm-surge, and by storm surge ebb flows. Wherever dune removal for development has occurred, the probability is increased for the likelihood of complete overwash and possible inlet formation. Hard Shoreline Stabilization

Hard shoreline stabilization includes various fixed, immovable structures designed to hold an eroding shoreline in place. Hard stabilization is one of the most common modifiers of topography in the coastal zone and is discussed in more detail below. Seawalls, jetties, groins, and offshore breakwaters interrupt sediment exchange and reduce shoreline flexibility to respond to wave and tidal actions. Armoring the shoreline changes the location and intensity of erosion and deposition. Indirectly, hard shoreline stabilization gives a false sense of security and encourages increased development landward of the walls, placing more and more people and property at risk from coastal hazards including waves, storm surge, and wind. Eventual loss of the recreational beach as shoreline erosion continues and catches up with the static line of stabilization is almost a certainty. In addition, structures beget more structures as small walls or groins are replaced by larger and larger walls and groins.


584


Figure 3 A compilation of many of the impacts humans have on the coastal topography. In this fictional barrier island, roads have been cut through excavated dunes, maritime forest removed for building sites, finger canals dredged, structures built too close to the water, and several types of coastal engineering projects undertaken.

Soft Shoreline Stabilization

The most common forms of soft shoreline stabilization are beach nourishment, dune building, sand fencing, beach bulldozing (beach scraping), and planting of vegetation to grow or stabilize dunes. Direct effects of such manipulations are changes in sedimentation rates and severity of erosion, and interruption of the onshore–offshore sediment transfer, similar in that respect to hard shoreline stabilization. Indirectly, soft shoreline stabilization may make it more difficult to recognize the severity of an erosion problem, i.e., ‘masking’ the erosion problem. Moreover, as with hard shoreline stabilization, development is actually encouraged in the high-hazard zone behind the beach. Coastal Engineering Construction Projects

The construction of harbors, port facilities, waterways (e.g., shipping channels, canals) and inlet channel alterations significantly change the coastal

outline as well as eliminating land topographic features or erecting artificial shorelines and dredge spoil banks. The Intracoastal Waterway of the Atlantic and Gulf Coasts is one of the longest artificial coastal modifications in the world. Large harbors in many places around the world represent significant alteration of the landscape. Many examples of coastal fill or artificial shorelines exist, but one of the best examples of such a managed shoreline is Chicago’s 18 miles of continuous public waterfront. Major canals such as the Suez, Panama, Cape Cod, or Great Dismal Swamp Canal also represent major modifications in the coastal zone. The Houston Ship Channel made the city of Houston, Texas, a major port some 40 miles from the Gulf of Mexico. Tidal inlets, either on the mainland or between barrier islands, can be altered by dredging, relocation, or artificial closure. Direct effects of dredging tidal inlets are changes in current patterns, which may change the location and degree of erosion and deposition events, and prevention of sand



transfer across inlets. In either case, additional shoreline hardening is a common response.

The Scope of Coastal Engineering Impacts Between 80 and 90% of the American open-ocean shoreline is retreating in a landward direction because of sea-level rise and coastal erosion. Because more static buildings are being sited next to this moving and constantly changing coastline, our society faces major problems. Various coastal engineering approaches to dealing with the coastal erosion problem have been developed (Figure 3). More than a century of experience with seawalls and other engineering structures in New Jersey and other coastal developments shows that the process of holding the shoreline in place leads to the loss of the beach, dunes, and other coastal landforms. The real societal issue is how to save both buildings and beaches. The action taken often leads to modifications to the coast that limit the natural flexibility of the coastal zone to respond to storms, that inhibit the natural onshore– offshore exchange of sand, and that interrupt the natural alongshore flow of sand. Seawalls

Seawalls include a family of coastal engineering structures built either on land at the back of the beach or on the beach, parallel to the shoreline. Strictly defined, seawalls are free-standing structures near the surf-zone edge. The best examples are the giant walls of the northern New Jersey coast, the end result of more than a century of armoring the

Figure 4 Cape May, New Jersey has been a popular seaside resort since 1800. Several generations of larger and larger seawalls have been built as coastal erosion caught up with the older structures. Today in many places there is no beach left in front of the seawall.

585

shoreline (Figure 4). If such walls are filled in behind with soil or sand, they are referred to as bulkheads. Revetments, commonly made of piled loose rock, are walls built up against the lower dune-face or land at the back of the beach. For the purpose of considering their alteration of topography both at their construction site and laterally, the distinction between the types of walls is gradational and unimportant, and the general term seawall is used here for all structures on the beach that parallel the shoreline. Seawalls are usually built to protect the property, not to protect the beach. Sometimes low seawalls are intended only to prevent shoreline retreat, rather than to block wave attack on buildings. Seawalls are successful in preventing property damage if built strongly, high enough to avoid being overtopped, and kept in good repair. The problem is that a very high societal price is paid for such protection. That price is the eventual loss of the recreational beach and steepening of the shoreface or outer beach. This is why several states in the USA (e.g. Maine, Rhode Island, North Carolina, South Carolina, Texas, and Oregon) prohibit or place strict limits on shoreline armoring. Three mechanisms account for beach degradation by seawalls. Passive loss is the most important. Whatever is causing the shoreline to retreat is unaffected by the wall, and the beach eventually retreats up against the wall. Placement loss refers to the emplacement of walls on the beach seaward of the high-tide line, thus removing part or all of the beach when the wall is constructed (Figure 5). Seawall placement was responsible for much of the beach loss in Miami Beach, Florida, necessitating a major beach nourishment project, completed in 1981. Active loss is the least understood of the beach

Figure 5 An example of placement loss in Virginia Beach, Virginia. The seawall was built out on the recreational beach, instantaneously narrowing the beach in front of the wall.


586


degradation mechanisms. Seawalls are assumed to interact with the surf during storms, which enhances the rate of beach loss. This interaction can occur in a number of ways including seaward reflection of waves, refraction of waves toward the end of the wall, and intensification of surf-zone currents. By the year 2000, 50% of the developed shoreline on Florida’s western (Gulf of Mexico) coast was armored, the same as the New Jersey coast. Similarly 45% of developed shoreline on Florida’s eastern (Atlantic Ocean) coast was armored, in contrast to 27% for South Carolina, and only 6% of the developed North Carolina open-ocean shoreline. These figures represent the armored percentage of developed shorelines and do not include protected areas such as parks and National Seashores. Shoreline stabilization is a difficult political issue because seawalls take as long as five or six decades to destroy beaches, although the usual time range for the beach to be entirely eroded at mid-to-high tide may be only one to three decades. Thus it takes a politician of some foresight to vote for prohibition of armoring. Another issue of political difficulty is that there is no room for compromise. Once a seawall is in place, it is rarely removed. The economic reasoning is that the wall must be maintained and even itself protected, so most walls grow higher and longer. Groins and Jetties

Groins and jetties are walls or barriers built perpendicular to the shoreline. A jetty, often very long (thousands of feet), is intended to keep sand from flowing into a ship channel within an inlet and to reduce the cost of channel maintenance by dredging. Groins are much shorter structures built on straight stretches of beach away from inlets. Groins are intended to trap sand moving in longshore currents. They can be made of wood, stone, concrete, steel, or fabric bags filled with sand. Some designs are referred to as T-groins because the end of the structure terminates in a short shore-parallel segment. Both groins and jetties are very successful sand traps. If a groin is working correctly, more sand should be piled up on one side of the groin than on the other. The problem with groins is that they trap sand that is flowing to a neighboring beach. Thus, if a groin is growing the topographic beach updrift, it must be causing downdrift beach loss. Per Bruun, past director of the Coastal Engineering program at the University of Florida, has observed that, on a worldwide basis, groins may be a losing proposition, i.e. more beach may be lost than gained by the use of groins. After one groin is built, the increased rate of

Figure 6 A groin field along Pawleys Island, South Carolina. Trapping of sand on the updrift side of a groin, and erosion of the beach on the downdrift side usually results in a sawtooth pattern to the beach. Note that in this example the beach is the same width on both sides of each groin, indicating little or no longshore transport of sand.

erosion effect on adjacent beaches has to be addressed. So other groins are constructed, in selfdefense. The result is a series of groins sometimes extending for miles (Figure 6). The resulting groin field is a saw-toothed beach in plan view. Groins fail when continued erosion at their landward end causes the groin to become detached, allowing water and sand to pass behind the groin. When detachment occurs, beach retreat is renewed and additional alteration of the topography occurs. Jetties, because of their length, can cause major topographic changes. After jetty emplacement, massive tidal deltas at most barrier island inlents will be dispersed by wave activity. In addition, major buildout of the updrift and retreat of the downdrift shorelines may occur. In the case of the Charleston, SC, jetties noted earlier, beach accretion occurred on the updrift Sullivans Island and Isle of Palms.



Offshore Breakwaters

Offshore breakwaters are walls built parallel to the shoreline but at some distance offshore, typically a few tens of meters seaward of the normal surf zone. These structures dampen the wave energy on the ‘protected’ shoreline behind the breakwater, interrupting the longshore current and causing sand to be deposited and a beach to form. Sometimes these deposits will accumulate out to the breakwater, creating a feature like a natural tombolo. As in the case of groins, the sand trapped behind breakwaters causes a shortage of sediment downdrift in the directions of dominant longshore transport, leading to additional shoreline retreat (e.g. beach and dune loss, scarping of the fastland, accelerated mass wasting).

587

longshore transport. Removal of this lower beach sand deprives downdrift beaches of their natural nourishment, steepens the beach topographic profile, and destroys beach organisms. Dune building is often an important part of beach nourishment design, or it may be carried out independently of beach nourishment. Coastal dunes are a common landform at the back of the beach and part of the dynamic equilibrium of barrier beach systems. Although extensive literature exists about dunes, their protective role often is unknown or misunderstood. Frontal dunes are the last line of defense against ocean storm wave attack and flooding from overwash, but interior dunes may provide high ground and protection against penetration of overwash, and against the damaging effects of stormsurge ebb scour.

Beach Nourishment

Beach nourishment consists of pumping or trucking sand onto the beach. The goal of most communities is to improve their recreational beach, to halt shoreline erosion, and to afford storm protection for beachfront buildings. Many famous beaches in developed areas, in fact, are now artificial! The beach or zone of active sand movement actually extends out to a water depth of 9–12 m below the low-tide line. This surface is referred to as the shoreface. With nourishment, only the upper beach is covered with new sand so that a steeper beach is created, i.e. the topographic profile is modified on land and offshore. This new steepened profile often increases the rate of erosion; in general, replenished beaches almost always disappear at a faster rate than their natural predecessors. Beach scraping (bulldozing) should not be confused with beach nourishment. Beach sand is moved from the low-tide beach to the upper back beach (independent of building artificial dunes) as an erosion-mitigation technique. In effect this is beach erosion! A relatively thin layer of sand (r30 cm) is removed from over the entire lower beach using a variety of heavy machinery (drag, grader, bulldozer, front-end loader) and spread over the upper beach. The objectives are to build a wider, higher, high-tide dry beach; to fill in any trough-like lows that drain across the beach; and to encourage additional sand to accrete to the lower beach. The newly accreted sand in turn, can be scraped, leading to a net gain of sand on the manicured beach. An enhanced recreational beach may be achieved for the short term, but no new sand has been added to the system. Ideally, scraping is intended to encourage onshore transport of sand, but most of the sand ‘trapped’ on the lower beach is brought in by the

Human Impact on Sand Supply In most of the preceding discussion the impact of humans on beaches and shoreline shape and position was emphasized. The beach plays a major role in supplying sand to barrier islands and, in fact, is important in supplying sand and gravel to any kind of upland, mainland, or island. In this sense, any topographic modification, however small, that affects the sand supply of the beach will affect the topography. In beach communities, sand is routinely removed from the streets and driveways after storms or when sand deposited by wind has accumulated to an uncomfortable level for the community. This sand would have been part of the island or coastal evolution process. Often, dunes are replaced by flat, well-manicured lawns. Sand-trapping dune vegetation is often removed altogether. The previously mentioned Civilian Conservation Corps construction of the large dune line along almost the entire length of the Outer Banks of North Carolina is an example of a major topographic modification that had unexpected ramifications, namely the increased rate of erosion on the beach as well as on the backside of the islands. Prior to dune construction, the surf zone, especially during storms, expended its energy across a wide band of island surface which was overwashed several times a year. After construction of the frontal dune, wave energy was expended in a much narrower zone, leading to increased rates of shoreline retreat, and overwash no longer nourished the backside of the island. Now that the frontal dune is deteriorating, North Carolina Highway-12 is buried by overwash sand in a least a dozen places 1–4 times each year. Overwash sand is an important part of the island migration process,


588


because these deposits raise the elevation of islands, and when sediment extends entirely across an island, widening occurs. If not for human activities, much of the Outer Banks would be migrating at this point in time, but because preservation of the highway is deemed essential to connect the eight villages of the southern Outer Banks, the NC Department of Transportation removes sand and places it back on the beach. As a result, the island fails to gain elevation. Inlet formation also is an important part of barrier island evolution. Each barrier island system is different, but inlets form, evolve, and close in a manner to allow the most efficient means of moving water in and out of estuaries and lagoons. Humans interfere by preventing inlets from forming, by closing them after they open naturally (usually during storms), or by preventing their natural migration by construction of jetties. The net result is clogging of navigation channels by construction of huge tidal deltas and reduced water circulation and exchange between the sea and estuaries. Globally shoreline change is being affected by human activity that causes subsidence and loss of sand supply. The Mississippi River delta is a classic example. The sediment discharge from the Mississippi River has been substantially reduced by upstream dam construction on the river and its tributaries. Large flood-control levees constructed along the lower Mississippi River prevent sediment from reaching the marshes and barrier islands along the rim of the delta in the Gulf of Mexico. Natural land subsidence caused by compaction of muds has added to the problem by creating a rapid (1–2 m per century) relative sea-level rise. Finally, maintaining the river channel south of New Orleans has extended the river mouth to the edge of the continental shelf, causing most remaining sediment to be deposited in the deep sea rather than on the delta. The end result is an extraordinary loss rate of salt marshes and very rapid island migration. The face of the Mississippi River delta is changing with remarkable rapidity. Other deltas around the world have similar problems that accelerate changes in the shape of associated marshes and barrier islands. The Niger and Nile deltas have lost a significant part of their sediment supply because of trapping sand behind dams. Land loss on the Nile delta is permanent and not just migration of the outermost barrier islands. On the Niger delta the lost sediment supply is compounded by the subsidence caused by oil, gas, and water extraction. The barrier islands there are rapidly thinning. Sand mining is a worldwide phenomenon whose quantitative importance is difficult to guage. Mining dunes, beaches, and river mouths for sand has

reduced the sand supply to the shoreface, beaches, and barrier islands. In developing countries beachsand mining is ubiquitous, while in developed countries beach and dune mining often is illegal and certainly less extensive, although still a problem. For example, sand mining has adversely affected the beaches of many West Indies nations going through the growing pains of development. Dune mining has been going on for so long that many current residents cannot remember sand dunes ever being present on the beaches, although they must have been there at one time, given the sand supply and the strong winds. For example, on the dual-island nation of Antigua and Barbuda, beach ridges – evidence of accumulating sand – can be observed on Barbuda, but are missing on Antigua. The beach ridges of Barbuda have survived to date only because it is much less heavily developed and populated. Sadly, Barbuda’s beach ridges are being actively mined. Puerto Rico is a heavily developed Caribbean island, much larger than Antigua or Barbuda, and with a more diversified economy. Many of Puerto Rico’s dunes have been trucked away (Figure 7). East of the capital city of San Juan, large sand dunes were mined to construct the International Airport at Isla Verde by filling in coastal wetlands. As a result of removing the dunes, the highway was regularly overwashed and flooded during even moderate winter storms. First an attempt was made to rebuild the dune, then a major seawall was built to protect the lone coastal road. Dredging or pumping sand from offshore seems like a quick and simple solution to replace lost beach sand; however, such operations must be considered with great care. The offshore dredge hole may allow larger waves to attack the adjacent beach. Offshore sand may be finer in grain size, or it may be

Figure 7 The dunes here near Camuy, Puerto Rico, used to be over 20 m high. After mining for construction purposes, all that remains is a thin veneer of sand over a rock outcrop.



composed of calcium carbonate, which breaks up quickly under wave abrasion. In all of these cases, the new beach will erode faster than the original beach. Dredging also may create turbidity that can kill bottom organisms. Offshore, protective reefs may be damaged by increased turbidity. Loss of reefs will mean faster beach erosion, as well as the obvious loss to the fishery habitat. Sand can also be brought in from land sources by dump truck, but this may prove to be more expensive. Sand is a scarce resource, and beaches/dunes have been regarded as a source for mining rather than areas that need artificial replenishment. Past beach and dune mining may well be a principal cause of present beach erosion. In some cases, gravel may be better for nourishment than sand, but the recreational value of beaches declines when gravel is substituted. Sand mining of beaches and dunes accounts for many of Puerto Rico’s problem erosion areas. Such sand removal is now illegal, but permits are given to remove sand for highway construction and emergency repair purposes. However, the extraction limits of such permits are often exceeded – and illegal removal of sand for construction aggregate continues. In all cases, the sand removal eliminates natural shore protection in the area of mining, and robs from the sand budget of downdrift beaches, accelerating erosion. Even a small removal operation can set off a sequence of major shoreline changes. The Caribe Playa Seabeach Resort along the southeastern coast illustrates just such a chain reaction. Located west of Cabo Mala Pascua, the resort has lost nearly 15 m of beach in recent years according to the owner. The problem dates to the days before permits and regulation, when an updrift property owner sold beach sand for 50 cents per dump truck load; a bargain by anyone’s standards but a swindle to downdrift property owners. Where the sand was removed the beach eroded, resulting in shoreline retreat and tree kills. In an effort to restabilize the shore, and ultimately protect the highway, a rip-rap groin and seawall were constructed. Today, only a narrow gravel beach remains. Undoubtedly much of the aggregate in the concrete making up the buildings lining the shore, and now endangered by beach erosion, was beach sand. What extreme irony: taking sand from the beach to build structures that were subsequently endangered by the loss of beach sand.

Conclusion The majority of the world’s population lives in the coastal zone, and the percentage is growing. As this

589

trend continues, the coastal zone will see increased impact of humans as more loss of habitats, more inlet dredging and jetties, continued sand removal, topography modification for building, sand starvation from groins and jetties, and the increased tourism and industrial use of coasts and estuaries. Our society’s history illustrates the impact of humans as geomorphic agents, and nowhere is that fact borne out as it is in the coastal zone. The ultimate irony is that many of the human modifications on coastal topography actually decrease the esthetics of the area or increase the potential hazards.

See also Beaches, Physical Processes Affecting. Coastal Zone Management. Sandy Beaches, Biology of. Viral and Bacterial Contamination of Beaches.

Further Reading Bush DM and Pilkey OH (1994) Mitigation of hurricane property damage on barrier islands: a geological view. Journal of Coastal Research Special issue no. 12: 311– 326. Bush DM, Pilkey OH, and Neal WJ (1996) Living by the Rules of the Sea. Durham, NC: Duke University Press. Bush DM, Neal WJ, Young RS, and Pilkey OH (1999) Utilization of geoindicators for rapid assessment of coastal-hazard risk and mitigation. Ocean and Coastal Management 42: 647--670. Carter RWG and Woodroffe CD (eds.) (1994) Coastal Evolution: Late Quaternary Shoreline Morphodynamics. Cambridge: Cambridge University Press. Carter RWG (1988) Coastal Environments: An Introduction to the Physical, Ecological, and Cultural Systems of Coastlines. London: Academic Press. Davis RA Jr (1997) The Evolving Coast. New York: Scientific American Library. French PW (1997) Coastal and Estuarine Management, Routledge Environmental Management Series. London: Routledge Press. Kaufmann W and Pilkey OH Jr (1983) The Beaches are Moving: The Drowning of America’s Shoreline. Durham, NC: Duke University Press. Klee GA (1999) The Coastal Environment: Toward Integrated Coastal and Marine Sanctuary Management. Upper Saddle River, NJ: Prentice Hall. Nordstrom KF (1987) Shoreline changes on developed coastal barriers. In: Platt RH, Pelczarski SG, and Burbank BKR (eds.) Cities on the Beach: Management Issues of Developed Coastal Barriers, pp. 65--79. University of Chicago, Department of Geography, Research Paper no. 224.


590


Nordstrom KF (1994) Developed coasts. In: Carter RWG and Woodroffe CD (eds.) Coastal Evolution: Late Quaternary Shoreline Morphodynamics, pp. 477--509. Cambridge: Cambridge University Press. Nordstrom KF (2000) Beaches and Dunes of Developed Coasts. Cambridge: Cambridge University Press. Pilkey OH and Dixon KL (1996) The Corps and the Shore. Washington, DC: Island Press.

Platt RH, Pelczarski SG and Burbank BKR (eds.) (1987) Cities on the Beach: Management Issues of Developed Coastal Barriers. University of Chicago, Department of Geography, Research Paper no. 224. Viles H and Spencer T (1995) Coastal Problems: Geomorphology, Ecology, and Society at the Coast. New York: Oxford University Press.


COASTAL TRAPPED WAVES J. M. Huthnance, CCMS Proudman Oceanographic Laboratory, Wirral, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 489–496, & 2001, Elsevier Ltd.

Introduction Many shelf seas are dominated by shelf-wide motions that vary from day to day. Oceanic tides contribute large coastal sea-level variations and (on broad shelves) large currents. Atmospheric pressure and (especially) winds generate storm surges; strong currents and large changes of sea level. Other phenomena on these scales are wind-forced upwelling, along-slope currents and poleward undercurrents common on the eastern sides of oceans, responses to oceanic eddies, and alongshore pressure gradients. All these responses depend on natural waves that travel along or across the continental shelf and slope. These waves, which have scales of about one to several days and tens to hundreds of kilometers according to the width of the continental shelf and slope, are the subject of this article. Also included are ‘Kelvin’ waves, also coastally trapped, that travel cyclonically around ocean basins but with typical scales of thousands of kilometers both alongshore and for offshore decrease of properties. The waves have been widely observed through their association with the above phenomena. In fact they have been identified along coastlines of various orientations and all continents in both the Northern and Southern Hemispheres. Typically, the identification involves separating forced motion from the accompanying free waves. The ‘lowest’ mode with simplest structure (see below) has been most often identified; its peak coastal elevation is relatively easily measured. More complex forms need additional offshore measurements (usually of currents) for identification. This has been done (for example) off Oregon, the Middle Atlantic Bight and New South Wales (Australia). Observations substantiate many of the features described in the following sections.

Formulation Analysis is based on Boussinesq momentum and continuity equations for an incompressible sea of near-uniform density between a gently-sloping

seafloor z ¼ hðxÞ and a free surface z ¼ Zðx; tÞ where the surface elevation Z ¼ 0 for the sea at rest. Cartesian coordinates x ðx; yÞ; z (vertically up) rotate with a vertical component f =2. The motion, velocity components ðu; v; wÞ, is assumed to be nearly horizontal and in hydrostatic balance. (These assumptions are almost always made for analysis on these scales; they are probably not necessary but certainly simplify the analysis.) At the surface, pressure and stress match atmospheric forcing (for free waves). There is no component of flow into the seabed (generalizing to zero onshore transport uh at the coast); u-0 far from the coast (the trapping condition) or is specified by forcing.

Straight Unstratified Shelf This is the simplest context. Taking x offshore (and y alongshore; Figure 1) the depth is hðxÞ. Uniformity along shelf suggests wave solutions fuðxÞ; vðxÞ; ZðxÞg expðiky þ istÞ. For positive wave frequency, s, k > 0 corresponds to propagation in y, with the coast on the right (‘forward’ in the Northern Hemisphere). Then the momentum equations give u; v in terms of Z satisfying 0

ðhZ0 Þ þ KZ ¼ 0

½1

where KðaÞ kfh0 =s þ ðs2 f 2 Þ=g k2 h (uniform f ); primes (0 ) denote cross-shelf differentiation q=qx. The boundary conditions become hðsZ0 þ fkZÞ-0ðx-0Þ;

Z-0ðx-NÞ

½2

Free wave modes are represented by eigensolutions of eqns. [1] and [2]. Successive modes with more offshore nodes correspond to large positive K and arise in two ways. The term ðs2 f 2 Þ=g k2 h in K represents the gravity wave mechanism, modified by rotation; it increases with frequency s. For kf > 0 it gives rise to the ‘Kelvin wave’ – the mode with simplest offshore form; decay but no zeros of elevation. Other forms depending on this term (kf o0 and/or nodes of elevation offshore) are termed edge waves and discussed elsewhere. If there is no slope or boundary, K equals this term alone and plane inertiogravity waves are solutions of eqn. [1]. The term kfh0 =s in K increases with decreasing s if everywhere h increases offshore and kf > 0. It represents the following potential vorticity (angular


591

592

COASTAL TRAPPED WAVES

y z t

as

Co

0 (Northern Hemisphere).

Shallow Time t Deep

Shallow Deep

Time t + π/2σ Deep

Figure 2 Topographic wave mechanism. m displacement, velocity, sr relative vorticity (Northern Hemisphere).

momentum) restoring mechanism. If fluid is displaced into shallower water, it spreads laterally to conserve volume and therefore spins more slowly in total; taking account of the earth’s rotation, it acquires anticyclonic relative vorticity. ‘Forwards’ along the slope, the resulting up-slope velocity implies an up-slope displacement in time. Hence the upslope displacement propagates ‘forwards’ along the slope. (Behind it, the anticyclonic relative vorticity implies down-slope flow restoring the fluid location from its previous up-slope displacement.) This sequence is depicted in Figure 2. These modes are referred to as continental shelf waves. The mode forms and frequencies are known for several analytic models, e.g., level, uniformly sloping, exponential concave and convex shelves bordering an ocean of uniform depth. Numerical solutions for the waveforms and dispersion relations sðkÞ are easily found for any depth profile hðxÞ. For any monotonic profile hðxÞ the following have been proved. Phase propagates ‘forwards’ for all modes with so7f 7. Waves forms with 1, 2,y nodes offshore have frequencies 7f 7 > s1 > s2 > y defined for all k (subject to kf > 0). The Kelvin wave

frequency s0 ðkÞ > s1 ðkÞ is likewise defined for all kðkf > 0Þ and passes smoothly through 7f 7 to s0 > 7f 7 for large enough k. Edge waves with 0 (if kf o0), 1, 2,y nodes in the offshore form have increasing frequencies s > 7f 7; however, low wavenumbers (and frequencies) are excluded where the dispersion curves break the trapping criterion 0oKðNÞ ðs2 f 2 Þ=g k2 hðNÞ (Figure 3). Besides these properties, the following features are typical. Bounded h0 =h ensures a maximum sM in sðkÞ; near s1M, mode 1 velocity tends to be maximal near the shelf edge, and polarized anticyclonically; the nearest approach to inertial motion in this topographic context; here the group velocity qs=qk of energy propagation (in y) reverses through zero. If k-N, then sn -f =ð2n þ 1Þ for the n-node continental shelf wave which becomes concentrated over the ‘beach’ at the coast. As s, k-0, the shelf wave and Kelvin wave speeds s=k approach constant (maximum) values and u=s; v; n approach constant forms so that cross-slope velocities tend to zero. Variables v, Z and Z0 are in phase or antiphase, v and Z0 being near the geostrophic balance fv ¼ gZ0 ; u is 901 out of phase. Typically, Kelvin-wave currents in shallow shelf waters are polarized cyclonically but first-mode continental shelf wave currents are anticyclonic. Continental shelf wave-forms depend on the shape rather than the horizontal scale L of the depth profile. Phase speeds scale as fL and ðu; v; ZÞ scale as ðsU=f ; U; fUL=gÞ where the velocity scale U may be typically 0.1 m s1. Kelvin wave forms depend more on the depth; usually the phase speed is just less than ½ghðNÞ1=2 and ðu; v; ZÞ scale as ðsZL=h; ðg=hÞ1=2 Z; ZÞ where Z is typically 0.1 to 1 m. Quantitative results depend on the strength of forcing and accurate profile modeling; numerical calculations should be used for real shelves. The typical maximum in sðkÞ and associated reversal of group velocity qs=qk appears to be of practical significance. Shelf waves with frequency near the maximum (for some mode) appear in several observations, e.g., North Carolina sea levels, Scottish and Vancouver Island diurnal tides, wind-driven flow north of Scotland and over Rockall Bank. There may be a bias in seeking motion correlated with local forcing, i.e., responses with nonpropagating energy. Figure 4 shows modeled rotary currents over the shelf edge, continental shelf waves near the maximum frequency with slow energy propagation, as a response to impulsive wind forcing.

Other Geometry Continental shelf waves exist in more general contexts than a straight shelf, as identification in nature



by continuity. The boundary conditions for no flow through the bottom, zero pressure at the surface and trapping become

testifies. Analyses also verify their possibility in rectangular and circular basins. Perfect trapping around islands is only possible if so7f 7; then results are qualitatively as for a straight shelf except that wavelength around the island (and hence frequency) is quantized. For a broad shelf (distant coast), with the continental slope regarded as a scarp, again only waves in so7f 7 are trapped and results are qualitatively as for a straight shelf. An exception is the lowest mode, a ‘double’ Kelvin wave decaying to both sides. A seamount is again similar but introduces the same quantization as an island. A ridge comprises two scarps back-to-back. Each has its set of waves propagating ‘forwards’ (relative to the local slope) in so7f 7; a double Kelvin wave is associated with any net depth difference. Edge waves also propagate in both senses for s=7f 7 large enough to make K40 (see eqn. [1]). Similarly, a trench has sets of waves appropriate to each side. All cases hðxÞ and radial geometry hðrÞ can be treated numerically in the same way as a straight monotone profile.

qh=qxðqp=qx þ fkp=sÞ þ f 2 s2 N 2 qp=qz ¼ 0 ðz ¼ hÞ

½4

qp=qz þ N 2 p=g ¼ 0

ðz ¼ 0Þ p-0ðx-NÞ ½5

A flat bottom (uniform h) with uniform N25g/h admits the simplest Kelvin wave and (for n 0) internal Kelvin wave solutions cos½N ðz þ hÞ=cn exp ½ist þ isy=cn fx=cn

We consider the simplest context, a straight shelf with rest state density r0 ðzÞ; let N 2 g=r0 dr0 =dz. A wave form fuðx; zÞ; vðx; zÞ; wðx; zÞ; rðx; zÞ; pðx; zÞg expðiky þ istÞ is posed. Hydrostatic balance, density, and momentum equations give r; w; u and v, respectively, in terms of p; then ½3

0 1

f

Kel vin

2 1 Edge waves

Edge waves

½6

where cn ¼ ðghÞ1=2 ðn ¼ 0Þ; cn ¼ Nh=np ðn ¼ 1; 2; yÞ. Similar solutions, with vertical structure distributed roughly as N, exist for any NðzÞ > 0. Propagation is ‘forwards’ (cyclonic around the deep sea; anticyclonic around a cylindrical island) but depends only on density stratification. For a sloping bottom hðxÞ with offshore scale L (shelf width) and depth scale H, the parameter S N 2 H 2 =f 2 L2 indicates the importance of stratification. For small S, the solutions in so7f 7 are depthindependent Kelvin and continental shelf waves. As S increases, wave speeds s=k increase and nodes of u; v; p in the ðx; zÞ cross-section tilt outwards from the vertical towards horizontal. Correspondingly, seasonal changes have been observed in the vertical structure and offshore scale of currents on the Oregon shelf (for example) and off Vancouver Island, where stratification increases the offshore decay scale of upper-level currents. Quite moderate S may imply

Stratification

q2 p=qx2 þ f 2 s2 q=qz N2 qpqz k2 p ¼ 0

593

=f 1

2

3

Coastal trapped waves

k

O

Figure 3 Qualitative dispersion diagram. —— trapped waves, - - - - nearly trapped waves, - - - s2 ¼ f 2 þ k 2 ghðNÞ.



Norway

594

1200GMT 5/1/76

Scotland

Figure 4 Modeled continental shelf waves around Scotland at 1200Z, 5 January 1976 after impulsive wind forcing near Shetland on 3 January 1976. Dashed line shows 200 m depth contour. (Reproduced with permission from Huthnance JM (1995) Circulation, exchange and water masses at the ocean margin: the role of physical processes at the shelf edge. Progress in Oceanography 35: 353–431.)

s monotonic increasing in k, contrasting with the maximum in sðkÞ common among unstratified modes. For large S, the modes in so7f 7 become internal Kelvin-like waves with x replaced by x h1 ðzÞ; S-N corresponds to a shelf width L much less than the internal deformation scale NH=f ; the slope is ‘seen’ only as a coastal wall. Internal Kelvin waves have been observed in the Great Lakes and around Bermuda, where the bottom slope is steep. Similarly, records off Peru show an offshore scale B70 km, greater than the shelf width, because cn =f is large near the equator. For s > 7f 7 and nonzero S, trapping is imperfect. However, there are frequencies sðkÞ at which waves are almost trapped, radiate energy only slowly or respond with maximal amplitude to sustained forcing. These sðkÞ appear to correspond to dispersion curves in so7f 7 but there is still some uncertainty about the role of these waves; they may need some oceanic forcing. Bottom-trapped waves are an idealized form with motion everywhere parallel to a plane sloping seafloor (in uniform N 2 ); they decay away from the seafloor. They may propagate for so7f 7 or s > 7f 7 and up or down the slope, but always with a

component ‘forwards’ along the slope. If this phase propagation direction is f relative to the along-slope direction, then the velocity being transverse is at angle f to the slope and the frequency is s ¼ Nqh=qx cosf. In so7f 7 the general coastal trapped wave form tends for large k to bottomtrapped waves confined near the seafloor maximum of Nqh=qxo7f 7; then s ¼ Nqh=qx. If the maximum Nqh=qxo7f 7, then s increases to 7f 7 as k increases, a qualitative difference from unstratified behavior; formally there is a smooth transition at s ¼ 7f 7. Bottom-trapped wave identification may be difficult (despite the apparent prevalence of near-bottom currents), requiring knowledge of the local slope and stratification. There is evidence from continental slopes off the eastern USA and NW Africa.

Friction Friction causes cross-shelf phase shifts and significant damping of coastal trapped waves. The depth-integrated alongshore momentum balance R for idealized uniform conditions ðqp=qy ¼ 0; udz ¼ 0Þ is qv=qt þ rv=h ¼ t=rh suggesting that the flow v lags the forcing stress t less for low frequency, shallow



water, and large friction r. For example, nearshore currents lag the wind less than currents in deeper water offshore. Damping rates may be estimated as r=h ¼ Oð0:003 Uh1 Þ, i.e., a decay time less than 4 days for a typical current U ¼ 0.1 m s1 and depth h ¼ 100 m. In this estimate of friction, U should represent all currents present, e.g., tidal currents can provide strong damping. This decay time converts in to a decay distance cg h=r for a wave with energy propagation speed cg . Such decay distances are largest (hundreds to a thousand kilometers or more) for long waves with ‘forward’ energy propagation; much less for (short) waves with ‘backward’ energy propagation.

Mean Flows Mean currents are significant in many places where continental shelf waves have been observed, e.g., adjacent to the Florida current. Waves with phase speeds of a few meters per second or less may be significantly affected by boundary currents of comparable speed, or by vorticity of order f. By linear theory, advection in a uniform mean current V is essentially trivial, but could reverse the propagation of slower waves (higher modes or short wavelengths). Shear V 0 dV=dx modifies the background potential vorticity to PðxÞ ðf þ V 0 Þ=h; then the gradient of P (rather than f =h) underlies continental shelf wave propagation. ‘Barotropic’ instability is possible if V is strong enough; necessary conditions are P0 ðxs Þ ¼ 0 (some xs ) and P0 ½Vðxs Þ V > 0 for some x; the growth rate is bounded by max 7V 0 =27. Gulf Stream meanders have been interpreted as barotropically unstable shelf waves from Blake Plateau. In a stratified context, these effects of mean flow V may still apply. Additionally, vertical shear qV=qz is associated with horizontal density gradients: f qV=qz ¼ gr1 0 qr=qx

½7

and associated ‘baroclinic’ instability extracting gravitational potential energy. Two-layer models represent qr=qx by a sloping interface; the varying layer depths are another source of gradients qP=qx in each layer. Thus baroclinic instability may occur even if the current and total depth are uniform. More generally, slow flows over a gently sloping bottom are unstable only if qP=qx has both signs in the system. Such a two-layer channel model predicts instability at peak-energy frequencies in Shelikov Strait, Alaska (for example) and a corresponding wavelength roughly matching that observed. However, we caution that two-layer models may unduly

595

segregate internal Kelvin and continental shelf wave types, say, exaggerating the multiplicity of wave forms and scope for instability. In continuous stratification, the equivalent potential vorticity gradient qf =qx þ q2 V=qx2 þ f 2 ðN 2 qV=qzÞ=qz may support waves in the interior. Density contours rising coastward in association with a shelf edge surface jet modify the fastest continental shelf wave to an inshore ‘frontal-trapped’ form. The bottom slope also supports waves. Over a uniform bottom slope, additional bottom features can couple and destabilize the interior and bottom modes, even if Vðx; zÞ is otherwise stable. However, the bottom slope stabilizes bottom-intensified waves under an intermediate uniformly stratified layer.

Non-linear Effects These mirror typical nonlinear effects for waves. For example, each part of the nonlinear Kelvin wave form moves with the local speed v þ ½gðh þ ZÞ1=2 ; crests (Z and v positive) gain on troughs (Z and v negative) and wave fronts steepen. Similarly, for internal Kelvin waves between an upper layer, depth h, and a deep lower layer (density difference Dr) the local speed is ðghDr=rÞ1=2 everywhere; troughs gain on crests. Dispersion limits this nonlinear steepening; the associated shorter wavelengths propagate more slowly; and the steepening is left behind by the wave. For small amplitudes and long waves, steepening and dispersion are small and can balance in permanentform sech2 ðky þ stÞ solutions. Steepening may be important for internal Kelvin waves, but for typical continental shelf wave amplitudes (near linear) it takes weeks or months, longer than likely frictional decay times. Mean currents may be forced via frictional contributions to the time-averaged nonlinear convective derivatives in the momentum equations. The mean flows, scale h1 h0 f 1 uˆ2 , are typically confined close to the coast or the shelf break (uˆ denotes on–offshore excursion in the waves). Another mechanism is wave-induced form drag over an irregular bottom, giving a biased response to variable forcing; a ‘forward’ flow along the shelf. For example, low-frequency sinusoidal wind forcing may give a mean current up to a maximum fraction ð2pÞ1 of the value under a steady wind, i.e., some centimeters per second, principally in shallower shelf waters. Three-way interactions between coastal-trapped waves can occur if s3 ¼ s1 7s2 , k3 ¼ k1 7k2 , possible for particular combinations according to the shape of the dispersion curve. Typical timescales for energy exchange are many days; effects may be masked by frictional decay. Near a group velocity of


596


zero, there may be more response to a range of energy inputs.

Alongshore Variations If changes in the stratification and continental shelf form are small in one wavelength, then individual wave modes conserve a longshore energy flux; local wave forms are as for a uniform shelf. Thus Kelvin wave amplitudes increase as f 1=2 and are confined closer to the coast at higher latitudes. Energy flux conservation implies a large amplitude increase if waves of frequency s approach a shelf region where the maximum (sM ) for their particular mode is near s, as for Scottish and Vancouver Island diurnal tides; the energy tends to ‘pile up.’ If shelf variations cause sM to fall well below s, then the waves are totally reflected, with large amplitudes near where sM ¼ s. Poleward-propagating waves experience changing conditions. Near the equator, f is small, S (effective stratification) is large and internal Kelvin-like waves are expected, as off Peru; as f increases poleward, waves evolve to less stratified forms, more like continental shelf waves. (Variations of f are special in supporting offshore energy leakage to Rossby waves in the ocean.) Small irregularities in the shelf (lateral, vertical scales eðgHÞ1=2 =f ; eH) generally cause Oðe2 Þ effects, but OðeÞ nearby and in phase shifts after depth changes. Scattering occurs, preferentially to adjacent wave modes and (if unstratified) the highest mode at the incident frequency (having near-zero group velocity). However, long waves, LW , on long topographic variations (as above; LT ) adopt the appropriate local form; scattering is slow unless LW BLT . If the depth profile has a self-similar form h½ðx cðyÞÞ=LðyÞ then long (44L) continental shelf waves propagate with changes of amplitude but no scattering or change of form, provided that c and L also vary slowly ðc=c0 ; L=L0 44LÞ. Likewise, there is no scattering if the depth is hðxÞ where r2 xðx; yÞ ¼ 0, representing approximately uniform topographic convexity. In these cases, stronger currents and shorter wavelengths are implied on narrow sections of shelf. Abrupt features are apt to give the strongest scattering, substantial local changes or eddies on the flow. Scattering is the means of slope–current adjustment to a changed depth profile. However, all energy must remain trapped in so7f 7; even in s > 7f 7 special interior angles p=ð2n þ 1Þ can give perfect Kelvin wave energy transmission (for example). Successive reflections in a finite shelf (embayment) may synthesize near-resonant waves with small energy leakage. A complete barrier across the shelf implies reflection into (short, slow) waves of

opposite group velocity. There is a considerable literature of particular calculations. However, it is difficult to generalize, because of the several nonscattering cases interspersed among those with strong scattering.

Generation and Role of Coastaltrapped Waves Oceanic motion may impinge on the continental shelf. Notably at the equator, waves travel eastward to the coast and divide to travel north and south. In general, oceanic motions accommodate to the presence of the coast and shelf; at a wall, by internal Kelvin waves (vertical structure modes); for more realistic shelf profiles, by coastal trapped waves. Oceanic signals tend to be seen at the coast if alongshelf scale 4wave-decay distance or if the feature is shallower than the shelf-water depth. Natural modes of the ocean are significantly affected by continental shelves. Modes depending on f=h gradients have increased frequencies and forms concentrated over topography. Numerical models have shown 13 modes with periods between 30 and 80 hours, each mode being localized over one shelf area. Atmospheric pressure forcing the sea surface can be effective in driving Kelvin and edge waves, especially if there is some match of speed and scale, more likely in shallower (shelf) seas. Longshore wind stress s is believed to be the most effective means of generating coastal trapped waves. Within the forcing region, the flow tends to match the wind field; when or where the forcing ceases, the wave travels onwards (‘forwards’) and is then most recognizable. A simple view of this forcing is that s accelerates the alongshore transport hu. A more sophisticated view is that s induces a cross-shelf surface transport 7t7=rf ; coastal blocking induces a compensating return flow beneath, which is acted upon by the Coriolis force to give the same accelerating alongshore transport. A typical stress 0.1 Nm2 for 105 s (B1 day) accelerates 100 m water to 0.1 ms1. Winds blowing across depth contours may be comparably effective if the coast is distant. Other generation mechanisms include scattering (of alongshore flow, especially) by shelf irregularities as above, variable river runoff, and a co-oscillating sea. Shelf-sea motion is often dominated by tides and responses to wind forcing. On a narrow shelf, oceanic elevation signals penetrate more readily to the coast as the higher-mode decay distances are short; the tide is represented primarily by a Kelvin wave spanning the ocean and shelf. Model fits to semidiurnal measurements, showing a dominant



1983 31 Aug

1984 19 Nov

10 Oct

597

29 Dec

18 Mar

7 Feb

50

Longshelf velocity (cm/s)

_ 50

Hindcast

f25/1000

Observed less eddy mode 50

_ 50

f22/125

50

_ 50

f21/125 0

20

40

60

80

100 120 Days

140

160

180

200

220

Figure 5 Comparison of hindcast along-shelf currents using three coastal-trapped wave modes with measured currents in a crossshelf section after band-pass filtering and removing an eddy mode. (Reproduced with permission from Church JA, White NJ, Clarke AJ, Freeland HJ and Smith RL (1986) Coastal-trapped waves on the East Australian continental shelf. Part II: model verification. Journal of Physical Oceanography 16: 1945–1957.)

Kelvin wave, have been made off California, Scotland and north-west Africa, for example. Several areas at higher latitudes show dominant continental shelf wave contributions to diurnal tidal currents, e.g., west of Scotland, Vancouver Island, Yermak Plateau. On a wide shelf, there is correspondingly greater scope for wind-driven elevations and currents. The extensive forcing scale, typically greater than the shelf width, induces flow with minimal structure (typically no reversals across the shelf) corresponding to the lowest-mode continental shelf wave with maximum elevation signal at the coast. In stratified conditions, the upwelling or downwelling response also corresponds to a wave (or waves). As a wave travels, its amplitude is continually incremented by local forcing. A model based on this approach was used to make the hindcast of measured currents on the south-east Australian shelf shown in Figure 5. At any fixed position, the motion results from local forcing and from arriving waves, bringing the influence of forcing (e.g., upwelling) ‘forwards’ from the ‘backward’ direction. In the Peruvian upwelling regime, for example, variable currents are not well correlated with local winds but include internal Kelvin-like features coming from nearer the equator. This is hardly compatible with (common) simplifications of a zero alongshore pressure

gradient. Moreover, the waves carry the influence of assumed ‘backward’ boundary conditions far into a model. The same applies for steady flow. Friction introduces a ‘forward’ decay distance for a coastaltrapped wave; this distance has a definite low-frequency limit. Currents decay over these distances according to their structure as a wave combination. Thus alongshore evolution or adjustment of flow (however forced) is affected by coastal-trapped waves whose properties should guide model design.

Summary This article considers waves extending across the continental shelf and/or slope and having periods of the order of one day or longer. Their phase propagation is generally cyclonic, with the coast to the right in the Northern Hemisphere, a sense denoted ‘forward’; cross-slope displacements change watercolumn depth and relative vorticity, causing crossslope movement of adjacent water columns. At short-scales, energy propagation can be in the opposite ‘backward’ sense. Strict trapping occurs only for periods longer than half a pendulum day; shorterperiod waves leak energy to the deep ocean, albeit only slowly for some forms. The waves travel faster in stratified seas and on broad shelf-slope profiles;


598


speeds can be affected, even reversed, by along-shelf flows and reverses of bottom slope. Large amplitudes and abrupt alongshore changes in topography cause distortion and transfers between wave modes. The waves form a basis for the behavior (response to forcing, propagation) of shelf and slope motion on scales of days and the shelf width. Hence, they are important in shelf and slope–sea responses to forcing by tides, winds (e.g., upwelling), density gradients, and oceanic features. Their propagation (distance before decay) implies nonlocal response (over a comparable distance), especially in the ‘forward’ direction.

See also Coastal Circulation Models. Internal Tides. Internal Waves. Regional and Shelf Sea Models. Rossby Waves. Storm Surges. Tides. Upper Ocean Structure: Responses to Strong Atmospheric Forcing Events. Vortical Modes. Wind Driven Circulation.

Further Reading Brink KH (1991) Coastal trapped waves and wind-driven currents over the continental shelf. Annual Review of Fluid Mechanics 23: 389--412. Brink KH and Chapman DC (1985) Programs for computing properties of coastal-trapped waves and wind-driven motions over the continental shelf and slope. Woods Hole Oceanographic Institution, Technical Report 85–17, 2nd edn, 87–24. Dale AC and Sherwin TJ (1996) The extension of baroclinic coastal-trapped wave theory to superinertial frequencies. Journal of Physical Oceanography 26: 2305--2315. Huthnance JM, Mysak LA and Wang D-P (1986) Coastal trapped waves. In: Mooers CNK (ed.) Baroclinic Processes on Continental Shelves. Coastal and Estuarine Sciences, 3, pp. 1–18. Washington DC. American Geophysical Union. LeBlond PH and Mysak LA (1978) Waves in the Ocean. New York: Elsevier.


COASTAL ZONE MANAGEMENT D. R. Godschalk, University of North Carolina, Chapel Hill, NC, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Home to over half the world’s population, coastal area environments and economies make critical contributions to the wealth and well-being of maritime countries. Market values stem from fisheries, tourism and recreation, marine transportation and ports, energy and minerals, and real estate development. Nonmarket values include life support and climate control through evaporation and carbon absorption, provision of productive marine and estuarine habitats, and enjoyment of nature. Cumulative effects of human activities and coastal engineering threaten the sustainability of coastal resources. We are depleting coastal fisheries, degrading coastal water quality, draining coastal wetlands, blocking natural beach and barrier island movements, and spoiling recreational areas. Population growth is putting pressure on fragile ecosystems and increasing the number of people exposed to natural hazards, such as hurricanes, typhoons, and tsunamis. Maintaining sustainability demands active intervention in the form of coastal management. Managing the conservation and development of coastal areas is, however, a challenging enterprise. Coastal land and water resources follow natural system boundaries rather than governmental jurisdiction boundaries, fragmenting management authority and responsibility. Coastal scientists and coastal managers operate in different spheres, fragmenting knowledge creation and dissemination. Coastal resource programs tend to focus on single sectors, such as water quality or land use, fragmenting efforts to manage holistic ecosystems. The field of coastal management has evolved to deal with the challenges of managing fragmented transboundary coastal resources.

Evolution of Coastal Zone Management Coastal zone management developed to protect coastal resources from threats to their sustainability and to overcome the ineffectiveness of single function management approaches. Unlike ocean

management issues, such as freedom of navigation and conservation of migratory species, coastal management traditionally focused on issues related to the land–sea interface, such as shoreline erosion, wetland protection, siting of coastal development, and public access. Important conceptual landmarks in the field’s development are the intergovernmental framework of the 1972 US Coastal Zone Management Act, the 1987 Brundtland report proposing sustainable development, and the 1992 Earth Summit recommendation for initiation of integrated coastal management (ICM). Coastal Zone Management

In 1972, the United States enacted the Coastal Zone Management Act to create a formal framework for collaborative planning by federal, state, and local governments. To receive federal funding incentives, states were required to set coastal zone boundaries, define permissible land and water uses, and designate areas of particular concern, such as hazard areas. An additional incentive was the promise that federal government actions would be consistent with the state’s approved coastal zone management program. In practice, state programs tended to focus on landuse planning and regulation. Many other developed countries followed suit, establishing their own coastal zone management programs. North Carolina offers an example of a state program developed under the Coastal Zone Management Act. Its 1974 Coastal Area Management Act established a coastal resource management program for the 20 coastal counties influenced by tidal waters. Policy is made by the Coastal Resources Commission, whose members are appointed by the governor. The act is administered by the state Division of Coastal Management, which issues coastal development permits, manages coastal reserves, and provides financial and technical assistance for local government planning. Required local land-use plans must meet approved standards, include a post-storm reconstruction policy, and identify watershed boundaries, but localities have considerable implementation flexibility. Four Areas of Environmental Concern are designated: estuarine and ocean, ocean hazard, public water supplies, and natural and cultural resources areas. All major development projects within an Area of Environmental Concern must receive a state permit. Small structures must be set back beyond the 30-year line of estimated annual


599

600

COASTAL ZONE MANAGEMENT

beach erosion and large structures must be set back beyond the 60-year line. Shore-hardening structures are banned. While the North Carolina program has enjoyed considerable success, in practice many coastal localities continue to put economic development ahead of environmental protection.

The main ideas of sustainable development can be stated in terms of questions to be asked of every environment and development decision:

• • •

How does it improve the quality of human life? How does it affect the environment and natural resources? Are its benefits distributed equitably?

The New Paradigm of Sustainable Development

In 1987, the UN World Commission on Environment and Development published its report Our Common Future, referred to as the Brundtland report in recognition of its chairman, Norwegian Prime Minister Gro Harland Brundtland. The report called for global sustainability, a concept which is now the dominant paradigm in coastal management. Sustainability requires economic development which meets the needs of the present generation to be achieved without compromising the ability of future generations to meet their own needs. Sustainability demands balance among the elements of a triple bottom line, sometimes referred to as ‘the three e’s’: ecological sustainability, social equity, and economic enterprise. The triple bottom line has become a touchstone for accountability reporting of business as well as government actions. The 2002 UN World Summit on Sustainable Development revalidated the concept as a collective responsibility at local, national, regional, and global levels.

The Rise of Integrated Coastal Management

The 1992 United Nations Conference on Environment and Development in Rio de Janeiro – the Earth Summit – recommended principles to guide actions on environment, development, and social issues. It also approved Agenda 21, an action plan for sustainable development. While both the principles and actions are nonbinding, they represent a major shift toward understanding that sustainability must address the interdependence of environment and development in both developed countries (North) and developing countries (South). As shown in Figure 1, global environmental problems, such as greenhouse gases, generated in the North threaten the ability of the South to develop, while poverty and overpopulation in the South lead to local environmental stresses, such as air and water pollution. Interdependence generates the need for integration between environment and development in sectors

International trade system Developed countries (North)

Export natural resources Export manufactured goods

Developing countries (South)

Poverty Emissions and overconsumption

International finance system

Overpopulation

Lending through multilateral institutions Repayment of debt

Environmental stresses (primarily global)

,

Threaten the South s ability to develop

Environmental stresses (primarily local)

Figure 1 Interdependence of environment and development. Source: Cicin-Sain B and Knecht RW (1998) Integrated Coastal and Ocean Management: Concepts and Practices, p. 83. Washington, DC: Island Press.



and nations. The ocean and coastal issues chapter of Agenda 21 calls for ocean and coastal management to be integrated in content and precautionary and anticipatory in ambit. The precautionary principle holds that lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation. In other words, a conservative regulatory and management approach should be taken, in the absence of convincing evidence to the contrary. The Agenda 21 section on integrated management and sustainable development of coastal areas calls on each coastal state to establish coordination mechanisms at both the local and national levels. It suggests undertaking coastal- and marine-use plans, environmental impact assessment and monitoring, contingency planning for both human-induced and natural disasters, improvement of coastal human settlements, conservation and restoration of critical habitats, and integration of sectoral programs such as fishing and tourism into a coordination framework. It also calls for national guidelines for ICM and actions to maintain biodiversity and productivity of marine species and habitats. The section highlights the need for information on coastal and marine systems and natural science and social science variables, along with education and training and capacity building. The European Commission defines integrated coastal zone management (ICZM) as ‘‘a dynamic, multidisciplinary and iterative process to promote sustainable management of coastal zones. It covers the full cycle of information collection, planning (in its broadest sense), decision making, management and monitoring of implementation. ICZM uses the informed participation and cooperation of all stakeholders to assess the societal goals in a given coastal area, and to take actions towards meeting these objectives. ICZM seeks, over the long-term, to balance environmental, economic, social, cultural and recreational objectives, all within the limits set by natural dynamics. Integrated in ICZM refers to the integration of objectives and also to the integration of the many instruments needed to meet these objectives. It means integration of all relevant policy areas, sectors, and levels of administration. It means integration of the terrestrial and marine components of the target territory, in both time and space.’’ Coastal managers worldwide face many of the same problems: environmental degradation, marine pollution, fishery depletion, and loss of marine habitat. However, each nation’s coastal management programs differ due to unique geography, development issues, and political system. Still, similar packages of tools and techniques are seen in many countries. The packages include combinations of

601

regulations, national policies, planning, diagnostic studies of natural and socioeconomic systems and government capacity, incentives provision, and consensus building and participation to respond to conflicts. The ICM influence often is less evident in those developed countries with coastal programs already in place, than in less-developed countries where external funding assistance is needed to generate new coastal management institutions and to sustain traditional coastal livelihoods. How well does ICM work? Evaluation efforts typically take the form of best practice stories, assessments of process outputs, or descriptions of levels of implementation. Program success indicators based on management evaluations indicate that successful programs tend to be comprehensive (based on ecosystem boundaries, such as watersheds or river basins), participatory (involving stakeholders), cooperative (networking among organizations), contingent (allowing for uncertainty and change), precautionary (acting conservatively to preserve the environment in the absence of conclusive evidence), long-term (with time horizons beyond project deadlines), focused (aimed at perceived issues or problems), incremental (taking small steps toward objectives, as in managed retreat from the shore), and adaptive (using learning to redirect program efforts). When more experience has been acquired, it may be possible to evaluate program outcomes or long-term, large-scale successes. What types of ICM actions have been taken? Practices may take the form of innovative programs or institutions, as well as innovative applications to particular problems or content areas. Examples of both types of successful ICM best practices are illustrated in Table 1. Regional Integrated Coastal Management Initiatives

With the maturing of ICM, its scope has expanded to include regional or transnational projects and programs. Inherent linkages among terrestrial and ocean processes have led to a number of regional ocean initiatives; 16 regional action plans have been formally adopted as of 2001 under the UN’s Regional Seas Program. Each plan focuses on the unique problems of its region. For example, the East Asian Seas region has tackled problems of coastal aquaculture, fisheries exploitation, coral reef restoration, and resource extraction, using an ecoregion approach. Nations in the Western Indian Ocean region, where 70% of the African continent earn their living from natural resources, focus on simultaneously alleviating poverty and conserving natural resources. These developing regions promote sustainable


602


Table 1

Successful ICM practices

Long-range planning and marine zoning for Australia’s Great Barrier Reef Marine Park, where use and protection zones were mapped and participatory visioning was used to prepare a 25-year long-range plan Bringing ocean and coastal management together in the Republic of Korea’s new Ministry of Maritime Affairs and Fisheries, which integrated coastal zone management at the national level with navigation, ports, and fisheries programs Involving publics in operation of the special area management zones in Ecuador, through education, training, and outreach Determining the value of coastal ocean utilization in the pilot ICM program in Xiamen, China, through valuing the use of ocean space and charging users for mariculture and anchorage space Incorporating tradition management practices in American Samoa’s villages, which are responsible for monitoring and enforcing ICM measures on their land and waters Controlling coastal erosion in Sri Lanka through establishment of a coastal zone development permit system, prohibition of coral mining except for research, and recruiting coastal communities into the development of management plans Control of nonpoint marine pollution in the Chesapeake Bay, where a partnership among three US states (Virginia, Maryland, and Pennsylvania) and the federal government applies science to restore America’s largest and most productive estuary Protecting coastal resources in Turkey though designating specially protected areas to address rampant tourism development and its impacts on coastal waters and fisheries Community-based coral reef protection in Phuket Island, Thailand, with a bottom-up program of education and outreach to protect the main tourist attraction, 60% of which has been seriously damaged or degraded Excerpted from Cicin-Sain B and Knecht RW (1998) Integrated Coastal and Ocean Management: Concepts and Practices, pp. 297–299. Washington, DC: Island Press.

livelihoods in order to meet both conservation and development goals. The example of the European Union illustrates the evolution of a regional program. They conducted a 1996 demonstration with 35 projects in member countries and six thematic analyses of key ICM factors, including legislation, information, EU policy, territorial and sectoral cooperation, technical solutions, and participation. Their ICM strategy, approved in 2000, recommended that member states commit to a common vision for the future of their coastal zones, based on durable economic opportunities, functioning social and cultural systems in local communities, adequate open space, and maintenance of ecosystem integrity. It set out principles for a long-term holistic perspective, adaptive management, local specificity, working with natural processes, participatory planning, involvement of relevant administrative bodies, and use of a combination of instruments.

The EU strategy was able to draw on the experience gained from the 1975 Mediterranean Action Plan (MAP), one of the regional seas programs of the United Nations Environment Program (UNEP). Originally focused on marine pollution control, the MAP widened to deal with the land-based sources of 80% of pollution sources. It created coastal area management programs and, following Agenda 21, shifted them to an integrated basis. The Mediterranean region faces rapid coastal urbanization, growing tourist activities, high water consumption, concentrated pollution hot spots, biodiversity losses, and soil erosion. Lessons learned from the MAP experience emphasize the need for building an evaluation and monitoring mechanism at the beginning. Programs start out oriented to specific issues and problems, but must become more comprehensive at later stages in order to deal with complex linkages and provide integrated solutions. Strong political commitment at all levels to the preparation and initiation of initiatives, along with participation of stakeholders and end-users from program design through implementation, are of utmost importance. While ICM has provided a sturdy conceptual framework, full integration remains an elusive goal, as discovered in evaluation studies. Ten MAP projects implemented during 1987–2001 were evaluated. The projects were located in three entire national coastal areas (Albania, Syria, and Israel), two areas of large parts of the national coast (Malta and Lebanon), two semi-enclosed bays polluted by urban and industrial expansion (Kastela Bay in Croatia and Izmir Bay in Turkey), two islands (Rhodes and Malta), a polluted industrialized urban coast (Sfax, Tunisia), and a relatively virgin coastal area under threat of uncontrolled development (FukaMatrouh, Egypt). The evaluation found that many changes occurred in the region during the multiyear implementation period, requiring the ICM projects to adapt to the new conditions, institutional structures, and coastal area priorities. Initial pilot projects focused on data collection, studies of pollution and ecosystem protection, and capacity building. Firstand second-cycle projects focused on sectoral studies, resource management, implementation and its basic tools (e.g., environmental impact assessment and geographic information system or GIS), and introducing carrying capacity assessment. Third-cycle projects focused on sustainable development, applying ICM, and newer tools (strategic environmental assessment, systemic sustainability analysis, sustainability indicators, resource valuation, and socioeconomic evaluation). The evaluation reported that the projects had an overall significant impact on national systems in the



Mediterranean region, with a more significant impact on the developing countries. The impacts directly influenced the practices, capacities, and awareness of the national and local authorities, institutions, teams, and stakeholders, as well as the resident coastal populations. The result was a strengthened awareness of the importance of coastal areas and the need for sustainable development, as well as an improvement in the ICM concepts and practices. Three of the 10 projects were also evaluated by a World Bank/Mediterranean Environmental Technical Assistance team in 1997. They found that, despite weaknesses in preparation and financial support, the overall impact of the projects was good, especially in terms of increased institutional capacity, impact on decision makers, and catalytic role of the projects. Performance problems were weak sectoral integration, weak participation and absence of domestic financial support, and uncertain project follow-up. Assessments of ICM bring out both its strengths and weaknesses. Its strengths are based on the rationality of an integrated attack on interdependent coastal problems aimed at balancing ecological, economic, and equity values. No single government agency or program is equipped to deal with the complex interactions of oceans and coasts, yet it is precisely these interactions that pose such serious challenges to global sustainability. Its weaknesses stem from its need to reform and reinvent longstanding institutions, political turfs, and development practices. The times necessary to achieve such ambitious reforms far exceed the times of the usual 4–6-year ICM project. In absence of a crisis event to catalyze, and a huge infusion of funds to pay for, the necessary institutional change, progress proceeds slowly and long-term coordination takes a back seat to short-term demands. These same weaknesses plague efforts to manage all transboundary resources: air, water, and land.

603

green buildings and smart growth by professional organizations, private businesses, and governmental bodies should provide a positive foundation for strengthening ICM around the world. Still the questions remain: Can we expect the emergence of sufficient political will to reform decades of territorial and institutional balkanization? Can we overcome the perception that ICM actions are an obstacle to economic development of the coast? Indicators of a Looming Coastal Crisis

One possible catalyst for a worldwide surge in ICM is the impacts of the dangerous combination of climate change, natural hazards, resource depletion, poverty, and coastal population growth. Scientific studies have made it increasingly difficult for politicians to deny the reality of global warming. The transformation of the Earth’s atmosphere through carbon dioxide emissions is driving up global temperature, which is expected to cause a string of disasters, including hurricanes, droughts, glacial melting, sea level rise, and ocean acidification. Since the turn of the century, the world experienced two of history’s most catastrophic coastal disasters – the 2004 Indian Ocean tsunami with a death toll of 230 000 people triggered by an earthquake off the coast of Indonesia and the 2005 hurricane Katrina which devastated New Orleans and coastal Mississippi causing 1836 deaths and $81.2 billion in property damage. Yet people, rich and poor alike, continue to move to the ocean’s edge, fueling the growth of vulnerable settlements and megacities. An estimated 40% of world cities of 1–10 million population and 75% of cities over 10 million population were located on the coast in 1995 and the number is growing. Clearly, continuation of these trends is a recipe for future disasters unless coastal settlement patterns resilient to natural hazards can be ensured. Increasing Integration Efforts

Future Directions All indications are that the need for ICM will continue to grow along with continued integration efforts. Serious global environmental problems are on the rise and more and more people are flocking to coastal areas where they will be exposed to potential catastrophes. The natural response by scientists and planners, governments and nongovernmental organizations, and other concerned stakeholders will be to call for further integration of coastal and ocean programs, institutions, and strategies. Increasing support for sustainable development in the form of

One example of a major new ICM initiative is the 2004 report of the US Commission on Ocean Policy, An Ocean Blueprint for the 21st Century. The report acknowledges the magnitude of the crisis but insists that all is not lost. It envisions a future in which the coasts are attractive places to live, work, and play with clean water, public access, strong economies, safe harbors, adequate roads and services, and special protection for sensitive habitats and threatened species. Future management boundaries coincide with ecosystem regions, managers balance competing considerations and proceed with caution, and ocean governance is effective, participatory, and well


604


coordinated. The report recommends creation of a National Ocean Council within the Executive Office of the President to coordinate federal policy, encourages groups of states to form networked regional ocean councils, and calls for federal laws and programs to be consolidated and modified to improve sustainability, including: amending the Coastal Zone Management Act to add measurable goals and performance measures and to implement watershedbased coastal zone boundaries; changing federal funding and infrastructure programs to discourage growth in fragile or hazard-prone coastal areas; and revising the National Flood Insurance Program to reduce incentives to develop in high-hazard areas. Response to the report has been generally positive, but it remains to be seen as to whether the recommendations will be implemented. Possible Futures

The future of ICZM is closely tied to the future directions of global development. The Millennium Ecosystem Assessment has attempted to imagine possible future scenarios of global development pathways between 2000 and 2050. The scenarios represent different combinations of governance and economic development (global vs. regional) and of ecosystem management (reactive vs. proactive). They focus on ecosystem services and the effects of ecosystems on human well-being. The four scenarios are:

• • • •

Global Orchestration (socially conscious globalization, with emphasis on equity, economic growth, and public goods, and reactive toward ecosystems); Order from Strength (regionalized, with emphasis on security and economic growth, and reactive toward ecosystems); Adapting Mosaic (regionalized, with proactive management of ecosystems, local adaptation, and flexible governance); and TechnoGarden (globalized, with emphasis on using technology to achieve environmental outcomes and proactive management of ecosystems).

None of the scenarios represents a best or worst path, although ecosystems do better in the TechnoGarden scenario. However, they illuminate the potential consequences of choices in institutional design and management which could inform future decisions about coastal zone management. Whether the catalysts of environmental threats, integrated management reforms, and scenarios of potential ecosystem outcomes can turn the tide toward fully integrated and effective coastal

management is an open question. What is clear is that a strong worldwide movement is actively pursuing this goal.

See also Coastal Topography, Human Impact on. Glacial Crustal Rebound, Sea Levels and Shorelines. Economics of Sea Level Rise. Tsunami.

Further Reading Barbiere J and Li H (2001) Third Millennium Special Issue on Megacities. Ocean and Coastal Management 44: 283--449. Beatley T, Brower DJ, and Schwab AK (2002) An Introduction to Coastal Zone Management, 2nd edn. Washington, DC: Island Press. Belfiore S (ed.) (2002) From the 1992 Earth Summit to the 2002 World Summit on Sustainable Development: Continuing Challenges and New Opportunities for Capacity Building in Ocean and Coastal Management. Special Issue on Capacity Building Ocean and Coastal Management 45: 541--718. Burbridge P and Humphrey S (eds.) (2003) Special Issue: The European Demonstration Programme on Integrated Coastal Zone Management. Coastal Management 31: 121--212. Cicin-Sain B and Knecht RW (1998) Integrated Coastal and Ocean Management: Concepts and Practices. Washington, DC: Island Press. Cicin-Sain B, Pavlin I, and Belfiore S (eds.) (2002) Sustainable Coastal Management: A Transatlantic and Euro-Mediterranean Perspective. Dordrecht: Kluwer. Francis J, Tobey J, and Torell E (eds.) (2006) Special Issue: Balancing Development and Conservation Needs in the Western Indian Ocean Region. Ocean and Coastal Management 49: 789--888. Glavovic B (2006) Coastal sustainability – an elusive pursuit: Reflections on South Africa’s coastal policy experience. Coastal Management 34: 111--132. Heilerman S (ed.) (2006) IOC Manuals and Guides 46, ICAM Dossier 2: A Handbook for Measuring the Progress and Outcomes of Integrated Coastal and Ocean Management. Paris: UNESCO. Kay R and Alder J (2005) Coastal Planning and Management, 2nd edn. London: Taylor and Francis. Mageau C (ed.) (2003) Special Issue: The Role of Indicators in Integrated Coastal Management. Ocean and Coastal Management 46: 221--390. Sain B and Knecht RW (1998) Integrated Coastal and Ocean Management: Concepts and Practices. Washington, DC: Island Press. Stojanovic T, Ballinger RC, and Lalwani CS (2004) Successful integrated coastal management: Measuring it with research and contributing to wise practice. Ocean and Coastal Management 47: 273--298.



Tobey J and Lowry K (eds.) (2002) Learning from the practice of integrated coastal management. Coastal Management 30: 283--345. US Commission on Ocean Policy (2004) An Ocean Blueprint for the 21st Century. Final Report. Washington, DC: US Commission on Ocean Policy. Vallega A (2002) The regional approach to the ocean, the ocean regions, and ocean regionalization – a post-modern dilemma. Ocean and Coastal Management 45: 721--760. Williams E, Mcglashan D, and Finn J (2006) Assessing socioeconomic costs and benefits of ICZM in the European Union. Coastal Management 34: 65--86.

Relevant Websites http://web.worldbank.org – Coastal and Marine Management, World Bank. http://www.unesco.org – Coastal Regions and Small Islands (CSI) Portal, UNESCO. http://www.coastalguide.org – EUCC: Coastal Guide. http://ec.europa.eu – EUROPA Coastal Zone Policy. http://www.keysheets.org – Integrated Coastal Management Keysheet. http://www.coastalmanagement.com – Integrated Coastal Management Websites. http://www.icriforum.org – International Coral Reef Initiative.

605

http://ioc3.unesco.org – IOC Marine Sciences and Observations for Integrated Coastal Management. http://www.millenniumassessment.org – Millennium Ecosystem Assessment. http://www.netcoast.nl – NetCoast: A Guide to Integrated Coastal Management. http://www.oneocean.org – OneOcean: Coastal Resources and Fisheries Management of the Philippines. http://www.pemsea.org – PEMSEA Integrated Coastal Management. http://www.unep.org – Regional Seas Programme, United Nations Environment Programme. http://www.deh.gov.au – The Contribution of Science to Integrated Coastal Management. http://www.globaloceans.org – The Global Forum on Oceans, Coasts, and Islands. http://www.csiwisepractices.org – Wise Coastal Practices for Sustainable Human Development Forum. http://www.ngdc.noaa.gov – World Data Center for Marine Geology and Geophysics: National Geophysical Data Center. http://earthtrends.wri.org – World Resources Institute: Earth Trends.


COCCOLITHOPHORES T. Tyrrell, National Oceanography Centre, Southampton, UK J. R. Young, The Natural History Museum, London, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction Coccolithophores (Figure 1) are a group of marine phytoplankton belonging to the division Haptophyta. Like the other free-floating marine plants (phytoplankton), the coccolithophores are microscopic (they range in size between about 0.003 and 0.040 mm diameter) single-celled organisms which obtain their energy from sunlight. They are typically spherical in shape. They are distinguished from other phytoplankton by their construction of calcium carbonate (CaCO3) plates (called coccoliths) with

which they surround their cells. While not quite the only phytoplankton to use CaCO3 (there are also some calcareous dinoflagellates), they are by far the most numerous; indeed they are one of the most abundant of phytoplankton groups, comprising in the order of 10% of total global phytoplankton biomass. The first recorded observations of coccoliths were made in 1836 by Christian Gottfried Ehrenberg, a founding figure in micropaleontology. The name ‘coccoliths’ (Greek for ‘seed-stones’) was coined by Thomas Henry Huxley (famous as ‘Darwin’s bulldog’) in 1857 as he studied marine sediment samples. Both Ehrenberg and Huxley attributed coccoliths to an inorganic origin. This was soon challenged by Henry Clifton Sorby and George Charles Wallich who inferred from the complexity of coccoliths that they must be of biological origin, and supported this with observations of groups of coccoliths aggregated into empty spheres.

Figure 1 Electron microscope images of some major coccolithophore species: (a) Coccolithus pelagicus, (b) Calcidiscus quadriperforatus, (c) Emiliania huxleyi, (d) Gephyrocapsa oceanica, (e) Florisphaera profunda, (f) Discosphaera tubifera.

606


COCCOLITHOPHORES

Species and Distribution Approximately 200 species of coccolithophore have been formally described, separated into 65 genera. However, the true number of authentic modern coccolithophore species is rather unclear, for a couple of reasons. First, it is now realized that pairs of species, previously thought to be distinct and rather unrelated, are actually different life-cycle stages of the same species; coccolithophores typically have life cycles in which the haploid (single set of chromosomes, as in sex cells) and diploid (double set of chromosomes) phases can form different coccolith types: ‘heterococcoliths’ during the diploid life stage, and ‘holococcoliths’ during the haploid life stage. This type of life cycle has long been known from classic studies of laboratory cultures. It has only recently been appreciated that it is a very widespread pattern, as a result of observations of combination coccospheres representing the transition between the two life-cycle phases, that is to say possessing half a covering of heterococcoliths and half a covering of holococcoliths (Figure 2). Fifty of the 200 described coccolithophores are taxa known only from their holococcolith-producing phase and so may prove to be part of the life cycle of a heterococcolithproducing species. The second factor making

Figure 2 A combination coccosphere, upper half (inner layer) heterococcoliths, lower half (outer layer) holococcoliths, of the species Calcidiscus quadriperforatus (the two stages were previously regarded as two separate species – Calcidiscus leptoporus and Syracolithus quadriperforatus – prior to discovery of this combination coccosphere). Scale ¼ 1 mm. Electron microscope image provided by Markus Geisen (AlfredWegener-Institute, Germany).

607

diversity estimates difficult is that recent research combining studies of fine-scale morphology, biogeography, and molecular genetics has suggested that many described species are actually clusters of a few closely related, but genetically distinct, species. Indeed, as a result of such studies, numerous additional morphotypes have been recognized and await formal description. Coccolithophores occur widely throughout the world’s oceans, with the exception of the very-high-latitude polar oceans. Most individual species have more restricted biogeographical ranges than the range of the group as a whole, but still typically have interoceanic distributions. Unlike diatoms (the other major group of phytoplankton that make hard mineral shields), they are absent from almost all freshwater rivers and lakes. They occur in the brackish (more saline than freshwater but less saline than seawater) Black Sea, but not in the brackish Baltic Sea. In contrast once more to diatoms, coccolithophores are almost exclusively planktonic. There are very few bottomdwelling species, even at shallow depths experiencing adequate light levels. Most species today live in warm, nutrient-poor conditions of the subtropical oceanic gyres, where they form a prominent component of the phytoplankton; there are fewer species that inhabit coastal and temperate or subpolar waters. Emiliania huxleyi (Figure 1(c)) is the best-known species, primarily because it forms intense blooms which are clearly visible in satellite images, appearing as pale turquoise swirls in the ocean (Figure 3). While E. huxleyi frequently dominates phytoplankton counts in seawater samples, at least in terms of numbers of cells, their cells (and therefore also the coccoliths that surround them) are rather small, with the cells about 5 mm across and the coccoliths about 3 mm long. No other coccolithophore species regularly forms blooms, although occasional blooms of other species, for instance Gephyrocapsa oceanica and Coccolithus pelagicus, have been recorded. Many other species, for example Calcidiscus quadriperforatus and Umbilicosphaera sibogae, are most successful in low-productivity waters but do not bloom there. Although these species are almost always much less numerous than E. huxleyi in water samples, they are on the other hand also significantly larger, with typical cell diameters greater than 10 mm and correspondingly larger coccoliths. Most species of coccolithophores are adapted to life in the surface mixed layer, but some species, such as Florisphaeara profunda, are confined to the deep photic zone where they make an important contribution to the ‘shade flora’.


Figure 3 SeaWiFS satellite image from 15 June 1998 of E. huxleyi blooms (the turquoise patches) along the west coast of Norway and to the southwest of Iceland. The perspective is from a point over the Arctic Ocean, looking southward down the North Atlantic. The Greenland ice sheet is visible in the center foreground. Imagery provided with permission by GeoEye and NASA SeaWiFS Project.

608 COCCOLITHOPHORES


COCCOLITHOPHORES

In oligotrophic surface waters, typical concentrations of coccolithophores are in the range 5000– 50 000 cells per liter. To put this in context, a teaspoonful (5 ml) of typical surface open ocean seawater will contain between 25 and 250 coccolithophore cells. Blooms of E. huxleyi have been defined as concentrations exceeding 1 million cells per liter; the densest bloom ever recorded, in a Norwegian fiord, had a concentration of 115 000 000 cells per liter. Blooms of E. huxleyi can cover large areas; the largest ever recorded bloom occurred in June 1998 (see Figure 3) in the North Atlantic south of Iceland and covered about 1 million km2, 4 times the area of the United Kingdom.

Coccoliths Coccolithophores, and the coccolith shields with which they surround themselves, are incredibly small. And yet, despite their small size, coccoliths are elegant and ornate structures, which, if the water chemistry is suitable, are produced reliably with few malformations. This efficient manufacture occurs at a miniature scale: the diameter of an E. huxleyi coccolith ‘spoke’ (Figure 1(c)) is of the order 50 nm, considerably smaller than the wavelength range of visible light (400–700 nm). Calcite is mostly transparent to visible light (unsurprisingly, given that coccolithophores are photosynthetic) and the small coccoliths are often at the limit of discrimination, even under high magnification. However, under cross-polarized light, coccoliths produce distinctive patterns which are closely related to their structure. As a result most coccoliths can be accurately identified by light microscopy. However, the details and beauty of coccoliths can only be properly appreciated using electron microscopy (Figure 1). Coccolithophores synthesize different types of coccoliths during different life-cycle stages. Here we concentrate on the heterococcoliths associated with the diploid life stage. These heterococcoliths are formed from crystal units with complex shapes, in contrast to holococcoliths which are constructed out of smaller and simpler crystal constituents. Coccoliths are typically synthesized intracellularly (within a vesicle), probably one at a time, and subsequently extruded to the cell surface. The time taken to form a single coccolith can be less than 1 h for E. huxleyi. Coccoliths continue to be produced until a complete coccosphere covering (made up of maybe 20 coccoliths, depending on species) is produced. Most coccolithophores construct only as many coccoliths as are required in order to provide a

609

complete single layer around their cell. Emiliania huxleyi is unusual in that, under certain conditions, it overproduces coccoliths; many more coccoliths are built than are needed to cover the cell. In these conditions, multiple layers of coccoliths accumulate around the E. huxleyi cell until the excessive covering eventually becomes unstable and some of the coccoliths slough off to drift free in the water. The large number of unattached coccoliths accompanying an E. huxleyi bloom contributes to a great extent to the turbidity of the water and to the perturbations to optics that make the blooms so apparent from space. Curiously, the functions of coccoliths are still uncertain. It is probable that a major function is to provide some protection from grazing by zooplankton, but many alternative hypotheses have also been advanced. For instance, the coccoliths may increase the rate of sinking of the cells through the water (and therefore also enhance the rate at which nutrient-containing water flows past the cell surface) or they may provide protection against the entry of viruses or bacteria to the cell. At one time it was thought that coccoliths might provide protection against very high light intensities, which could explain the resistance to photoinhibition apparent in E. huxleyi, but various experimental results make this explanation unlikely. One species, F. profunda, a member of the deeper ‘shade flora’, orients its coccoliths in such a way that they conceivably act as a light-focusing apparatus maximizing photon capture in the darker waters it inhabits (Figure 1(e)). Some species produce trumpet-like protrusions from each coccolith (Figure 1(f)), again for an unknown purpose. Currently there is a paucity of hard data with which to discriminate between the various hypotheses for coccolith function, and the diversity of coccolith morphology makes it likely that they have been adapted to perform a range of functions.

Life Cycle Many details are still obscure, and data are only available from a limited number of species, but it appears that most coccolithophores alternate between fully armored (heterococcolith-covered) diploid life stages and less-well-armoured (either holococcolith-covered or else naked) haploid phases. Both phases are capable of indefinite asexual reproduction, which is rather unusual among protists. That sexual reproduction also occurs fairly frequently is evidenced by the observation of significant genetic diversity within coccolithophore blooms. Bloom populations do not consist of just one clone


610

COCCOLITHOPHORES

(just one genetic variant of the organism). Coccolithophore gametes (haploid stages) are radically different from those of larger (multicellular) organisms in the sense of being equipped for an independent existence: they can move about, acquire energy (photosynthesize), and divide asexually by binary fission. Naked diploid phases can be induced in cultures, but these may be mutations which are not viable in the wild. There are no confirmed identifications of resting spore or cyst stages in coccolithophores. Coccolithophores, in common with other phytoplankton, experience only an ephemeral existence. Typical life spans of phytoplankton in nature are measured in days. Comparison of the rate at which CaCO3 is being produced in open ocean waters (as measured by the rate of uptake of isotopically labeled carbon), to the amounts present (the ‘standingstock’), has led to the calculation that the average turnover (replacement) time for CaCO3 averages about 3 days, ranging between a minimum of o1 and a maximum of 7 days at different locations in the Atlantic Ocean. This implies that if a surfacedwelling coccolithophore synthesizes coccoliths on a Monday, the coccoliths are fairly unlikely to still be there on the Friday, either because they have redissolved or else because they have sunk down to deeper waters. The genome of one species, E. huxleyi, has recently been sequenced, but at the time of writing its analysis is at an early stage.

Calcification Calcification is the synthesis of solid calcium carbonate from dissolved substances, whether passively by spontaneous formation of crystals in a supersaturated solution (inorganic calcification) or actively through the intervention of organisms (biocalcification). The building of coccoliths by coccolithophores is a major fraction of the total biocalcification taking place in seawater. Inorganic calcification is not commonplace or quantitatively significant in the global budget, with the exception of ‘whitings’ that occur in just a few unusual locations in the world’s oceans, such as the Persian Gulf and the Bahamas Banks. The chemical equation for calcification is Ca2þ þ 2HCO3 ) CaCO3 þ H2 O þ CO2 Heterococcoliths are constructed out of calcite (a form of calcium carbonate; corals by contrast synthesize aragonite, which has the same chemical

composition but a different lattice structure). Heterococcolith calcite typically has a very low magnesium content, making coccoliths relatively dissolution-resistant (susceptibility to dissolution increases with increasing magnesium content). Dissolved inorganic carbon in seawater is comprised of three different components: bicarbonate ions (HCO3 ), carbonate ions (CO3 2 ), and dissolved CO2 gas (CO2(aq)), of which it appears that bicarbonate or carbonate ions are taken up to provide the carbon source for CaCO3 (coccoliths have a d13C isotopic composition that is very different from dissolved CO2 gas). The exact physiological mechanisms of calcium and carbon assimilation remain to be established. Calcification (coccolith genesis) is stimulated by light but inhibited in most cases by plentiful nutrients. Separate experiments have found that increased rates of calcification in cultures can be induced by starving the cultures of phosphorus, nitrogen, and zinc. Low levels of magnesium also enhance calcification, and high levels inhibit it, but in this case probably because Mg atoms can substitute for Ca atoms in the crystalline lattice and thereby ‘poison’ the lattice. Calcification shows the opposite response to levels of calcium, unsurprisingly. Progressive depletion of calcium in the growth medium induces progressively less normal (smaller and more malformed) coccoliths. The calcification to photosynthesis (C:P) ratio in nutrient-replete, Ca-replete cultures is often in the vicinity of 1:1 (i.e., more or less equivalent rates of carbon uptake into the two processes). Low levels of iron appear to depress calcification and photosynthesis equally. Measurements at sea suggest that the total amount of carbon taken up by the whole phytoplankton community to form new CaCO3 is rather small compared to the total amount of carbon taken up to form new organic matter. Both calcification carbon demand and photosynthetic carbon demand have recently been measured on a long transect in the Atlantic Ocean and the ratio of the two was found to average 0.05; or, in other words, for every 20 atoms of carbon taken up by phytoplankton, only one on average was taken up into solid CaCO3.

Ecological Niche In addition to our lack of knowledge about the exact benefit of a coccosphere, we also have rather little definite knowledge as to the ecological conditions that favor coccolithophore success. There is certainly variation between species, with some being adapted to relatively eutrophic conditions (although diatoms invariably dominate the main spring blooms in


COCCOLITHOPHORES

temperate waters, as well as the first blooms in nutrient-rich, recently upwelled waters) and some to oligotrophic conditions. Most species are best adapted to living near to the surface, but some others to the darker conditions prevailing in the thermocline. Most species today live in warm, nutrient-poor, open ocean conditions; the highest diversity occurs in subtropical oceanic gyres, whereas lower diversity occurs in coastal and temperate waters. Much of our knowledge of coccolithophore physiology and ecology comes from studies of E. huxleyi, which has attracted more scientific interest than the other coccolithophore species because of its ease of culturing and the visibility of its blooms from space. The ability to map bloom distributions from space provides unique information on the ecology of this species. Blooms of the species E. huxleyi occur preferentially in strongly stratified waters experiencing high light levels. Coccolithophore success may be indirectly promoted by exhaustion of silicate, due to exclusion of the more competitive diatoms. By analogy with diatoms, whose success is contingent on silicate availability for their shell building, coccolithophores might be expected to be more successful at high CaCO3 saturation state O ð¼ ½CO3 2 ½Ca2þ =Ksp Þ, because the value of O controls inorganic calcification and dissolution. Such a dependency would render coccolithophores vulnerable to ocean acidification, as discussed further below. It was formerly thought that E. huxleyi was particularly successful in phosphate-deficient waters, but a reassessment has suggested that this is not a critical factor. Many coccolithophores are restricted to the warmer parts of the oceans, although this may be coincidental rather than due to a direct temperature effect. Emiliania huxleyi is found to grow well at low iron concentrations, in culture experiments.

611

coccoliths precludes the likelihood of single coccoliths sinking at all rapidly under gravity, because of the considerable viscosity of water with respect to such small particles (Stokes’ law). Stokes’ law can be overcome if coccoliths become part of larger aggregates, either marine snow or zooplankton fecal pellets. Another possible fate for coccoliths is to become incorporated into the shells of tintinnid microzooplankton, which when grazing on coccolithophores make use of the coccoliths in their own shells (Figure 4). Regardless of their immediate fate, the coccoliths must eventually either dissolve or else sink toward the seafloor. The construction of CaCO3 coccoliths (calcification) leads to additional impacts, over and above those associated with the photosynthesis carried out by all species. The first and perhaps the most important of these is that CaCO3 contains carbon and the vertical downward flux of coccoliths thereby removes carbon from the surface oceans. It might be expected that this would lead to additional removal of CO2 from the atmosphere to the oceans, to replace that taken up into coccoliths, but in fact, because of the complex effect of calcification (CaCO3 synthesis) on seawater chemistry, the production of coccoliths actually increases the partial pressure of CO2 in surface seawater and promotes outgassing rather than ingassing. Determining the exact nature and magnitude of the overall net effect is complicated by a possible additional role of coccoliths as ‘ballast’ (coccoliths are denser than water and hence when

Biogeochemical Impacts Coccolithophores assimilate carbon during photosynthesis, leading to similar biogeochemical impacts to other phytoplankton that do not possess mineral shells. They also, however, assimilate carbon into biomass. Following death, some of the coccolith CaCO3 dissolves in the surface waters inhabited by coccolithophores, with the rest of the coccolith CaCO3 sinking out of the surface waters within zooplankton fecal pellets or marine snow aggregates. The exact means by which some coccoliths are dissolved in near-surface waters are unclear (dissolution within zooplankton guts may be important), but regardless of mechanisms several lines of evidence suggest that near-surface dissolution does occur. The size of

Figure 4 Tintinnid lorica (casing) with embedded coccoliths.


612

COCCOLITHOPHORES

incorporated into aggregates of particulate fecal material may drag down extra organic carbon into the ocean interior). Microscopic examination of seafloor sediments (if shallow enough that the CaCO3 does not dissolve) and of material caught in sediment traps has revealed that much of the calcium carbonate in the samples consists of coccoliths. The flux of coccoliths probably accounts for c. 50% of the total vertical CaCO3 flux in open ocean waters (in other words, about 50% of the inorganic carbon pump), with foraminifera shells responsible for most of the rest. It is usually not the most numerous species (E. huxleyi) but rather larger species (e.g., Calcidiscus quadriperforatus and Coccolithus pelagicus) that make the greatest contributions to the total coccolith flux. Coccolithophores also impact on climate in other ways, ones that are unconnected with carbon. Coccolithophores are intense producers of a chemical called dimethylsulfoniopropionate (DMSP). The production of DMSP leads eventually (via several chemical transformations) to additional cloud condensation nuclei in the atmosphere and thereby to increased cloud cover. Coccoliths also scatter light, polarizing it in the process. They do not reflect or block light (this would clearly be disadvantageous for the photosynthetic cell underneath), but the difference between the refractive indices of water and of calcium carbonate means that the trajectories of photons are deflected by encounters with coccoliths. A small proportion of the scattering (deflection) events are through angles greater than 901, leading to photons being deflected into upward directions and eventually passing back out through the sea surface. Because of this light-scattering property of coccoliths, their bulk effect is to make the global oceans slightly brighter than they would otherwise be. It has been calculated that the Earth would become slightly dimmer (the albedo of the Earth would decrease by about 0.1% from its average global value of about 30%) were coccolithophores to disappear from the oceans. The effect of coccoliths in enhancing water brightness is seen in its most extreme form during coccolithophore blooms (Figure 3).

The Past Coccolithophores are currently the dominant type of calcifying phytoplankton, but further back in time there were other abundant calcifiying phytoplankton, for instance the nannoconids, which may or may not have been coccolithophores. The fossil

calcifying phytoplankton are referred to collectively as calcareous nannoplankton. The first calcareous nannoplankton are seen in the fossil record c. 225 Ma, in the late Triassic period. Abundance and biodiversity increased slowly over time, although they were at first restricted to shallow seas. During the early Cretaceous (145–100 Ma), calcareous nannoplankton also colonized the open ocean. They reached their peak, both in terms of abundance and number of different species (different morphotypes) in the late Cretaceous (100–65 Mya). ‘The Chalk’ was formed at this time, consisting of thick beds of calcium carbonate, predominantly coccoliths. Thick deposits of chalk are most noticeable in various striking sea cliffs, including the white cliffs of Dover in the United Kingdom, and the Isle of Rugen in the Baltic Sea. The chalk deposits were laid down in the shallow seas that were widespread and extensive at that time, because of a high sea level. Calcareous nannoplankton, along with other biological groups, underwent long intervals of slowly but gradually increasing species richness interspersed with occasional extinction events. Their heyday in the late Cretaceous was brought to an abrupt end by the largest extinction event of all at the K/T boundary (65 Ma), at which point B93% of all species (B85% of genera) suddenly went extinct. Although biodiversity recovered rapidly in the early Cenozoic, calcareous nannoplankton have probably never since re-attained their late Cretaceous levels. Because the chemical and isotopic composition of coccoliths is influenced by the chemistry of the seawater that they are synthesized from, coccoliths from ancient sediments have the potential to record details of past environments. Coccoliths are therefore a widely used tool by paleooceanographers attempting to reconstruct the nature of ancient oceans. Some of the various ways in which coccoliths are put to use in interpreting past conditions are as follows: (1) elemental ratios such as Sr/Ca and Mg/Ca are used to infer past seawater chemistry, ocean productivity, and temperatures; (2) the isotopic composition (d13C, d18O) of the calcium carbonate is used to infer past carbon cycling, temperatures, and ice volumes; (3) the species assemblage of coccoliths (some species assemblages are characteristic of eutrophic conditions, some of oligotrophic conditions) is used to infer trophic status and productivity. Some of the organic constituents of coccolithophores are also used for paleoenvironmental reconstructions. In particular, there is a distinctive group of ketones, termed long-chain alkenones, which are specific to one family of coccolithophores and closely related


COCCOLITHOPHORES

haptophytes. These alkenones tend to survive degradation in sediments, and the ratio of one type of alkenones to another (the U37 K index) can be used to estimate past ocean temperatures. Calcareous nannofossils are also extremely useful in determining the age of different layers in cores of ocean sediments (biostratigraphy).

The Future The pH of the oceans is falling (they are becoming increasingly acidic), because of the invasion of fossil fuel-derived CO2 into the oceans. Surface ocean pH has already dropped by 0.1 units and may eventually drop by as much as 0.7 units, compared to preindustrial times, depending on future CO2 emissions. The distribution of dissolved inorganic carbon (DIC) between bicarbonate, carbonate, and dissolved CO2 gas changes with pH in such a way that carbonate ion concentration (and therefore saturation state, O) is decreasing even as DIC is increasing due to the invading anthropogenic CO2. It is predicted that, by the end of this century, carbonate ion concentration and O may have fallen to as little as 50% of preindustrial values. If emissions continue for decades and centuries without regulation then the surface oceans will eventually become undersaturated with respect to calcium carbonate, first with respect to the more soluble aragonite used by corals, and some time later also with respect to the calcite formed by coccolithophores. There has been an increasing appreciation over the last few years that declining saturation states may well have significant impacts on marine life, and, in particular, on marine organisms that synthesize CaCO3. Experiments on different classes of marine calcifiers (CaCO3 synthesizers) have demonstrated a reduction in calcification rate in high CO2 seawater. One such experiment showed a strong decline in coccolithophore calcification rate (and a notable increase in the numbers of malformed coccoliths) at high CO2 (low saturation state), although some other experiments have obtained different results. If coccolithophore biocalcification is controlled by O then the explanation could be linked to the importance of O in controlling inorganic calcification, although coccolithophores calcify intracellularly and so such a link is not guaranteed. At the time of writing, further research is being undertaken to determine whether, as the oceans become more acidic, coccolithophores will continue to be able to synthesize coccoliths and subsequently maintain them against dissolution. Our ability to predict the consequences of ocean acidification on coccolithophores is hampered by our

613

poor understanding of the function of coccoliths (what they are for, and therefore how the cells will be affected by their absence), and also by our poor understanding of the possibilities for evolutionary adaptation to a low-pH ocean. These constraints can be overcome to an extent by examining the geological history of coccolithophores, and their (in)ability to survive previous acid ocean events in Earth history. Although coccoliths (and other calcareous nannofossils) have been widely studied by geologists, it is only recently that there has been a concerted effort to study their species turnover through events in Earth history when the oceans were more acidic than now. Although many authors have taken the success of coccolithophores during the high-CO2 late Cretaceous as reassuring with respect to their future prospects, the reasoning is fallacious. Levels of calcium are thought to have been higher than now during the Cretaceous, and the CCD (the depth at which CaCO3 disappears from sediments due to dissolution, which is a function of deep-water O) was only slightly shallower than today, indicating that Cretaceous seawater conditions were not analogous to those to be expected in a future high-CO2 world. It turns out that coccolithophores survived the Paleocene–Eocene Thermal Maximum event (thought to more closely resemble the predicted future) fairly well, with a modest increase in extinction rates matched by a similar increase in speciation rates. On the other hand, the environmental changes at the Cretaceous–Tertiary boundary (the K/T impact event), which also appears to have induced acidification, led to a mass extinction of 93% of all coccolithophore species, as well as to extinction of many other calcifying marine organisms including ammonites. It is necessary to more accurately characterize the environmental changes that took place across such events, in order to better determine how well they correspond to the ongoing and future ocean acidification.

See also Calcium Carbonates. Phytoplankton Blooms. Phytoplankton Size Structure. Plankton and Climate. Marine Plankton Communities. Protozoa, Planktonic Foraminifera. Small-Scale Physical Processes and Plankton Biology. Benthic Foraminifera.

Further Reading Gibbs SJ, Bown PR, Sessa JA, Bralower TJ, and Wilson PA (2007) Nannoplankton extinction and


614

COCCOLITHOPHORES

origination across the Paleocene–Eocene Thermal Maximum. Science 314: 1770--1773 (doi: 10.1126/ science.1133902). Holligan PM, Fernandez E, Aiken J, et al. (1993) A biogeochemical study of the coccolithophore Emiliania huxleyi in the North Atlantic. Global Biogeochemical Cycles 7: 879--900. Paasche E (2002) A review of the coccolithophorid Emiliania huxleyi (Prymnesiophyceae), with particular reference to growth, coccolith formation, and calcification–photosynthesis interactions. Phycologia 40: 503--529. Poulton AJ, Sanders R, Holligan PM, et al. (2006) Phytoplankton mineralization in the tropical and subtropical Atlantic Ocean. Global Biogeochemical Cycles 20: GB4002 (doi: 10.1029/2006GB002712). Riebesell U, Zonderva I, Rost B, Tortell PD, Zeebe RE, and Morel FMM (2000) Reduced calcification in marine plankton in response to increased atmospheric CO2. Nature 407: 634--637. Thierstein HR and Young JR (eds.) (2004) Coccolithophores: From Molecular Processes to Global Impact. Berlin: Springer. Tyrrell T, Holligan PM, and Mobley CD (1999) Optical impacts of oceanic coccolithophore blooms. Journal of Geophysical Research, Oceans 104: 3223--3241.

Winter A and Siesser WG (eds.) (1994) Coccolithophores. Cambridge, UK: Cambridge University Press. Young JR, Geisen M, Cros L, et al. (2003) Special Issue: A Guide to Extant Coccolithophore Taxonomy. Journal of Nannoplankton Research 1–125.

Relevant Websites http://cics.umd.edu – Blooms of the Coccolithophorid Emiliania huxleyi in Global and US Coastal Waters, CICS. http://www.ucl.ac.uk – Calcareous Nannofossils, MIRACLE, UCL. http://www.nanotax.org – Calcareous Nannofossil Taxonomy. http://www.emidas.org – Electronic Microfossil Image Database System. http://www.noc.soton.ac.uk – Emiliania huxleyi Home Page, National Oceanography Centre, Southampton. http://www.nhm.ac.uk – International Nannoplankton Association page, hosted at Natural History Museum website.


COLD-WATER CORAL REEFS J. M. Roberts, Scottish Association for Marine Science, Oban, UK

now beginning to reveal the true extent of cold-water coral ecosystems (Figure 3).

& 2009 Elsevier Ltd. All rights reserved.

Cold-Water Corals Introduction and Historical Background Corals and coral reefs are not restricted to shallow, tropical waters. Deep-ocean exploration around the world is now revealing coral ecosystems at great depths in the cooler waters of the continental shelf, slope, and seamounts. Here, in permanent darkness and without the algal symbionts (zooxanthellae) of many tropical species, cold-water corals grow to form true deep-water scleractinian reefs or ‘forests’ of flexible gorgonian, black, gold, and bamboo corals. Such corals have been known since the eighteenth century; the Reverend Pontoppidan, Bishop of Bergen, discussed corals as ‘sea-vegetables’ in his 1755 book The Natural History of Norway and Linnaeus subsequently described several cold-water coral species. The following century, the British naturalist Philip Henry Gosse summarized British corals in his 1860 book Actinologica Britannica. A History of the British Sea-Anemones and Corals (Figure 1). Further records and samples were obtained during the pioneering nineteenth century expeditions of HMS Porcupine (1869, 1870) and HMS Challenger (1872–76), and in the first half of the twentieth century scientific dredging studies by Dons and Le Danois identified sizeable coral banks off the Norwegian and Celtic margins, respectively. However, until relatively recently, cold-water coral banks remained best known to fishermen, especially trawlermen, who risked damaging their nets and marked coral areas on fishing charts. In the latter half of the twentieth century, advances in acoustic survey techniques and research submersibles allowed the first mapping and direct observations of coral colonies in deep water. Using the early research submersible Pisces, Wilson described cold-water coral patch development on the Rockall Bank west of the UK and was among the first to show the value of video surveys to document and understand these structurally complex habitats (Figure 2). In the last 20 years, there have been further advances in deep-ocean exploration, notably the development of multibeam echo sounders, remotely operated vehicles, and, most recently, autonomous underwater vehicles that are

Cold-water corals are all members of the phylum Cnidaria and include species belonging to a number of lower taxonomic groups. Among these the colonial scleractinian or stony corals can develop sizable reef frameworks and are the focus of this article. Lophelia pertusa dominates reef frameworks in the Northeast Atlantic where Madrepora oculata is an important secondary species. L. pertusa is also abundant on the other side of the Atlantic Ocean in the US South Atlantic Bight where other framework corals

Figure 1 Color plate from Gosse’s 1860 book Actinologica Britannica, showing a colony of Lophohelia prolifera (Lophelia pertusa) together with cup corals and zoanthids.


615

616

COLD-WATER CORAL REEFS

Figure 2 Lophelia pertusa reef patches on Rockall Bank in the Northeast Atlantic were among the first cold-water coral habitats observed from a manned submersible. (a) The Pisces III submersible in 1973. (b) Live coral framework and surrounding rubble. (c) Large antipatharian coral colony, possibly Parantipathes hirondelle. Images courtesy of Dr. John Wilson. (b) Reproduced from Wilson JB (1979) ‘Patch’ development of the deep-water coral Lophelia pertusa (L.) on Rockall Bank. Journal of the Marine Biological Association of the UK 59: 165–177, with permission of the Marine Biological Association of the UK.

include Enallopsammia profunda, M. oculata, and Solenosmilia variabilis. Along the eastern Florida shelf, the facultatively zooxanthellate coral Oculina varicosa forms banks up to 35 m in height at depths of 70–100 m. S. variabilis is also reported forming tightly branched frameworks on the Little Bahama Bank and Reykjanes Ridge in the Atlantic and on Tasmanian seamounts in the South Pacific. Goniocorella dumosa is only reported from the Southern Hemisphere, where it forms reef frameworks around New Zealand on the Chatham Plateau. Thus, there are just six cold-water scleractinian coral species currently known to form significant reef frameworks in deep water, compared to more than 800 species of shallow reef-building tropical corals. While the scleractinian reef framework-forming species form the basis of this article, other coldwater corals, notably gorgonians, antipatharians, and hydrocorals, can develop dense assemblages that also provide significant structural habitat for other species.

Reef Distribution and Development The robust anastomosing skeletons of colonial coldwater scleractinians produce dense frameworks that over time can develop structures with significant topographic expression from the seafloor that alter

local sedimentary conditions, are subject to the dynamic process of (bio)erosion, and provide habitat to many other species (Figure 4). By these criteria, coldwater scleractinian corals form reefs and these reefs can persist for tens of thousands of years. Cold-water coral reef distribution is controlled by a suite of environmental factors. They are largely restricted to water masses with temperatures of 4–12 1C and salinities of 35 psu. Although they are often generically referred to as deep-sea corals, their wide depth distribution, from shallow fiordic sills at just 40 m to shelf and slope depths of 200–1000 m where the majority of cold-water coral reefs are found, reflects bathyal environments (200–2000 m) and not the abyssal depths (2000–6000 m) more often associated with the term ‘deep sea’. On a global scale, the importance of seawater carbonate chemistry as a control on cold-water coral occurrence is now becoming apparent. Scleractinian corals secrete calcium carbonate skeletons in the form of aragonite. The boundary between saturated and undersaturated seawater, the aragonite saturation horizon (ASH), is relatively deep (42000 m) in the Northeast Atlantic and relatively shallow (50–500 m) in the North Pacific. As shown in Figure 5, there are many records of reef frameworkforming cold-water corals from the Northeast Atlantic and very few from the North Pacific where coral assemblages are dominated by octocorals and



617

Figure 3 Multibeam echosounders are valuable tools for wide area mapping in deep waters. (a) Three-dimensional bathymetry, exaggerated sixfold in the vertical, showing the many seabed mounds formed by Lophelia pertusa reefs of the Mingulay Reef Complex, Northeast Atlantic. (b) Photograph of a L. pertusa colony from these mounds. Reproduced from Roberts JM, Brown CJ, Long D, and Bates CR (2005) Acoustic mapping using a multibeam echosounder reveals cold-water coral reefs and sourrounding habitats. Coral Reefs 24: 654–669.

hydrocorals that do not form robust aragonitic skeletons. Indeed, the scleractinians beneath the ASH from the Aleutian Islands in the North Pacific are dominated by species adapted to low-calciumcarbonate environments and the majority of the hydrocorals found there form calcitic rather than aragonitic coralla. Compellingly, calcite is c. 50% less soluble in seawater than aragonite. In the 1990s, the hydraulic theory was advanced proposing that cold-water coral reefs, notably the well-developed Lophelia reefs of the Norwegian continental margin, were coupled to the geosphere via the seepage of light hydrocarbons. This intriguing idea led to a large research effort searching for evidence that the corals and associated fauna were reliant on local seepage. Despite proximity to seeps in

certain areas, stable isotope analyses of coral tissue reflect material derived from surface productivity and recent ocean drilling through a coral carbonate mound in the Porcupine Seabight failed to find any evidence of gas accumulation or that mound growth had been initiated by local hydrocarbon seepage. Thus while broad trends in the factors controlling cold-water coral reef distribution are becoming apparent, our understanding of their global distribution remains biased to those parts of the world, notably the Northeast and Northwest Atlantic, where most surveys have been carried out. Predictive modeling approaches suggest that suitable conditions for coldwater coral reef development may be found in areas of the continental slope that have not yet been surveyed and mapped in sufficient detail to test for their


618


occurrence. Similarly, only a small proportion of the tens to hundreds of thousands of seamounts that have been estimated to exist have ever been surveyed. Many of the seamounts that have been examined reveal abundant cold-water corals like the S. variabilis reefs on Tasmanian seamounts in the South Pacific or the assemblages of gorgonian, black, and bamboo corals on the Davidson Seamount off Monterey in the East Pacific.

Feeding, Growth, and Reproduction

Figure 4 Cold-water coral reefs form highly complex threedimensional structural habitat. (a) The sloping flank of a giant coral carbonate mound in the Porcupine Seabight, Northeast Atlantic. (b) Dense scleractinian coral framework (Lophelia pertusa and Madrepora oculata) with purple octocorals (probably Anthothelia grandiflora) and glass sponges (probably Aphrocallistes). (c) Live polyps of L. pertusa. Images courtesy of VICTOR-Polarstern cruise ARKXIX/3a, Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung and the Institut Français de Recherche pour l’Exploitation de la Mer.

Cold-water corals are typically reported from areas with locally accelerated currents or from regions offshore where internal tidal waves impinge on the slope and break, thus enhancing mixing and flux of food material from the surface to the seabed. Both productive surface waters and hydrographic conditions that transport this material to the benthos are needed to support cold-water coral growth. The reef framework-forming corals L. pertusa and M. oculata seem able to use both zoo- and phytoplankton and, as with other cnidarians, are likely to feed from a cosmopolitan mix of zooplankton prey, detritus, and dissolved organic matter. Understanding cold-water coral growth has largely been limited to observations derived in two ways from dead coral skeletons. First, coral colonies found on man-made structures can be measured and used to estimate approximate extension rates based on the age of the man-made structure. Second, cycles in the stable isotopes of carbon and oxygen in the coral’s skeleton can be used to infer annual extension rates, although recent work has shown this method is complicated by poor understanding of skeletal banding patterns. Using these approaches to study L. pertusa, annual extension rates of between 0.5 and 3 cm have been derived, with the faster growth rates associated with shallower colonies growing on North Sea oil platforms where enhanced food availability and competition for space may combine to accelerate linear extension rates. To date there have been no detailed reports on cold-water coral calcification rate or mode – a worrying deficiency in a time of predicted climate change and ocean acidification, discussed below. While we lack detailed understanding of coldwater coral growth and calcification, some promising initial progress in unraveling the reproductive ecology of some species has been made. Histological studies show that L. pertusa polyps have separate sexes (gonochoristic) and that in the Northeast Atlantic they produce gametes over winter following phytodetrital flux to the seafloor (Figure 6). Spawning has



70° 0′ 0′′ N

180° 0′ 0′′

619

70° 0′ 0′′ N

50° 0′ 0′′ N

50° 0′ 0′′ N

30° 0′ 0′′ N

30° 0′ 0′′ N

10° 0′ 0′′ N

10° 0′ 0′′ N

10° 0′ 0′′ S

10° 0′ 0′′ S

30° 0′ 0′′ S

30° 0′ 0′′ S

50° 0′ 0′′ S

50° 0′ 0′′ S 70° 0′ 0′′ S

180° 0′ 0′′

70° 0′ 0′′ S

Figure 5 Global distribution of reef framework-forming cold-water corals. Map courtesy of Dr. Max Wisshak, University of Erlangen, and Dr. Andrew Davies, Scottish Association for Marine Science.

Female mesenteries

Vitellogenic oocytes 100 μm

300 μm

Male mesenteries

Spermacysts 300 μm

100 μm

Figure 6 Histological sections showing female and male reproductive tissues of Lophelia pertusa sampled from a North Sea oil platform. Image courtesy of Dr. Rhian Waller, University of Hawaii.

not been observed yet and while no coral larvae have been sampled, the widespread occurrence and rapid colonization of man-made structures both point to a dispersive planula larva capable of remaining competent in the water column for several weeks.

Hidden Diversity and Molecular Genetics Molecular genetic analysis has tremendous potential to reveal both species- and population-level information. Few molecular studies of cold-water corals


620


have been completed and, as with most deep-sea biological research, those that have been attempted have been limited by small sample sizes. In spite of these constraints, interesting patterns are emerging. For example, at a species level, molecular analysis using partial sequences of the mitochondrial 16S ribosomal RNA encoding gene suggests that M. oculata may have been misclassified on the basis of its skeletal morphology. Rather than grouping with the Oculinidae, M. oculata may be more closely related to the Caryophyliidae and Pocilloporidae. On a population level, microsatellite and ribosomal internal transcribed spacer sequence analyses have shown that in the Northeast Atlantic some slope populations of L. pertusa are predominantly clonal and this species forms discrete fiord and shelf populations reflecting geographical isolation in fiord settings. Information like this is vital to develop management strategies to protect and conserve coral populations.

Habitats and Biodiversity Coral reefs are renown for their structural complexity and cold-water coral reefs are no exception. Corals are ecological engineers, their skeletons forming complex three-dimensional structures that provide a multitude of surfaces for attached epifauna and shelter for mobile fauna. Coral frameworks trap resuspended sand and mud, forming sedimentclogged frameworks and providing further niches for

infaunal species. On a larger scale, clear habitat zones develop around a cold-water coral reef. Live coral is largely unfouled by other organisms and supports relatively few species. Over time, as older polyp generations die back and exposed skeleton is (bio)eroded, coral frameworks degrade to a coral rubble apron that can extend for considerable distances downslope at the foot of the reef. The small-scale structural complexity of coral skeletal frameworks and the larger-scale diversity of habitat types combine to support highly diverse associated animal communities. For example, a recent compilation of European studies showed that cold-water coral reefs along the Northeast Atlantic margin supported 1317 other animal species (Figure 7). However, significant gaps in our understanding remain. While it is clear that cold-water coral reefs sustain many species, we have very limited understanding of the functional relationships between these species. Largely because of the great expense and technical difficulties of working on structurally complex habitats in deep water, few examples of the natural history of these systems have been described. To date, only the most ubiquitous relationships have been examined in any detail. Of these the symbiosis formed between reef framework-forming scleractinians and eunicid polychaetes appears particularly significant. In the Northeast Atlantic, the large worm Eunice norvegica is very frequently found with both L. pertusa and M. oculata. The worm develops a fragile parchment tube through the coral

Figure 7 Examples of diverse fauna sampled from a giant coral carbonate mound in the Porcupine Seabight, Northeast Atlantic. (a) Isopod Natatolana borealis. (b) Gastropod Boreotrophon clavatus. (c) Brachiopod Macandrevia cranium. (d) Hydrocoral Pliobothrus symmeticus. Images courtesy of Dr. Lea-Anne Henry, Scottish Association for Marine Science.



framework that becomes calcified by the coral. As well as apparently strengthening the overall framework, recent aquarium observations have shown that the worms repeatedly aggregate small coral colonies. In situ, this behavior could be a significant factor in enhancing patch formation and accelerating reef growth. We also know little of what organisms prey upon cold-water corals with only a few direct submersible observations showing asteroids apparently grazing on live coral colonies. However, some parasitic relationships have now been described from samples recovered from cold-water coral reefs. L. pertusa is parasitized by the foraminiferan Hyrrokkin sarcophaga and gall-forming copepods are sometimes associated with large gorgonian corals such as Paragorgia arborea. It is intriguing to think what other interactions and behaviors might remain to be described and in situ video and photographic records from unobtrusive benthic landers have great potential to provide these observations (Figure 8). There has been great interest in whether cold-water coral reefs form essential fish habitat, for example, providing areas for spawning or nursery areas for juvenile fish. Investigations are again at an early stage but some broad trends are becoming apparent. The degree to which cold-water corals provide habitat to fish seems to depend on the coral habitat in question. Coral assemblages dominated by gorgonian ‘forests’ seem to sustain fish communities similar to those found near seabeds with other structural habitat such as large rocks. In contrast, there is reasonably

621

compelling evidence that the large, structurally complex reefs formed by scleractinians can provide important fish habitat. For example, higher numbers of gravid female redfish (Sebastes marinus) were found associated with large Lophelia reefs than neighboring off-reef areas. However, the studies to date rely on sparse data collected in different ways (e.g., long line catches vs. submersible observations) in different regions, making it hard to draw out clear patterns. Thus although cold-water coral reefs form biodiversity hotspots on the continental slope, it is proving hard to understand their true significance in biogeographic and speciation terms. Studies are typically biased by the methodology used to collect samples and the taxonomic expertise applied to the fauna. Cold-water coral reef biodiversity studies rely on material gathered by trawl, dredge, box core, grab, or on megafaunal descriptions from photographs or lower-resolution video records. Each technique differs in the type of sample it can recover and further biases may be introduced by different sample-processing methods. Despite these frustrations, when the animal communities recovered from cold-water coral reefs are examined by taxonomic specialists, they frequently reveal undescribed species and often new records or range extensions of species known from other areas. For example, one study examining just 11 box core samples from coral carbonate mounds in the Porcupine Seabight reported 10 undescribed species and that coral-rich cores on-mound were 3 times more species-rich than off-mound cores.

Figure 8 Developments in benthic landers and seafloor observatories will allow long-term environmental data recording and unobtrusive observations of cold-water coral reef fauna. (a) A ‘photolander’ deployed at 800 m water depth on a giant coral carbonate mound. (b) Lander photograph showing coral framework, glass sponges (probably Aphrocallistes), and deep-sea red crab (Chaceon affinis). (a) Courtesy of VICTOR-Polarstern cruise ARKXIX/3a, Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung and the Institut Français de Recherche pour l’Exploitation de la Mer.


622


Timescales and Archives As individual animals, corals can be extremely longlived. 14C dating has shown that one colony of the gold coral Gerardia (Zoanthidea) was approximately 1800 years old when collected, making it possibly the oldest marine animal known. Although scleractinian polyps are unlikely to live for more than 10–20 years, reef frameworks can persist for thousands to tens of thousands of years. In the Northeast Atlantic, the ages of scleractinian reefs clearly correspond to glacial history with relatively young reefs (8000 yr BP) found at latitudes affected by Pleistocene glaciation and far older reef frameworks (50 000 yr BP) further south well beyond glacial influence. Initial interpretation of the cores obtained by drilling through a large carbonate mound in the Porcupine Seabight suggest that the entire mound structure has developed over 1.5–2 My with periods of interglacial coral framework growth interspersed with periods of glacial sediments when conditions were unsuitable for corals. An extensive literature now exists, showing the value of shallow, tropical corals as paleoenvironmental archives, and interest in studying historical patterns in ocean temperature and circulation has increased as evidence of anthropogenic climate change grows. Cores extracted from glacial ice sheets provide invaluable high-resolution climate archives but the potential of deep-ocean sediments to reveal similar high-resolution temporal patterns is limited by the mixing action of bioturbating infauna. Recent research has shown that cold-water coral skeletons not only provide a long-lasting archive but one that can be analyzed at high temporal resolution without confounding effects of bioturbation. Studies fall into two categories: those that use coral skeletal chemistry to estimate past seawater temperature and those that use combinations of dating techniques to trace ocean ventilation. The most convincing paleotemperature estimates come from gorgonian and bamboo corals, which, unlike reef framework-forming scleractinians, contain clear skeletal banding patterns that allow a good chronology to be developed (Figure 9). We continue to learn more about the importance of deep-ocean circulation in regulating global climate and once again recent research demonstrates that cold-water corals provide a unique archive of ocean circulation patterns. By dating coral skeletons with both 14C and U/Th techniques, it is possible to discriminate the coral’s actual age from the age of the seawater in which it grew. This is possible because as corals calcify they use dissolved carbon from the ambient seawater, thus providing material that can later be 14C-dated. This means that cold-water coral

skeletons can record ocean ventilation history as water masses of differing 14C age exchange. This approach has so far successfully followed ventilation patterns in the Southern Ocean and North Atlantic and offers great potential to study past ocean circulation at key oceanographic gateways.

Threats As shallow-water fish stocks have diminished on continental shelves around the world, the fishing industry has expanded into deeper slope and even seamount waters. This move, made possible with larger, powerful refrigerated vessels and improved navigational technology (Global Positioning System), has seen the development of a series of boom-and-bust deep-water fisheries for species such as the orange roughy (Hoplostethus atlanticus) around New Zealand, the roundnose grenadier (Coryphaenoides rupestris) in the Northwest Atlantic, marbled rockcod (Notothenia rossii) around Antarctica, and pelagic armorheads (Pseudopentoceros pectoralis) on Pacific seamounts. Trawling for fish in deep waters requires heavier ground gear and large trawl doors that plough across the seafloor and can easily damage epifaunal communities dominated by corals and sponges (Figure 10). Visual and acoustic evidence of damage to cold-water coral habitats has now been recorded from the territorial seas of many nations including the Canada, Ireland, New Zealand, Norway, UK, and USA, and in each case measures to limit or ban trawling in some coral-rich areas have been instigated. However, on the High Seas beyond the jurisdiction of any one nation, deep-water trawling continues without regulation causing unknown damage to benthic communities. As well as direct physical impacts, deep-water trawling disturbs sediment producing a plume that could smother epifauna over a wider area. To date, few studies have examined this wider area impact. While evidence for damage from deep-water trawling is visually clear and now known to be a concern in several regions, the impacts of other human activities are harder to pin down. As with the fishing industry, technological developments and reduced shallow-water reserves have made it economically viable for the oil industry to explore progressively deeper and deeper waters. At the time of writing, there are producing oil fields in the deep waters of the Atlantic Frontier (UK), Campos Basin (Brazil), and Gulf of Mexico (USA), where exploratory drilling has now taken place to depths over 3000 m. For example, in late 2003, the Toledo well was drilled at 3051 m water depth by Chevron Texaco in the Gulf of Mexico.



623

2 mm

30 (1973) 25 (1979)

20 (1984)

15 (1989)

10 (1994)

5 (1999)

Figure 9 Cross section of the gorgonian Primnoa resedaeformis showing clear growth banding. Courtesy of Dr. Owen Sherwood, Memorial University of Newfoundland, reproduced from Sherwood OA, Scott DB, Risk MJ, and Guilderson TP (2005) Radiocarbon evidence for annual growth rings in the deep-sea octocoral Primnoa resedaeformis. Marine Ecology Progress Series 301: 129–134, with permission of Inter-Research.

As with all suspension-feeding invertebrate animals, cold-water corals are vulnerable to increased sediment loads that could smother polyps or even bury whole colonies. Tropical studies have shown corals are vulnerable to discharges, notably the muds and cuttings released during drilling operations, but it is frequently hard to disentangle physical effects of increased particle exposure from any toxic effects of exposure to drill muds or other additives. Intriguingly, both tropical and cold-water corals will settle and grow on producing oil platforms, sometimes close to drilling discharge points. Some of the older platforms in the northern North Sea support large colonies of L. pertusa that must have grown there continuously for over 20 years to have reached their present size

and now form a reproductive population (Figure 6). Visual surveys have shown that coral polyps directly exposed to drill cuttings may be smothered but other polyps, even on the same colonies, that are not directly exposed can survive. This and the relatively small geographical extent of drilling suggest that if these activities are restricted in areas supporting coldwater coral reefs, their impacts will be considerably less than those of deep-water trawling now known to abrade large areas of the slope and to target seamount fish stocks. However, as tropical studies have shown, coral responses to drill discharges are hard to interpret and to date no detailed studies of reef framework-forming cold-water corals exposed to muds and cuttings have been carried out.


624


Figure 10 Fisheries damage to cold-water corals from coral carbonate mounds in the Porcupine Seabight and Porcupine Bank, Northeast Atlantic. (a) Nets and ropes with crushed coral rubble. (b) Closer view showing scavenging crabs. (c) Lost trawl net. (d) Abandoned trawl rope. (a), (b) Reproduced from Grehan et al. (2004) Proceedings, ICES Annual Science Conference, 22–25 Sep., Vigo, Spain, ICES CM 2004/AA07, with permission of International Council for the Exploration of the Sea (ICES). (c), (d) Courtesy of Dr. Anthony Grehan, National University of Ireland; VICTOR-Polarstern cruise ARKXIX/3a, Alfred-Wegener-Institut fu¨r Polar- und Meeresforschung and the Institut Français de Recherche pour l’Exploitation de la Mer.

Deep seabed mining for valuable materials such as manganese nodules or the rich mineral deposits of ridge and vent systems has yet to develop as a commercial proposition, although some forecasts suggest this could happen within the next 20 years. As with trawling and drilling, mining activities would have localized impacts on the area mined but could also disturb a sediment plume that would disperse affecting a wider area. Again, great care is needed to limit the impact of any developments in deep seabed mining on cold-water coral ecosystems. Recent analysis suggests that rising atmospheric carbon dioxide levels will not only cause global warming, but also could cause the most rapid ‘acidification’ of the oceans seen in the last 300 My. Once again, no studies of cold-water coral response to ocean acidification have been carried out but tropical coral calcification could be reduced by over 50% if atmospheric carbon dioxide concentrations doubled. Perhaps of greatest concern are modeled predictions that the depth of the ASH could shallow by several hundred meters in as little as next 50–100 years, leading to concerns that regions currently suitable for cold-water coral reef development will become inhospitable. Corals living in lowered aragonite

saturation states produce weaker skeletons more vulnerable to (bio)erosion; consequently, entire reef systems may shift from a phase of overall growth to erosion with severe implications for habitat integrity.

Conclusions The last decade has seen an explosion of interest in cold-water coral reefs. This article attempts to summarize the many and exciting advances in our understanding of reef development, longevity, and diversity in deep waters while realizing that this work has only just begun. Work to map and characterize cold-water coral reefs is still needed, as shown by the geographical bias in studies so far, and efforts to unify sampling methodologies and species identification are vital. However, in the coming years, we will enter a phase in research on cold-water coral ecosystems that requires a shift away from baseline mapping to one centered on processoriented questions that will examine the factors driving reef development in the deep ocean and allow us to really understand the significance and vulnerability of cold-water coral reefs and develop protected areas for their conservation.



See also Coral Reefs. Deep-Sea Fauna. Deep-Sea Fishes. Deep Submergence, Science of. History of Ocean Sciences. Law of the Sea. Manganese Nodules. Manned Submersibles, Deep Water. Manned Submersibles, Shallow Water. Marine Protected Areas. Past Climate from Corals. Platforms: Autonomous Underwater Vehicles. Platforms: Benthic Flux Landers. Seamounts and Off-Ridge VolcanismSonar Systems. Vehicles for Deep Sea Exploration.

Further Reading Cairns SD (2007) Deep-water corals: An overview with special reference to diversity and distribution of deepwater Scleractinia. Bulletin of Marine Science 81: 311--322. Expedition Scientists (2005) Modern carbonate mounds: Porcupine drilling. Integrated Ocean Drilling Program Report Number 307. Washington, DC: Integrated Ocean Drilling Program. Freiwald A, Fossa˚ JH, Grehan A, Koslow T, and Roberts JM (2004) Cold-Water Coral Reefs. Cambridge, UK: United Nations Environment Programme – World Conservation Monitoring Centre. Freiwald A and Roberts JM (eds.) (2005) Cold-Water Corals and Ecosystems. Berlin: Springer.

625

Gage JD and Tyler PA (1991) Deep-Sea Biology: A Natural History of Organisms at the Deep-Sea Floor. Cambridge, UK: Cambridge University Press. Grehan A, Unnithan V, Wheeler A, et al. (2004) Proceedings, ICES Annual Science Conference, 22–25 Sep., Vigo, Spain, ICES CM 2004/AA07. Roberts JM, Brown CJ, Long D, and Bates CR (2005) Acoustic mapping using multibean echosounder reveals cold-water coral reefs and surrounding habitats. Coral Reefs 24: 654–669. Roberts JM, Wheeler AJ, and Freiwald A (2006) Reefs of the deep: The biology and geology of cold-water coral ecosystems. Science 312: 543--547. Rogers AD (1999) The biology of Lophelia pertusa (LINNAEUS 1758) and other deep-water reef-forming corals and impacts from human activities. International Review of Hydrobiology 4: 315--406. Sherwood OA, Scott DB, Risk MJ, and Guilderson TP (2005) Radiocarbon evidence for annual growth rings in the deep-sea octocoral Primnoa resedaeformis. Marine Ecology Progress Series 301: 129--134. Wilson JB (1979) ‘Patch’ development of the deep-water coral Lophelia pertusa (L.) on Rockall Bank. Journal of the Marine Biological Association of the UK 59: 165--177.

Relevant Website http://www.lophelia.org – Lophelia.org, an information resource on the coldwater coral ecosystems of the deep ocean.


CONSERVATIVE ELEMENTS D. W. Dyrssen, Gothenburg University, Go¨teborg, Sweden Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 499–502, & 2001, Elsevier Ltd.

Introduction If 1 kg of sea water is evaporated and ignited according to a special procedure 35 g of solids are obtained. This is the normal (standard) salinity. Since the salinity is mainly changed by evaporation or by dilution with practically ion-free rain water the composition of the major ions in sea water is not changed by such processes. These constituents are considered to be conservative, and as a consequence their ratios are constant. Thus the concentration of a conservative constituent (element) at a salinity S is obtained by multiplying the values in Table 1 by S/ 35.

Strontium may be determined with three significant figures by various procedures. Fluoride can be determined with three significant figures using a fluoride electrode. Otherwise, the recommended procedure is a spectrophotometric determination with lanthanum alizarin complexone. Sulfate may be determined gravimetrically with a precision of 0.14% using precipitation of barium sulfate. Boric acid, B(OH)3, together with borate, B(OH) 4 , may be determined with three significant figures by the spectrophotometric curcumin method. Alkalinity is determined by titration with hydrochloric acid of the main basic constituent, HCO 3, together with minor basic components such as CO2 3 , 2 B(OH) 4 , SiO(OH)3 , H2PO4 , and HPO4 . It can be determined with four significant figures. When accurate methods are used for the determinations of the main constituents slight deviations from a conservative behavior may be detected. The deviations, which are due to some fundamental processes, will be discussed below.

Determinations Plankton Production

The salinity can be determined with five significant figures from conductivity measurements as well as by potentiometric titration of chloride þ bromide in m g of sea water with v ml of t molar silver nitrate. Thereby the chlorinity is given by:

Plankton production involves the formation of hard parts (biogenic calcium carbonate and biogenic opal) in addition to soft material. The stoichiometry varies around:

Cl ¼ vt 107:87 328:5233=1000m

ðCH2 OÞ106 ðNH3 Þ16 H3 PO4 ðCaCO3 Þ20 ðSiO2 Þ20

where 107.87vt/1000 represents the mass in grams of pure silver that is necessary to precipitate the halogens in 328.5233 g of sea water. The relationship between salinity and chlorinity is:

With this stoichiometry the increase in alkalinity due to the uptake of nitrate

S ¼ 1:80655Cl

is almost balanced by the biogenic formation of calcium carbonate

Sodium cannot be determined with four significant figures and the value in Table 1 has been calculated from the ion balance X X n½Xn n Xnþ ¼ Potassium can be determined gravimetrically with a precision of 0.26% by precipitation with sodium tetraphenylborate. Calcium ( þ strontium) and magnesium can be determined with four significant figures by titration procedures.

626

þ NO 3 þ H þ H2 O ) NH3 ðorgÞ þ 2O2

þ Ca2þ þ HCO 3 ) CaCO3 ðsÞ þ H

However, when the production sinks below the euphotic zone the soft parts deteriorate NH3 ðorgÞ þ 2O2 ) Hþ þ NO 3 þ H2 O lowering the increase in alkalinity due to the dissolution of calcium carbonate CaCO3 ðsÞ þ CO2 þ H2 O ) Ca2þ þ 2HCO 3


CONSERVATIVE ELEMENTS

Table 1 The major constituents of average sea water with a salinity of 35

Table 2 oceans

627

World average major ions and silica drained into the

Constituent

g kg 1 sea water

mol kg 1 sea water

Constituent

Concentration (mmol l 1)

Sodium (Naþ) Potassium (Kþ) Magnesium (Mg2þ) Calcium (Ca2þ) Strontium (Sr2þ) Fluoride (F) Chloride (Cl) Bromide (Br) Sulfate (SO2 4 ) Alkalinity (At) Boron (B)

10.76 0.3992 1.292 0.4128 0.00815 0.00141 19.344 0.06712 2.712 (0.143)a 0.00445

0.4680 0.01021 0.05315 0.01030 0.000093 0.000074 0.54563 0.00084 0.02823 0.00234b 0.000412

Sodium (Naþ) Potassium (Kþ) Calcium (Ca2þ) Magnesium (Mg2þ) Chloride (Cl) Sulfate (SO2 4 ) Hydrogen carbonate (HCO 3) Silica (Si(OH)4)

0.252 0.0486 0.353 0.148 0.192 0.104 0.902 0.198

a b

Calculated from HCO 3 , the principle base. Mol HCl kg1 needed to titrate all bases to pH 4.5.

Ion balance:

P

n½Xnþ ¼

P

n½Xn ¼ 1:302 meq l 1

exchange with the sediments: NaR þ Kþ 3KR þ Naþ

The CO2 is supplied by microbial decomposition of carbohydrates CH2 OðorgÞ þ O2 ) CO2 þ H2 O In the past, the alkalinity (At) was considered to be a conservative property of sea water. With high precision titration techniques variations in AtS/35 can be measured, as well as in (Cat þ Srt)S/35. The most important carbonate-secreting organisms in the oceans are foraminifera, coccolithophorides, and pteropods. The carbonate tests vary in size, appearance, crystal form (calcite or aragonite), and magnesium content. The solubility depends on the depth (pressure), temperature, and concentration of CO2 besides the crystal form. For example, the pteropods which secrete shells of aragonite undergo dissolution at shallower depths than the coccolithophorides which secrete calcite shells.

River Inputs The river inputs into the oceans vary between the oceans. Ten percent of the total inflow of 106 m3 s1 (1 Sverdrup, Sv) flows into the Arctic Ocean, whereby the normalized alkalinity (AtS/35) is increased in the outflow along the east coast of Greenland. The average composition of the major ions in river water is presented in Table 2. The ratios are quite different from the ratios for the major elements in sea water that may be calculated from the concentrations in Table 1. For example, the Na/K ratio in river water is 5.2 while the ratio in sea water is 46. This is most likely due to ion

where R represents an aluminosilicate. The ratio between magnesium and calcium is 0.42 in river water, but 5.2 in sea water. This may be explained by the fact that the organisms use only small amounts of magnesium, while the biogenic formation of calcium carbonate is a major process. The average ratios of SO2 4 /Cl and HCO3 /Cl are 0.54 and 4.7 in river water; and 0.052 and 0.0043 in sea water. Only small amounts of sulfate are used by the organisms to produce essential sulfur-containing compounds, but when sea water reacts with hot basalt in rift zones sulfate is removed (see below). Hydrogen carbonate is removed upon the formation of biogenic calcium carbonate (see above). The rivers also carry clay minerals into the ocean. The cation exchange capacity corresponds to 5.2 1015 meq y1. This may be compared with the river input of cations of 41 1015 meq y1 (1.302 meq l1 from Table 2 in a flow of 106 m3 s1). In the ocean the sodium, potassium, and magnesium displace calcium in the clay minerals by ion exchange. Man is changing the composition of river waters. Besides pollutants and an increase of particulate matter, acid rain and increased concentration of carbon dioxide in the atmosphere causes the following reactions, especially with limestone in southern Europe: CaCO3 ðsÞ þ Hþ ) Ca2þ þ HCO 3 CaCO3 ðsÞ þ CO2 ðgÞ þ H2 O ) Ca2þ þ 2HCO 3 CaCO3 ðsÞ þ CH2 OðorgÞ þ O2 ) Ca2þ þ 2HCO 3 This has caused an increase in the normalized calcium concentration as well as the normalized alkalinity in the Baltic Sea. Weathering of silicate rocks is


628


also dependent on the partial pressure of CO2 in the atmosphere, for example: CaAl2 Si2 O8 ðsÞ þ 6H2 O þ 2CO2 ðgÞ ) Ca2þ þ 2HCO 3 þ 2AlOOHðsÞ þ 2SiðOHÞ4

Hot Vents About 5000 m3 s1 of sea water reacts with fresh hot basalt in the rift zones. Hydrothermal reactions produce large changes in the composition of the sea salts. The reactions lead to a loss of magnesium, sulfate, and fluoride, but to an addition of calcium and potassium. The hydrothermal activity balances some of the river inputs and contributes to maintenance of the steady state and the conservation of the constant composition of sea salts. If the hot vent lies in a depression such as the Atlantis II Deep in the Red Sea, the results of the hydrothermal reactions are very evident (see Figure 1). Water is removed together with magnesium and sulfate, and calcium is released from the basalt. Rock salt (NaCl), anhydrite (CaSO4), and silica (SiO2) are formed and the brine becomes saturated with these solids. Sulfide metals are coprecipitated with iron sulfide (FeS). Brines are also formed upon the formation of sea ice in polar regions; thereby calcium carbonate and calcium sulfate are precipitated. These processes also occur upon evaporation of sea water, e.g. for the production of sodium chloride.

Marine Aerosols Marine aerosols are formed by bursting of air bubbles at the sea’s surface. Although some separations of the

main constituents might occur by this process, the fallout of sea salts over the continents is in the order of 109 tons y1. Since the weight of the sea salts in the oceans is in the order of 13 1017 tons the recycling time would be 1300 million years (My). The residence times calculated from the river inputs are 210 My for sodium and 100 My for chloride. This implies that weathering influences the concentrations of the major ions in the runoff from the continents.

Minor Constituents Some minor constituents show a conservative type of distribution in the oceans. They are the alkali ions þ Liþ, Rbþ, and Csþ, besides MoO2 4 and Tl . Obviously, the organisms only use small amounts of molybdenum. However, in waters with anoxic basins (some fiords and the Black Sea), molybdate is de2 pleted. In addition, WO2 4 , ReO4 , and U(VI) show conservative-type oceanic distributions.

Conclusions The major constituents of sea water are present in practically constant proportions. The main ions include the alkali ions Naþ and Kþ and the alkaline earth ions Mg2þ and Ca2þ, in addition to chloride (Cl) and sulfate (SO2 4 ). Their concentrations are also proportional to the salinity. In spite of the weathering and recycling processes and the hydrothermal reactions with hot basalt in rift zones there are only slight perturbations of the steady state. Deviations from a conservative behavior may be detected by accurate analytical methods and studies of ancient marine waters.

See also

Depth m 1900

2+

Ca

SO 42+

Mg

2+

Cl

Breaking Waves and Near-Surface Turbulence. Calcium Carbonates. Carbon Cycle. Copepods. Dispersion from Hydrothermal Vents. Freshwater Transport and Climate. Hydrothermal Vent Deposits. Hydrothermal Vent Fluids, Chemistry of. Ice–ocean interaction. Iron Fertilization. Nitrogen Cycle. Open Ocean Convection. Bacterioplankton. River Inputs. Stable Carbon Isotope Variations in the Ocean. Water Types and Water Masses.

2000

Cl g kg 0

50

2100 0

50

100 mol kg

100

_1

150

_1

Figure 1 Depth profiles of chlorinity (Cl, g kg1), calcium, magnesium and sulfate (mol kg1) in the Red Sea Atlantis II deep on 22–23 March 1976.

Further Reading Broecker WS and Peng T-H (1982) Tracers in the Sea. Palisades, New York: Lamont-Doherty Geological Observatory.



Chester R (1990) Marine Geochemistry. London: Unwin Hyman. Degens ET and Ross DA (eds.) (1969) Hot Brines and Recent Heavy Metal Deposits in the Red Sea. New York: Springer-Verlag. Dyrssen D and Jagner D (eds.) (1972) The Changing Chemistry of the Oceans. Stockholm: Almqvist Wiksell. Dyrssen D (1993) The Baltic-Kattegat-Skagerrak estuarine system. Estuaries 16: 446--452. Goldberg ED (ed.) (1974) The Sea. Ideas and Observations on Progress in the Study of the Seas. New York: John Wiley Sons.

629

Grasshoff K, Kremling K, and Ehrhardt M (eds.) (1999) Methods of Seawater Analysis 3rd edn. Weinheim: Wiley-VCH. Kremling K and Wilhelm G (1997) Recent increase of calcium concentrations in Baltic Sea waters. Marine Pollution Bulletin 34: 763--767. Libes SM (1992) Marine Biogeochemistry. New York: John Wiley Sons. Riley JP and Chester R (eds.) (1975) Chemical Oceanography, vol. 1, 2nd edn. London: Academic Press Silleń LG (1963) How has the sea water got its present composition. Svensk Kemisk Tidskrift 75: 161--177.


CONTINUOUS PLANKTON RECORDERS A. John, Sir Alister Hardy Foundation for Ocean Science, Plymouth, UK P. C. Reid, SAHFOS, Plymouth, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 502–512, & 2001, Elsevier Ltd.

Introduction The Continuous Plankton Recorder (CPR) survey is a synoptic survey of upper-layer plankton covering much of the northern North Atlantic and North Sea. It is the longest running and the most geographically extensive of any routine biological survey of the oceans. Over 4 000 000 miles of towing have resulted in the analysis of nearly 200 000 samples and the routine identification of over 400 species/groups of plankton. Data from the survey have been used to study biogeography, biodiversity, seasonal and interannual variation, long-term trends, and exceptional events. The value of such an extensive time-series increases as each year’s data are accumulated. Some recognition of the importance of the CPR survey was achieved in 1999 when it was adopted as an integral part of the Initial Observing System of the Global Ocean Observing System (GOOS).

History The CPR prototype was designed by Alister Hardy for operation on the 1925–27 Discovery Expedition to the Antarctic, as a means of overcoming the problem of patchiness in plankton. It consisted of a hollow cylindrical body tapered at each end, weighted at the front and with a diving plane, horizontal tail fins, and a vertical tail fin with a buoyancy chamber on top (Figure 1A). Hardy designed a more compact version with a smaller sampling aperture for use on merchant ships and this was first deployed on a commercial ship in the North Sea in September 1931 (Figure 1B). During the 1980s the design was modified further to include a box-shaped double tail-fin that provides better stability when deployed on the faster merchant ships of today (Figure 1C). The space within this tailfin is used in some machines to accommodate physical sensors and flowmeters. The normal maximum tow distance for a CPR is approximately 450 nautical miles (834 km). By the late 1930s there were seven CPR routes in the North Sea and one in the north-east Atlantic; in 1938 CPRs were towed for over 30 000 miles. After a

630

break for the Second World War, the survey restarted in 1946 and expanded into the eastern North Atlantic. Extension of sampling into the western North Atlantic took place in 1958. The survey reached its greatest extent from 1962 to 1972 when CPRs were towed for at least 120 000 nautical miles annually. Sampling in the western Atlantic, which had been suspended due to funding problems in 1986, recommenced in 1991 and is still ongoing. Figure 2A shows the extent of the survey in 1999. Initially based at the University College of Hull, the survey moved to Leith, Edinburgh in 1950 under the management of the Scottish Marine Biological Association (now the Scottish Association for Marine Science). In 1977 it finally moved to Plymouth as part of the Institute for Marine Environmental Research (now Plymouth Marine Laboratory). After a short period of uncertainty in the late 1980s, when the continuation of the survey was threatened, the Sir Alister Hardy Foundation for Ocean Science (SAHFOS) was formed in November 1990 to operate the survey. Since 1931 more than 200 merchant ships, ocean weather ships, and coastguard cutters – known as ‘ships of opportunity’ – from many nations have towed CPRs in a voluntary capacity to maintain the survey. The Foundation is greatly indebted to the captains and crews of all these towing ships and their shipping and management companies, without whom the survey could not continue. During the 1990s CPRs were towed by SAHFOS in several other areas, including the Mediterranean (1998–99), the Gulf of Guinea (1995–99), the Baltic (1998–99), and the Indian Ocean (1999). A separate survey by the National Oceanic and Atmospheric Administration/National Marine Fisheries Service using CPRs along the east coast of the USA off Narragansett has been running since 1974; CPRs are currently towed on two routes in the Middle Atlantic Bight. Following a successful 2000 mile trial tow in the north-east Pacific from Alaska to California in July–August 1997, a 2-year survey by SAHFOS using CPRs in the north-east Pacific started in March 2000. In addition to five tows per year on the Alaska– California route, there is one 3000 mile tow annually east–west from Vancouver to the north-west Pacific (Figure 2B). A ‘sister’ survey, situated in the Southern Ocean south of Australia between 601E and 1601E, is operated by the Australian Antarctic Division. In this survey CPRs have been deployed since the early 1990s on voyages between Tasmania and stations in the Antarctic.


CONTINUOUS PLANKTON RECORDERS

631

Q R

V H J

T

E F

K G

A

B

D

C

P L

P (C)

(A)

Formalin tank

Propeller shaft Gears

Wire driven fusee Water tunnel

Water & plankton Water exit (B)

Plankton, between two silk bands, stored in tank of formalin

Figure 1 (A) Diagram of the first Continuous Plankton Recorder used on ‘Discovery’. (Reproduced with permission from Hardy, 1967). (B) The ‘old’ CPR, used up to around 1983, showing the internal filtering mechanism. (C) The CPR in current use, with the ‘box’ tail-fin.

As the operator of a long-term international survey, which has sampled in most of the world’s oceans, SAHFOS regularly trains its own staff in plankton identification. In recent years SAHFOS has also trained scientists from the following 10 countries: Benin, Cameroon, Coˆte d’Ivoire, France, Finland, Ghana, Italy, Nigeria, Thailand, and the USA. The Database and Open Access Data Policy

The CPR database is housed on an IBM-compatible PC and stored in a relational Microsoft Access DATABASE system. Spatial and temporal data are stored for every sample analysed by the CPR survey since 1948. This amounts to 4175 000 samples, with around 400 more samples added per month. There are more than two million plankton data points in the database, which also contains

supporting information, including sample locations, dates and times of samples, a taxon catalog, and analyst details. In the near future it will also hold additional conductivity, temperature, and depth (CTD) data. Routine processing procedures ensure that, despite various operational difficulties, the previous year’s data are usually available in the database within 9 months. In 1999 SAHFOS adopted a new open access data policy, i.e. data are freely available to all users worldwide, although a reasonable payment may be incurred for time taken to extract a large amount of data. The only stipulation is that users have to sign a SAHFOS Data Licence Agreement. Details of the database can be found on the web site: http:// www.npm.ac.uk/sahfos/. This site advertises the availability of data and allows requests for data to be made easily.


632


60˚N

60˚N

50˚N

50˚N

40˚N

40˚N

(A)

40˚W

60˚W

20˚W _ 145

_ 150

20˚E

0˚ _ 140

_ 135

_ 130

_ 125

_ 120 60

55 x 5 per year 50 60 45

x 1 per year

50

150

160

170

180

190

200

210

40

220

230

35

(B) Figure 2 (A) CPR routes in 1999/2000. (B) CPR routes in the Pacific Ocean towed in 2000 and 2001.

The CPR Bibliography, which is available on the SAHFOS web site, lists over 500 references using results from the survey. During the early years many of the papers based on CPR data were published in the ‘in-house’ journal Hull Bulletins of Marine Ecology, which from 1953 onwards became the Bulletins of Marine Ecology; this was last published in 1980.

Methods Merchant ships of many nations tow CPRs each month along 20–25 standard routes (Figure 2A) at a

depth of 6–10 m. Water enters the CPR through a 12.7 mm square aperture and travels down a tunnel that expands to a cross-section of 50 100 mm, where it passes through a silk filtering mesh with a mesh size of approximately 280 mm. The movement of the CPR through the water turns a propeller that drives a set of rollers and moves the silk across the tunnel. At the top of the tunnel the filtering silk is joined by a covering silk and both are wound onto a spool located in a storage chamber containing formaldehyde solution. The CPRs are then returned to SAHFOS in Plymouth for examination. The green



coloration of each silk is visually assessed by reference to a standard color scale; this is known as ‘Phytoplankton Color’ and gives a crude measure of total phytoplankton biomass. The silks are then cut into sections corresponding to 10 nautical miles (18.5 km) of tow and are distributed randomly to a team of 10–12 analysts. The volume of water filtered per 10-nautical-mile sample is approximately 3 m3. Phytoplankton, small zooplankton (o2 mm in size) and larger zooplankton (42 mm) are then identified and counted in a three-stage process. Over 400 different taxa are routinely identified during the analysis of samples and the recent expansion of the survey into tropical waters and the Pacific Ocean will certainly increase this figure. A detailed and thorough quality control examination is carried out by the most experienced analyst on the completed analysis data. Apparently anomalous results are rechecked by the original analyst and the data are altered accordingly where necessary. This system ensures consistency of the data and acts as ‘in-service’ training for the less experienced analysts. Instrumentation

On certain routes CPRs carry additional equipment to obtain physical data. In the past temperature has been recorded on certain routes in the North Sea using BrainconTM recording thermographs, prototype electronic packages, and AquapacksTM. Aquapacks record temperature, conductivity, depth, and chlorophyll fluorescence. These are now deployed on CPR routes off the eastern coast of the USA, in the southern Bay of Biscay and, until November 1999, in the Gulf of Guinea. VemcoTM minilogger temperature sensors are used on routes from the UK to Iceland, and from Iceland to Newfoundland. In order to measure flow rate through the CPR, electromagnetic flowmeters are used on some routes. Such recording of key physical and chemical variables simultaneously with abundance of plankton enhances our ability to interpret observed changes in the plankton.

Results and Applications of the Data The long-term time-series of CPR data acts as a baseline against which to measure natural and anthropogenic changes in biogeography, biodiversity, seasonal variation, inter-annual variation, long-term trends, and exceptional events. The results have applications to studies of eutrophication and are increasingly being applied in statistical analysis of plankton populations and modeling. Some examples are given below.

633

Another possible application, in the context of the new Pacific CPR programme, is an inter-comparison with data from the CalCOFI Program, the only other existing decadal-scale survey in the world sampling marine plankton. This survey has taken monthly or quarterly net samples from 1949 to the present over an extensive grid of stations off the west coast of California. In the majority of samples the zooplankton has been measured only as displacement volume, rather than being identified to species, but concurrently measured physical and chemical data are more extensive.

Biogeography of Marine Plankton

Much of the early work of the survey focused on biogeography. Using Principal Component Analysis, Colebrook was able to distinguish five main geographical distribution patterns in the plankton – northern oceanic, southern oceanic, northern intermediate, southern intermediate, and neritic. Two closely related species of calanoid copepod – Calanus finmarchicus and C. helgolandicus – which co-occur in the North Atlantic and are morphologically very similar, show very different distributions (Figure 3). C. finmarchicus is a cold-water species whose center of distribution lies in the north-west Atlantic gyre and the Norwegian Sea (‘northern oceanic’). In contrast, C. helgolandicus is a warm–temperate water species occurring in the Gulf Stream, the Bay of Biscay and the North Sea (‘southern intermediate’). These different distribution patterns are reflected in their life histories; C. finmarchicus overwinters in deep waters off the shelf edge, whereas C. helgolandicus overwinters in shelf waters. A new species of marine diatom, Navicula planamembranacea Hendey, was first described from CPR samples taken in 1962. The species was found to have a wide distribution in the western North Atlantic from Newfoundland to Iceland. An atlas of distribution of 255 species or groups (taxa) of plankton recorded by the CPR survey between 1958 and 1968 was published by the Edinburgh Oceanographic Laboratory in 1973. An updated version of this atlas, covering more than 40 years of CPR data and over 400 taxa, is in preparation.

Phytoplankton, Zooplankton, Herring, Kittiwake Breeding Data, and Weather

A study in the north-eastern North Sea found that patterns of four time-series of marine data and weather showed similar long-term trends. Covering


634

CONTINUOUS PLANKTON RECORDERS 60°W

40°W

20°W

0°

60°N

60°N

50°N

50°N

40°N

40°N

30°N

30°N 60°W

60°W

20°W

0°

20°W

0°

40°W Calanus finmarchicus 0.018 2.928 21.547 40°W

60°N

60°N

50°N

50°N

40°N

40°N

30°N

30°N 60°W

20°W 40°W Calanus helgolandicus 0.000 0.200 1.293

0°

Figure 3 Distribution of Calanus finmarchicus and C. helgolandicus recorded in CPR samples from 1958 to 1994.

the period 1955–87, these trends were found in the abundance of phytoplankton and zooplankton (as measured by the CPR), herring in the northern North Sea, kittiwake breeding success (laying date, clutch size, and number of chicks fledged per pair) at a colony on the north-east coast of England, and the frequency of westerly weather (Figure 4).

The mechanisms behind the parallelism in these data over the 33-year period are still not fully understood. Calanus and the North Atlantic Oscillation

The North Atlantic Oscillation (NAO) is a largescale alternation of atmospheric mass between



635

3 Westerly weather

2 1 0 _1 _2 1950

1960

1970

1980 3

3 Phytoplankton

2

0

0

_1

_1

_2

_2 1960

3

1970

1990

1980

1950

2

1

1

0

0

_1

_1

_2

_2 1960

3

1970

1980

1990

Kittiwake clutch

1950

2

1

1

0

0

_1

_1

_2

_2 1960

1970

1980

1990

1970

1950

1980

1990

Kittiwake laydate

1960

3

2

1950

1960

3

Herring

2

1950

Zooplankton

2 1

1

1950

1990

1970

1980

1990

Kittiwake chicks

1960

1970

1980

1990

Figure 4 Standardized time-series and 5-year running means for frequency of westerly weather, and for abundances of phytoplankton, zooplankton, herring, and three parameters of kittiwake breeding (laying date, clutch size, and number of chicks fledged per pair), from 1955 to 1987. (Reproduced with permission from Aebischer NJ et al. (1990) Nature 347: 753–755.)

subtropical high surface pressure, centred on the Azores, and subpolar low surface pressures, centred on Iceland. The NAO determines the speed and direction of the westerly winds across the North Atlantic, as well as winter sea surface temperature. The NAO index is the difference in normalized sea level pressures between Ponta Delgadas (Azores) and Akureyri (Iceland). There is a close association between the abundance of Calanus finmarchicus and C. helgolandicus in the north-east Atlantic and this index (Figure 5). At times of heightened pressure difference between the Azores and Iceland, i.e. a high, positive NAO index, there is low abundance of C. finmarchicus and high abundance of C. helgolandicus; during a low, negative NAO index the

reverse is true. However, since 1995 this strong Calanus/NAO relationship has broken down and the causes of this are presently unknown. It suggests a change in the nature of the link between climate and plankton in the north-east Atlantic. North Sea Ecosystem Regime Shift

Recent studies have shown changes in CPR Phytoplankton Color, a visual assessment of chlorophyll, for the north-east Atlantic and the North Sea. In the central North Sea and the central north-east Atlantic an increased season length was strikingly evident after the mid-1980s. In contrast, in the north-east Atlantic north of 591N Phytoplankton Color


Log Abundance

636


0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0_

and more southerly tracks of the westerly winds and higher temperatures in western Europe. These changes coincided with a series of other changes that affected the whole North Sea ecosystem, affecting many trophic levels and indicating a regime shift. 1997 1998

1996

_4

6

_2

1999

0

2

4

North Wall of the Gulf Stream and Copepod Numbers

6

NAO index

Figure 5 Annual log abundance of Calanus finmarchicus in the north-east Atlantic Ocean against the NAO winter index for the period 1962–99. (Adapted with permission from Fromentin JM, and Planque B (1996) Marine Ecology Progress Series 134: 111–118.)

declined after the mid-1980s (Figure 6). These changes in part appear to be linked to the recent high positive phase of the NAO index and reflect changes in mixing, current flow, and sea surface temperature. The increase in Phytoplankton Color and phytoplankton season length after 1987 coincided with a large increase in catches of the western stock of horse mackerel Trachurus trachurus in the northern North Sea, apparently connected with the increased transport of Atlantic water into the North Sea. From 1988 onwards the NAO index increased to the highest positive level observed in the twentieth century. Positive NAO anomalies are associated with stronger

Zooplankton populations in the eastern North Atlantic and the North Sea show similar trends to variations in the latitude of the north wall of the Gulf Stream, as measured by the Gulf Stream North Wall (GSNW) index, which is statistically related to the NAO 2 years previously. Figure 7 shows the close correlation between total copepods in the central North Sea and the GSNW index. This relationship is also evident in zooplankton in freshwater lakes and in the productivity of terrestrial environments, indicating a possible climatic control. Biodiversity

Analyses of long-term trends in biodiversity of zooplankton in CPR samples indicate increases in diversity in the northern North Sea. This may be related to distributions altering in response to climatic change as geographical variation in biodiversity of the plankton shows generally higher diversity at low latitudes than NNE Atlantic

CNE Atlantic

C North Sea 1995

1995

1995

1990

1990

1990

1985

1985

1985

1980

1980

1980

1975

1975

1975

1970

1970

1970

1965

1965

1965

1960

1960

1960

1955

1955

1955

1950

1950

1950

0

4 Months

8

12

0

4

8

12

Months

0

4

8

12

Months

Figure 6 Contour plots of mean monthly Phytoplankton Color during 1948–95 for the central North Sea, and for the central and northern north-east Atlantic. (Reproduced with permission from Reid PC et al. (1998) Nature 391: 546.)



637

Total copepods Central North Sea 1.1 r = 0.64

1.0

1

0.9 0.8

0

0.7 _1

Log10 abundance

Arbitrary units

2

0.6

_2

0.5 1965

1970

1975

1980

1985

1990

1995

2000

Figure 7 The latitude of the Gulf Stream (the GSNW index ‘arbitrary units’, broken line) compared with the abundance of total copepods in the central North Sea (solid line). Adapted with permission from Taylor AH et al. (1992) Journal of Mar. Biol. Ass. UK 72: 919–921.

60˚

N

50˚

N

40˚

N

40˚W

20˚W

Taxonomic richness per CPR sample

1

1.5

2

2.5

3

3.5

4

Figure 8 The biodiversity (taxonomic richness) of calanoid copepods in the CPR sampling area. (Adapted with permission from Beaugrand et al. (2000) Marine Ecology Progress Series 204: 299–303.)

at high latitudes. Calanoid copepods are the dominant zooplankton group in the North Atlantic and the large data set from the CPR survey has been used to map their diversity. This has demonstrated a pronounced local spatial variability in biodiversity. Higher

diversity was found in the Gulf Stream extension, the Bay of Biscay, and along the southern part of the European shelf. Cold water south of Greenland, east of Canada, and west of Norway was found to have the lowest diversity (Figure 8).



Area 1

Monitoring for Nonindigenous and Harmful Algal Blooms

0.11 0.04 _ 0.03

58

62

66 Ye 70 7 ar 4

Area 2

78

0.30

82

86

0.18 0.06 _ 0.06 54

58

62

66 Ye 70 7 ar 4 78

Area 3 0.07

82

86

0.03

3 5 M 79 on 1 th 1

50

1

46

_ 0.01 54

58

62

66 Ye 70 ar 74

Area 4 0.39

78

82

0.17

86

3 5 M 79 on 1 th 1

50

1

46

_ 0.05 54

58

62

66 Ye 70 74 ar

Area 5 0.11

78

0.07

82

86

3 5 M 79 on 1 th 1

50

1

46

0.03 46

50

54

58

62 66 Ye 70 ar 74

78

82

86

3 5 M 79 on 1 th 1

_ 0.01

Exceptional Events

Doliolids are indicators of oceanic water and in CPR samples are normally found to the west and southwest of the British Isles; they occur only sporadically in the North Sea and are rarely recorded in the central or southern North Sea. On two occasions in recent years, in October–December 1989 and September–October 1997, the doliolid Doliolum nationalis was recorded in CPR samples taken in the

54

3 5 M 79 on 1 th 1

50

1

46

Presence

The regularity of sampling by the CPR enables it to detect changes in plankton communities. Few case histories exist that describe the initial appearance and subsequent geographical spread of nonindigenous species. In 1977 the large diatom Coscinodiscus wailesii was recorded for the first time off Plymouth, when mucilage containing this species was found to be clogging fishing nets. C. wailesii was previously known only from southern California, the Red Sea, and the South China Sea and it is believed that it arrived in European waters via ships’ ballast water. Since then the species has spread throughout north-west European waters and has become an important contributor to North Sea phytoplankton biomass, particularly in autumn and winter. Such introduced species can, on occasions, have considerable ecological and economic effects on regional ecosystems. There has been an apparent worldwide increase in the number of recorded harmful algal blooms and the CPR survey is ideally placed to monitor such events. The serious outbreak of paralytic shellfish poisoning that occurred in 1968 on the north-east coast of England was shown by CPR sampling to have been caused by the dinoflagellate Alexandrium tamarense. Increased nutrient inputs into the North Sea since the 1950s have been linked with an apparent increase in the haptophycean alga Phaeocystis, particularly in Continental coastal regions where it produces large accumulations of foam on beaches. In contrast, long-term records (1946–87) from the CPR survey, which samples away from coastal areas, show that Phaeocystis has declined considerably in the open-sea areas of the north-east Atlantic and the North Sea (Figure 9). It is notable that the decline occurred both in areas not subject to anthropogenic nutrient inputs (Areas 1 and 2, west of the UK) and in the most affected area (Area 4, the southern North Sea). This decrease in Phaeocystis up to 1980 is also shown by many other species of plankton, suggesting a common causal relationship.

1

638

Figure 9 Presence of Phaeocystis in five areas of the northeast Atlantic Ocean and the North Sea. Data are plotted for each month for 1946–87 inclusive. (Reproduced with permission from Owens NJP et al. (1989) Journal of Mar. Biol. Ass. UK 69: 813–821.)



German Bight, accompanied by other oceanic indicator species, suggesting a strong influx of north-east Atlantic water into the North Sea. Both these occasions coincided with higher than average sea surface temperature and salinities.

Summary and the Future

639

Gelatinous Zooplankton. Large Marine Ecosystems. North Atlantic Oscillation (NAO). North Sea Circulation. Pelagic Fishes. Phytoplankton Blooms. Plankton. Plankton and Climate. Protozoa, Planktonic Foraminifera. Protozoa, Radiolarians. Satellite Remote Sensing: Ocean Color. Satellite Remote Sensing of Sea Surface Temperatures. Shelf Sea and Slope Sea Fronts.

The long-term time-series of CPR data have been used in many different ways:

• • • • •

mapping the geographical distribution of plankton a baseline against which to measure natural and anthropogenically forced change, including eutrophication and climate change linking of plankton and environmental forcing detecting exceptional events in the sea monitoring for newly introduced and potentially harmful species.

In the future new applications of CPR data may include:

• • • • •

use as ‘sea-truthing’ for satellites regional assessment of plankton biodiversity regional studies of responses to climate change as input variables to predictive modeling for fish stock and ecosystem management for construction and validation of new models comparing ecosystems of different regional seas.

The CPR survey has gathered nearly 70 years of data on marine plankton throughout the North Atlantic Ocean, and has recently extended into the North Pacific Ocean. Alister Hardy’s simple concept in the 1920s has succeeded in providing us with a unique and valuable long-term data set. There is increasing worldwide concern about anthropogenic effects on the marine ecosystem, including eutrophication, overfishing, pollution, and global warming. The data in the CPR time-series is being used more and more widely to investigate these problems and now plays a significant role in our understanding of global ocean and climate change.

Further Reading Colebrook JM (1960) Continuous Plankton Records: methods of analysis, 1950–59. Bulletins of Marine Ecology 5: 51--64. Gamble JC (1994) Long-term planktonic time series as monitors of marine environmental change. In: Leigh RA and Johnston AE (eds.) Long-term Experiments in Agricultural and Ecological Sciences, pp. 365--386. Wallingford: CAB International. Glover RS (1967) The continuous plankton recorder survey of the North Atlantic. Symp. Zoological Society of London 19: 189--210. Hardy AC (1939) Ecological investigations with the Continuous Plankton Recorder: object, plan and methods. Hull Bulletins of Marine Ecology 1: 1--57. Hardy AC (1956) The Open Sea: Its Natural History. Part 1: The World of Plankton. London: Collins. Hardy AC (1967) Great Waters. London: Collins. IOC and SAHFOS (1991) Monitoring the Health of the Ocean: Defining the Role of the Continuous Plankton Recorder in Global Ecosystem Studies. Paris: UNESCO. Oceanographic Laboratory, Edinburgh (1973) Continuous plankton records: a plankton atlas of the North Atlantic and the North Sea. Bulletins of Marine Ecology 7: 1--174. Reid PC, Planque B, and Edwards M (1998) Is observed variability in the observed long-term results of the Continuous Plankton Recorder survey a response to climate change? Fisheries Oceanography 7: 282--288. Warner AJ and Hays GC (1994) Sampling by the Continuous Plankton Recorder survey. Progress in Oceanography 34: 237--256.

See also Copepods. Diversity of Marine Species. Ecosystem Effects of Fishing. Eutrophication. Fish Larvae. Florida Current, Gulf Stream and Labrador Current.


COPEPODS R. Harris, Plymouth Marine Laboratory, Plymouth, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 512–523, & 2001, Elsevier Ltd.

and Misophrioida are primarily benthopelagic groups, the latter having two pelagic species. The Poecilostomatoida and Siphonostomatoida are commensal or parasitic groups. Finally, the Monstrilloidaare exclusively marine, with parasitic juveniles, but a pelagic adult stage.

Introduction Copepods are microscopic members of the phylum Crustacea, the taxonomic group that includes crabs, shrimps and lobsters and is the only large class of arthropods that is primarily aquatic. The name copepod comes from the Greek words kope (an oar) and podos (foot), the majority of members of the group having five pairs of flat paddle like swimming legs. About 10 000 species are currently known, and their numerical dominance as members of the marine plankton means that they are probably the most numerous metazoan – multicellular – animals on earth. In addition to forming a major component of marine plankton communities, copepods are also found in sea-bottom sediments, as well as associated with many marine plants and animals. They play a pivotal role in marine ecosystems by controlling phytoplankton production through grazing, and by providing a major food source for larval and juvenile fish. This article will place particular emphasis on the dominant group of planktonic copepods, known as the Calanoida (Figure 1), playing a central role in these processes inthe world’s oceans.

Taxonomy There are 10 taxonomic orders of copepods, of which 9 have marine representatives. Of these the most important marine orders are the Calanoida, Cyclopoida, and Harpacticoida. Calanoid copepods are primarily pelagic, 75% of the known species are marine, and some are benthopelagic or commensal. The group includes the species Calanus finmarchicus (Gunnerus), a dominant component of North Atlantic boreal ecosystems, first named nearly 250 years ago as Monoculus finmarchicus by Johan Ernst Gunnerus, Bishop of Trondheim in Norway (Figure 2). The Cyclopoida include pelagic commensal and parasitic species (Figure 3). Harpacticoid copepods are predominantly marine, with only 10% of species being freshwater. Most are benthic, with a few pelagic and commensal representatives, they represent the most abundant component of the meiofauna after nematode worms. The Platycopoida

640

Morphology Most copepods are small, requiring study with a microscope. Small planktonic cyclopoids may be only 0.2 mm long and similarly harpacticoid copepods found in the interstitial space of sandy sediments are among the smallest Metazoa. In contrast, some large deep-sea calanoids, such as Valdiviella, may exceed 20 mm in length. Calanus finmarchicus is often said to be about the size of a grain of rice (Figure 2). Parasitic forms are generally larger than the free living copepods. For example, species of the genus Penella, which is parasitic on fish and whales, may be over 30 cm in length. The body of a free-living copepod (Figure 4) is normally cylindrical, and is distinctly segmented. The head, which is the site of the median naupliar eye, is either rounded or may bear a pointed rostrum. The presence of at least two pairs of swimming legs is characteristic of nearly all copepods at some stage in their life cycle. Similarly, antennules with up to 27 (Figure 5) segments are general in the order, though segmentation may often be reduced. The body is divided into the prosome, which may be further subdivided into the cephalosome and metasome, and the urosome. The feeding appendages are on the cephalosome; in the calanoids these comprise, from anterior to posterior, the antennule, antenna, mandible, maxillulle, maxilla, and maxilliped (Figure 6). The swimming legs are attached to the metasome in adult calanoids, one pair for each of the five segments of the metasome. The urosome contains the genital and anal segments, and ends with the furca orcaudal rami, a series of spines or fine hairs. Most copepods are pale and transparent, though some species, particularly those living at the sea surface or in the deep sea, may be pigmented blue, red, orange, or black. The early developmental stages in the life history are the nauplii (Figure 7), with reduced numbers of appendages. In calanoids there are six naupliar stages, NI to NVI. Copepods, like all other Crustacea, molt by shedding their exoskeleton as they grow. Hence there is amolt between each naupliar


COPEPODS

641

(E) (D) (C)

(F)

(A)

(B)

(G)

(I) (H) (J)

(K)

(L)

Figure 1 The diversity of calanoid body form. (A) Diaixidae; (B) Calocalanidae; (C) Acartiidae; (D) Pseudocyclopidae; (E) Augaptilidae; (F) Pontellidae; (G) Metridinidae; (H) Eucalanidae; (I) Stephidae; (J) Euchaetidae; (K) Temoridae; (L) Calocalanidae. (Permission from Huys and Boxshall, 1991.)


642

COPEPODS

stage. Molting from NVI involves a radical change inmorphology (metamorphosis) to the first copepodite stage. The copepodites, of which there are normally six stages (CI–CVI) are like small adults, and gradually develop adult characteristics during successive molts. The adult stage is CVI, and no further molts occur.

Distribution and Habitats

Figure 2 Gunnerus’ sketches of Calanus. The smallest shows the natural size. (Permission from Marshall and Orr, 1955.)

(D)

(E)

As has already been noted, copepods are probably the most numerous multicellular organisms on earth. They are found throughout the marine and estuarine environments of the world’s oceans. Species inhabiting coastal and brackish waters have wider tolerances of environmental variables than, for example, deep sea copepods, which are specifically adapted to the special conditions of this environment. Generally, copepods are more abundant in coastal and productive upwelling environments than

(F)

(G)

(C)

(B)

(A)

Figure 3 The diversity of cyclopoid body form. (A) Cyclopidae; (B) Cyclopinidae; (C) Oithonidae; (D) Thespessiopsyllidae; (E) Asidicolidae; (F) Archinotodelphyidae; (G) Mantridae. (Permission from Huys and Boxshall, 1991.)


643

COPEPODS

rostrum antennule labrum antenna labium mandible maxillule maxilla maxilliped

Cephalosome (Ce) Prosome (Pr) Metasome (Me)

Urosome (Ur)

1

f.o. (A1)

m.e.

r.f.

(A2) b.r. es.

(Md) (Mx1) (Mx2) (Mxp)

leg 1

(P1)

2

leg 2

(P2)

3 4 5

leg 3

(P3)

leg 4 leg 5 genital somite

(P4) (P5)

mo. mx.g.

a.

g.

(Gn)

anal somite

odi.

o. v.n.c.

h.

furca or caudal ramus o.s. Figure 4 Diagrammatic illustration of the external morphology and appendages of a female calanoid copepod. The metasome has five clearly defined segments, numbered 1–5; this species has five pairs of swimming legs and so these five metasome segments are synonymous with pedigerous segments 1–5. Legs 1–5 are the swimming legs. (Permission from Mauchline, 1998.)

in the oligotrophic open ocean. Over the deep ocean, where the water column may extend to 8000 m, the abundance of copepods is highest in the surface layers, and then decreases almost exponentially. The number of species occurring in a particular environment varies. In some, for example brackishtidepools, a single species may dominate. In contrast, assemblages in the open ocean normally exceed 100 species (Figure 8). In addition to those that dominate the plankton, copepods also live in marine sediments, forming a major component of the meiofauna. They are found in all sediments from muds to coarse sands, and from the intertidal zone to the deep ocean. Harpactocoids are the dominant copepod component of the meiofauna. This group is also abundant on intertidal and subtidal macroalgae. Apart from free-living planktonic and benthic forms, almost half of the described species live in association with other marine animals. Copepods parasitize almost every phylum of marine animals, many as ectoparasites living on the external body surface, though others have exploited, for example, the internal surfaces of the gills of fish. In the majority of cases it is the adult copepods that are parasites, but the Monstrilloida are an exception, as the naupliar stages are internal parasites of polychaete worms and gastropod mollusks. The adults live in the plankton, but do not feed. Specialized habitats include marine caves that are home to a number of platycopoid and calanoid

od. sp.

an.

Figure 5 Diagram of the internal anatomy of a female Calanus from the side. a., aorta; an., anus; br., brain; f.o., frontal organ; g., gut; h., heart; m.e., median eye; mo., mouth; mx.g., maxillary gland; o., ovary; o.di., oviducal diverticula; od., oviduct; es., esophagous; o.s., oil sac; r.f., rostral filament; sp., spermathecal sac; v.n.c. ventral nerve cord. (Permission from Marshall and Orr, 1955.)

species, all living in association with the bottom sediments. Other such hyperbenthic copepods, living close to the sediment surface, are also found throughout shallow and deep seas. Deep-sea hydrothermal vents also have an associated copepod fauna, which is only now being described. Other interfaces in the marine environment that provide specialized habitats for copepods are the under ice environment in Polar regions, and the sea surface. The under-ice habitat supports a rich growth of microalgae, and in turn this food source is exploited by a large number of copepods. The sea surface habitat is that of the neuston, the group of animals and plants living in the extreme surface film. The calanoids of the family Pontellidae, such as members of the genera Pontella and Anomalocera are the commonest neustonic copepods. Many have


644

COPEPODS

gnb

B C

epi

C B

art

e

ex ex

en en

en

ex Maxillule Mandible Antenna Swimming leg Maxilliped

Antennule

C

PC

B 1

C Maxilla

2 1 PC

e

3 2

B

C B

3 en

e en

en

Figure 6 Diagrammatic representations of the appendages of a calanoid copepod. The swimming legs usually have developed endopods and exopods with upto three segments, numbered 1–3 here. art, arthrite; B, basis; C, coxa;e, endite; en, endopod; epi, epipodite; ex, exopod; gnb, gnathobase; PC, praecoxa. (Permission from Mauchline, 1998.)

100 µm (A) (B)

300 µm

d ae

(C)

(D)

(E) v Figure 7 Nauplii of calanoid copepods. (A) Clausocalanus furcatus, stage I (NI); (B) Paracalanus aculeatus, NV; (C) Rhincalanus cornutus, NIV; (D) Euchaeta marina, NVI; (E) antennule showing dorsal (d) and ventral (v) setae and terminal aesthetasc (ae). (Permission from Mauchline, 1998.)

strong blue pigmentation, which may be associated with protection against surface ultraviolet radiation, and also attachment structures on the back of the head by which the copepod suspends attached to the

surface film. A few species can move with such vigor that they can hurl themselves out of the water, and a shoal of these creatures can appear like a rain shower on the surface of the sea.


Species abundance (No./1000 m 3)

COPEPODS

645

(C)

1000 (B)

800

(A)

600 400 200 0

0

20

40

60 80 100 120 140 160 180 Rank of species

Figure 8 Abundances of copepod species in the open Pacific Ocean showing the species order. Data points are the overall means. Gray bars show the range of seven individual cruise mean abundances per species. (Permission from McGowan JA and Walker PW (1985) Dominance and diversity maintenance in an oceanic ecosystem. Ecological Monographs 55: 113–118.)

Feeding The majority of planktonic copepods were originally thought to be exclusively herbivorous, filtering phytoplankton fromsea water with the fine hairs of the mouth parts. In contrast, carnivorous copepods have more robust spines on the mouth parts. More recently it has been appreciated that many copepods are omnivores, feeding on a wide range of naturally occurring particulate material, phytoplankton, small planktonic animals of the microzooplankton, and detritus. The feeding appendages are the antennules, antennae, mandibles, maxillules and maxillae (Figure 5). These are often considerably reduced in adult males, which may not feed. The mouth parts of ectoparasites are adapted for piercing or sucking. Internal parasites have often lost their mouth parts, and food is absorbed directly from the host. Among planktonic calanoids there are three general mouth part patterns, related to feeding ecology: the true filter-feeders, the omnivores, and the true carnivores. The antennules are particularly involved in carnivorous feeding, having sensory organs that function in prey detection. Spacing between the hairs (the setae) of the maxillae has been considered to indicate the size of particles that can be filtered by a copepod (Figure 9). However, the model of copepod filter-feeding as a mechanical process depending on the morphological characteristics of the maxillae is no longer accepted. Direct studies of feeding behavior using high-speed microcinematography and video observations have shown that feeding behavior is complex, taking account of the viscous, low-Reynolds-number, environment that these small organisms inhabit. Feeding behavior, and adaptations of the appendages,

Figure 9 Left maxilla of Calanus helgolandicus female from the right. A, B and C represent the sizes of three algal cells: (A) Nannochloris oculata; (B) Syracosphaera elongata; (C) Chaetoceros decipiens. (Permission from Marshall SM and Orr AP (1956) On the biology of Calanus finmarchicus IX. Feeding and digestion in the young stages. Journal of the Marine Biological Association of the United Kingdom 35: 587–603.)

enable copepods to exploit particles such as detritus and phytoplankton, a few micrometers in size, while at the other extreme they can feed on other members of the zooplankton such as other copepods, chaetognaths, and fish larvae. Particles may be rejected during the feeding process, resulting in food selectivity. The feeding rate of planktonic copepods is dependent on type and size of food particle (Figure 10), as well as environmental factors such as temperature, light, and turbulence. The latter, in particular, can affect therate of encounter between predator and prey.

Growth and Development Copepods grow by molting, as do all other Crustacea. Normally the nauplius stage NI hatches from the egg; naupliar growth involves five molts to the sixth nauplius (NVI), and then after metamorphosis to copepodite stage one (CI) there are a further five molts until the adult, CVI stage, is reached. In a few groups the egg hatches directly into one of the later naupliar stages, for example NII. The development rate of copepod eggs is dependent on temperature within any one species. The relationship between development time D (days) and


COPEPODS

250 Generation time D (days)

1.0 T. fluviatilis C. angstii C. eccentricus Centric sp. 0 10

200 150 100 50 0

5

0

5

10

15 20 T (˚C)

25

30

35

0

5

10

15 20 T (˚C)

25

30

35

(A)

F (mL h

_1

1.2 _

Specific growth rate g (d 1)

per copepod)

_ I (µg h 1 per copepod)

646

0

0

400 600 200 _ Carbon concentration (µg l 1)

800

Figure 10 The effect of size (species) and concentration (as carbon) of food particles on ingestion rate, I, and volume swept clear, F, of adult females of Calanus. (Permission from Frost BW (1972) Effects of size and concentration of food particles on the feeding behavior of the marine planktonic copepod Calanus pacificus. Limnology and Oceanography 17: 805–815.)

1 0.8 0.6 0.4 0.2 0

temperature T (1C) is generally described by the empirical equation [1], in which a and b are fitted constants. D ¼ aðT aÞb

½1

Egg development times of egg sac-carrying groups are longer than those of free spawners. A number of models of development have beenapplied to the naupliar and copepodite stages of copepods. Equiproportional development considers that the duration of each developmental stage is proportional to the egg development time, determined by the equation above, at the same temperature. The is ochronal model of development describes those species for which all stages have almost the same duration, and development proceeds linearly with time. In sigmoidal development, the development rate of the early naupliar stages is significantly slower, and the later copepodite stages also have a longer relative development duration. Growth rates of copepods are temperaturedependent, and are most usefully expressed as the weight-specific growth rate (per day, d1), which is the increase in body weight per day as a proportion of the body weight of the developmental stage being considered (Figure 11). An adequate food supply, both quantitative and qualitative, is clearly necessary for proper development and growth. Ultimate body

(B)

Figure 11 (A) The generation D time (days) of different species of copepods related to environmental temperature T (1C). The relationship is described by the equation D ¼ 128.8 e0.120T (B) The specific growth rate g (d1) of species of copepods calculated from the weight of egg and the adult and the generation time of each species and related to environmental temperature T (1C). The equation for the relationship is g ¼ 0.0445e0.111T. (Permission from Huntley ME and Lopez MDG (1992) Temperature-dependent production of marine copepods: a global synthesis. American Naturalist 140: 201– 242.)

size, either as length or weight, is dependent on both temperature and food conditions. It has been suggested that for small planktonic copepods growth is optimized, and food is utilized more efficiently, at higher temperatures, whereas larger forms optimize growth and food utilization at lower temperature. This may explain some aspects of the geographical and vertical distribution patterns of copepods.

Metabolism Copepods have a variety of digestive enzymes in the gut, and there are both diel and seasonal changes in enzyme activity. In particular, overwintering animals in diapause may have considerably reduced digestive enzyme activity. The proportion of the ingested food


COPEPODS

that is assimilated, and therefore available for subsequent metabolism, ranges from 60% to 90% in herbivores, the remaining 10% to 40% being released as fecal pellets. The soluble excretory products of metabolism are generally excreted as ammonia or urea and dissolved phosphorus compounds, and this process is important in nutrient regeneration cycles supporting phytoplankton growth in marine ecosystems. There are no gills in free-living copepods, and respiratory exchange is supported by direct uptake of dissolved oxygen from sea water. Apart from in the calanoids and some parasitic species there is no heart or circulatory system. Both respiration and excretion are closely coupled to feeding activity and often exhibit diel cycles.

Reproduction Mating behavior follows a generally similar sequence in all copepods. Initially the male is attracted to the female, often by chemical attractants, pheromones (Figure 12), Then the male captures the female, adjusts to the mating position, and finally transfers

Search

Dance

647

and attaches a package of sperm, the spermatophore, to the female. Most species of planktonic calanoids lay their eggs directly into the water. However, harpacticoids and cyclopoids usually carry the eggs in a single or paired egg sacs and a number of calanoid genera, for example Euchaeta, Eurytemora, and Pseudocalanus also carry egg sacs. Individual eggs are usually spherical, ranging in size from 0.2 to over 0.6 mm, the eggs within egg sacs often being relatively larger than those that are freely spawned. Females of some freely spawning species, when well fed, produce over 100 eggs in a day. The daily rates of egg production expressed as a proportion of the female body weight are around 0.17 for copepods carrying their eggs, and 0.2 for free spawners. The lifetime fecundity of egg sac-bearers is lower than that of the free spawners; the latter have been observed in laboratory studies to produce over 2000 eggs in a female’s lifetime. A number of groups, including calanoids ofthe families Acartiidae, Centropagidae, Temoridae, and Pontellidae, produce resting eggs, often distinguishable from the normal eggs by having a thicker outer coating. These diapause eggs sink to the seabed and may become buried in bottom sediments until conditions are appropriate for hatching. It has been estimated that diapause eggs may remain viable in sediment, capable of hatching, for up to 40 years.

Male

Behavior Chase

Clasp and transfer

Approach and escape Female

Figure 12 Calanus marshallae. A conceptual interpretation of mate-attraction–mate-search behavior. The sequence ofevents is (1) a female generates a vertical pheromone trail; (2) a male is alerted by pheromone to females in the general vicinity and swims in smooth loops of mostly horizontal orientation; (3) on crossing a pheromone trail, the male performs a dance (or sometimes does not); (4) the male chases down the pheromone trail to the female; (5) the female jumps away repeatedly with the male pursuing, sometimes bumping her; and (6) a mating clasp is established and a spermatophore is transferred from the male to the female. (Permission from Tsuda A and Miller CB (1998) Mate-finding behavior in Calanus marshallae. Philosophical Transactions of the Royal Society of London, series B 353: 713–720.)

Perhaps the most striking aspect of the behavior of planktonic copepods is that of diel vertical migration. This behavior, characteristic of most planktonic organisms, involves the population remaining at depth during the daytime. As night falls, the copepods actively migrate upward to spend some hours in the surface during the hours of darkness, before descending at dawn to the original daytime depth (Figure 13). Although this is a general phenomenon, there are many variations, depending both on species, and the influence of environmental factors. Light is the dominant environmental factor controlling diel vertical migration, with populations following diel changes in light intensity, isolumes (layers of constant light intensity). Predator avoidance related to light is thought to be one of the major adaptive advantages of diel vertical migration. By migrating to deeper darker layers by day, copepods minimize mortality from visual predators, in particular fish. Predator avoidance has to be balanced against the need to feed and, as the phytoplankton is concentrated in the surface layers, migration to the surface by night is generally associated with active


648

COPEPODS

Day 1 A B

Night 1 A B

Day 2 A B

A

Night 2 B

Depth (m)

0 50 100 150 120 m

_3

200

0

Depth (m)

50 100 150 _3

200

100 m

Figure 13 Vertical distribution of adult females of Calanus pacificus (A) and Metridia lucens (B), 5 and 6 August 1986. (Permission from Dagg MJ, Frost BW and Walser WE Jr (1989) Copepod diel migration, feeding, and the vertical flux of pheopigments. Limnology and Oceanography 34: 1062–1071.)

feeding, diel cycles of digestive enzyme activity, and diel feeding rhythms. Where invertebrate predators detecting prey nonvisually are dominant, the phasing of migration may be reversed. Vertical migratory behavior involves active swimming. Most copepods swim by rapid beating of the appendages, the antennae, mandibular palps, the maxillules, and the maxillae. In some species of planktonic calanoids, such as the genera Metridia, Centropages, and Temora, these movements result in a smooth continuous swimming behavior. In others, periods of active swimming are interspersed with inactivity when the animal sinks. This hop-and-sink behavior is characteristic of Calanus finmarchicus. Rapid jumping, often as apredator-avoidance behavior, involves strong strokes of the antennules and the swimming legs. This results in very rapid jumps, which propel the copepod several body lengths from the source of stimulus. The benthic harpacticoids and some cyclopoids crawl over or burrow through sediment. The thoracic limbs are used in crawling, and this is accompanied in harpacticoids, by sideways undulations of the body. Nauplii use three pairs of appendages in swimming: the antenules, the antennae, and the mandibles. Three swimming behaviors have been recognized; a slow gliding motion propelled by the

antennae and mandibles; a rapid darting behavior driven by all three pairs of appendages beating together; a cruise and pause behavior. Swimming and feeding behaviors are interdependent in herbivores, omnivores, and carnivores. Swimming speeds of planktonic calanoids generally range from 1 to 20 mm s1 which is equivalent to 1–5 body lengths per second. Estimates based on field studies of oceanic diel vertical migrators, such as Pleuromamma, range from 10 to 50 mm s1, representing the ability of such copepods to migrate at rates in excess of 100 m h1. The predominant sensory mechanisms are mechanoreception and chemoreception, and receptor structures are found on the antennules. The antennules, particularly of males, are covered with sensory structures, aesthetascs, which are important in detecting water movement, food, predators, and potential mates. Detection of mechanical stimuli appears only to operate over short distances, often less than one body length. Chemoreception probably operates over longer distances and is involved in mate detection and response to food concentrations and to predators. Many planktonic copepods are bioluminescent. The families Megacalanidae, Lucicutiidae, Heterohabdidae, Augaptilidae, and Metridinidae have luminescent glands that produce luminous glandular secretions. The number of light organs varies from 10 to 70, and they may be distributed widely over the body surface. The function of copepod bioluminescence is not certain. It may deter predators in the dark water column of the deep sea, and may act as a warning signal between individuals of the same species.

Life Histories Copepods inhabit a wide range of environments, from the tropics to the Polar regions, and from the intertidal zone to the deep ocean, and their life histories accordingly vary considerably. In the tropics and subtropics there is no seasonality in breeding, most species breeding continuously. An exception to this pattern occurs in the upwelling system off the Gulf of Guinea, and in the Benguela Current off south-west Africa, where the dominant calanoid, Calanoides carinatus enters a diapause resting stage, at copepodite stage CV, at the end of the cold season and sinks to colder water until the next season. At high latitudes, diapause and overwintering strategies are the dominant responses to the highly seasonal environment (Figure 14). Breeding periods are restricted, often only one generation is produced each year, and growth and development rates are slowed. The most common diapause stage is copepodite CV.


649

COPEPODS

A

M

J

J

A

S

O

N

D

J

F

M

RECEPTION OF DIAPAUSE STIMULUS ?

?

CI CII CIII CIV CV

Depth

STAGE

INDUCTION

REFRACTORY

ACTIVATION

TERMINATION

INSECTS

ACCUMULATION OF FOOD RESERVES

GRADUAL CESSATION OF DNA, RNA SYNTHESES; LOWERING OF METABOLISM

ACTIVATION PROCESS; SUPERCOOLING REACTIONS

INDIVIDUALS WITH POTENCY TO DEVELOP; GENERAL PHYSIOLOGY AS BEFORE

ENDOCRINE REACTIVATION; BIOCHEMICAL READJUSTMENTS FOR DNA, RNA SYNTHESES

CALANUS FINM.

ARRESTED DEVELOPEMENT; LIPID ACCUMULATION; INGESTION STOP; DESCENT

REDUCED DIGESTIVE TRACT; REDUCED METABOLISM; INGESTION STOP

RESISTANCE TO ANAEROBIOSIS; TORPIDITY; INCAPACITY TO DEVELOP

GONADOGENESIS; MOULTING; COMPETENCY TO ASCENT DEVELOP

CYCLOPOIDS

INTESTINE EMPTIED; INCREASE OF NUMBER AND SIZE OF OIL GLOBULES; ACCUMULATION OF INSTARS ABOVE THE BOTTOM SEDIMENTS

INCREASE IN DORMANCY

DEEP TORPOR; DEGREE OF INCAPACITY TO TORPIDITY DIMINISHED DEVELOP; LOW METABOLIC RATE; ENHANCED ABILITY TO SURVIVE ANAEROBIOSIS; REDUCTION OF INTESTINAL EPITHELIUM

PHYSIOLOGICAL PATTERNS

DIAPAUSE PHASES

PREPARATORY

? ALL RESTING STAGES LEAVING THE SEDIMENTS

Figure 14 Generalized pattern of seasonal ontogenetic migration and physiological changes during overwintering of Calanus finmarchicus in relation to diapause phases in insects and comparison with insect and cyclopoid diapause. (Permission from Hirche HJ (1996) Diapause in the marine copepod, Calanus finmarchicus – a review. Ophelia 44: 129–143.)

Animals in this state overwinter in deep water with delayed development and reduced respiration and excretion rates, usually do not feed, and often show reduced digestive enzyme activity and changes in the digestive epithelium of the gut. Metabolism is sustained by the extensive lipid reserves, which in highlatitude copepods may exceed 75% of the total body weight, and these reserves give the body a brilliant red coloration in some species. Lipid stores can fuel egg laying in the spring before the spring phytoplankton bloom in species such as Cala nus glacialis and Calanus hyperboreus, ensuring that the resultant nauplii are able to exploit the spring pulse of phytoplankton production. In the bathypelagic environment of the deepsea below 500 m, copepods do not undertake diel vertical migration, and there is reduced seasonality with depth. The life histories of deep sea copepods are

relatively little known, but the majority probably breed continuously throughout the year, with slow development rates and long generation times, reflecting the low-temperature environment.

Copepods as Pests Some copepods are economically important pests. An example is the salmon louse Lepeophtheirus salmonis (Krøyer), which may have significant impact on the economics of salmon aquaculture. The copepods breed rapidly in the high fish densities of the salmon cages and may kill the fish either directly or by causing skin damage that in turn makes the fish susceptible to disease. Gill-parasitic copepods such as Lernaeocera may have significant effects on commercial fish species, and the shell fish parasites such as Myticola intestinalis may also have economic effects.


650

COPEPODS

Biogeochemical Role The production of fecal pellets by copepodsis an important source of sedimented material for benthic organisms and plays a significant role in nutrient cycling and in vertical flux of biogenic elements to the deep ocean. Fecal pellet production rates of actively feeding copepods may exceed ten pellets per hour. Such rates, combined with the abundance of copepods in some ecosystems, mean that a significant component of the small particulate food captured is transformed into much larger packages represented by the fecal pellets. These may have sinking rates greater than 100 md1, the rate being dependent both on size and composition, derived from the diet, of the pellets. Many of these rapidly sinking pellets may exit the surface layers of the ocean, and either reach the sea floor of the continental shelves or enter the bathypelagic zone. This pellet flux is sogreat that sinking pellets may form a significant part of the diet of other members of the plankton, including copepods, and of benthic organisms.

Role in the Ecosystem Planktonic copepods, through their grazing activity, are one of the major controls on the growth of phytoplankton, and quantitative understanding of grazing processes is central to modeling marine ecosystem dynamics. Similarly, copepods play a pivotal role in nutrient cycles, by excreting dissolved nitrogen and phosphorus compounds, which are then utilized by phytoplankton to support growth, and hence primary production. Pelagic cyclopoid and calanoid copepods form the first link in the marine food chain that leads from the single-celled plants of the phytoplankton to the fishes and marine mammals that form the exploitable living resources of the world’s oceans. Nauplii through to adult stages of copepods are the typical food of nearly all larvae of commercially exploited marine fish. Some adult fish, such as herring, continue to feed on them. Similarly, the harpacticoid copepods of the meiofauna are a food source for bottom-feeding flatfish. Copepods have been subjected to limited commercial exploitation. Although they are extremely abundant, their small size makes direct harvesting impractical. Limited fisheries for Calanus species, in areas of high coastal abundance, have provided

dietary supplements for salmon aquaculture and for pet food.

See also Biogeochemical Data Assimilation. Carbon Cycle. Continuous Plankton Recorders. Fish Feeding and Foraging. Fish Larvae. Fish Migration, Vertical. Gelatinous Zooplankton. Lagrangian Biological Models. Large Marine Ecosystems. Meiobenthos. Nitrogen Cycle. Ocean Gyre Ecosystems. Optical Particle Characterization. Particle Aggregation Dynamics. Plankton. Plankton and Climate. Polar Ecosystems. Population Dynamics Models. Temporal Variability of Particle Flux. Upwelling Ecosystems. Zooplankton Sampling with Nets and Trawls.

Further Reading Boxshall GA and Schminke HK (eds.) (1988) Biology of Copepods. Dordrecht: Kluwer. Corner EDS and O’Hara SCM (eds.) (1986) The Biological Chemistry of Marine Copepods. Oxford: Clarendon Press. Ferrari FD and Bradley BP (1994) Ecology and Morphology of Copepods. Dordrecht: Kluwer. Gotto RV (1979) The association of copepods with marine invertebrates. Advances in Marine Biology 16: 1--109. Hardy A (1956) The Open Sea: Its Natural History, Part 1: The World of Plankton. London: Collins. Harris RP (ed.) (1995) Zooplankton Production. ICES Journal of Marine Science 52: 261--773. Harris RP, Wiebe PH, Lenz J, Skjoldal HR, and Huntley M (eds.) (2000) ICES Zooplankton Methodology Manual. London: Academic Press. Huys R and Boxshall GA (1991) Copepod Evolution. London: The Ray Society. Kerfoot CW (1980) Evolution and Ecology of Zooplankton Communities. Hanover, NH: University Press of New England. Marshall SM (1973) Respiration and feeding in marine copepods. Advances in Marine Biology 11: 57--120. Marshall SM and Orr AP (1955) The Biology of a Marine Copepod, Calanus finmarchicus (Gunnerus). London: Oliver and Boyd. Mauchline J (1998) The biology of calanoid copepods. In: Blaxter JHS, Southward AJ, and Tyler PA (eds.) Advances in Marine Biology, 33, 1--710. Raymont JEG (1983) Plankton and Productivity in the Oceans, 2nd edn; vol. 2, Zooplankton. Oxford: Pergamon Press.


CORAL REEF AND OTHER TROPICAL FISHERIES V. Christensen, University of British Columbia, Vancouver, BC, Canada D. Pauly, University of British Columbia, Vancouver, BC, Canada Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 534–538, & 2001, Elsevier Ltd.

Introduction Until recently studying and reporting on tropical fisheries tended to be done in the context of development aid projects initiated with the perception that the transfer of technology, management approaches, and scientific models to tropical countries would assist them in a way that would help to raise the standard of living and the food supply. Many of these aims have been achieved, although not necessarily through such North/South transfer but rather through local growth of the preexisting fisheries, and through access rights granted to distant water fleets of developed countries to operate in the waters of developing countries. These developments have turned tropical and reef fisheries from the marginal activities they were in the 1960s and 1970s to key players in international fisheries. Catches in tropical and subtropical fisheries presently exceed those in developed countries, as do exports from developing to developed countries – which dwarf exports in the opposite direction. In the process tropical fisheries have become globalized to a much further extent than other food commodities such as rice, for example, which is overwhelmingly consumed locally. The experience gained in developing, managing, and studying developing country fisheries have in the process become part of the mainstream of fisheries research. Some of these findings are presented below to characterize trends in tropical and coral reef fisheries, which have paralleled or in some case preceded developments in higher latitude fisheries.

Coral Reef and Tropical Fisheries Fisheries Development

Through the first half of the last century the tropical seas were only lightly exploited by humans, even if coastal areas were harvested on a small scale using sustainable methods and often with locally based regulation methods in place. However, during the inter-war period many colonial states tried to

introduce more industrial fishing methods to increase productivity and food supply. In general, the early attempts were not particularly successful, an exception being the introduction of trawling in the South China Sea area by the Japanese in the late 1920s. After the Second World War surplus engines and crafts allowed for an expansion of effort in many areas and the following decades saw the introduction of trawling in most of the coastal shelf areas of the tropics. Stagnating catches and signs of overfishing soon followed the expansion of trawling; an early example of this came from the Manila Bay in the Philippines where trawling was introduced shortly after the war, and where overfishing was apparent by the late 1950s. The classical, well-documented case of how fisheries ‘development’ with the introduction of trawling quickly leads to stagnating catches in spite of continuing build-up of effort comes from the Gulf of Thailand. Here a development project in the early part of the 1960s initiated a demersal trawl fishery in hitherto unexploited parts of the Gulf. This led to a rapid build-up of commercial fisheries in the mid1960s, and has often been hailed as a leading example of a successful development project. Fortunately, a stratified research survey series has been in place in the Gulf continuously since 1966 documenting the ecological changes brought about by the fisheries. Initially, the catches increased steadily with effort, but stagnated after a decade and since then they have remained at about the same level in spite of increasing effort. At first glance this may be indicative of a sustainable fishery – after all the catch level has been maintained for several decades. However, a closer look shows a different and unfortunately very typical picture. As the fishing pressure increased, the longer-lived species rapidly declined, many by a factor of 10 or more within a decade. Sharks and the larger rays were among the first to disappear because of their low fecundity and long life spans, while the dominant group (by weight) in the ecosystem changed from ponyfishes to squids within a decade from the onset of the trawl fisheries. The average trophic level of the catches, as well as of the ecosystem resources in the Gulf, has decreased steadily as fishing pressure mounted. Hence, catch level may be maintained but what is being extracted is smaller and generally lower-value catches; the average fish is finger-sized. The groupers, snappers, and most other high-value


651

652

CORAL REEF AND OTHER TROPICAL FISHERIES

species are gone, and the catches include some 40% of ‘trashfish’, i.e. small species and juveniles of larger, commercial species. The trashfish are being used to produce fish oil and meal mainly to supply a growing aquaculture industry. The degradation of the Gulf of Thailand ecosystem as described is typical of what is happening to the fisheries of the world because of overexploitation, and is indicative of a process now generally being called ‘fishing down the food web’. This term recognizes how we change ecosystems by systematically eradicating the upper trophic levels, how we follow up by removing the more intermediate levels to end up catching squids, shrimps, and other organisms low in the food web. This may be considered an economically interesting alternative to the unexploited state given the high value of squids and shrimps, but a closer examination shows that it is a dangerous path, beset with increased risk of unwanted structural changes in the underlying ecosystems. The scientific and practical challenge lies in balancing the harvesting so as to maintain healthy ecosystems. Also, the path does not recognize the growing public concern for marine ecosystems – they may be largely out of sight but they are no longer out of mind. The Ecosystem Perspective

Fisheries in temperate areas are often described in the form of their target, e.g. herring or cod fisheries. In tropical fisheries the taxonomic diversity, until recently not fully mastered by fish taxonomists (but see FishBase at //www.fishbase.org), along with the unselective nature of the fishing gear used (e.g. lift-nets or trawls in shallow waters), result in a widely diverse catch, often comprising hundreds of species in a single haul. This precludes the notion of using single-species management procedures, which so far have dominated the management of temperate fisheries. From the onset, management in the tropics had to be based on the yields of aggregate of species of target groups (e.g. groupers), implicitly or explicitly taking account of their biological interactions. In contrast, in temperate waters managers have only recently fully realized the consequences of the nonselective nature of the gear used, and thus the ecological consequences of their impact through the food webs. The trend in fisheries research in recent years has been toward incorporating an ecosystem perspective into the assessment and management of living aquatic resources. This is done in recognition both of the fact that there is biological interaction among the resources (‘fish eat fish’) and acceptance that exploitation has consequences not just for the exploited resource but for their predators, competitors, and

prey as well. Of especial significance here is what may be termed ‘charismatic’ organisms, notably marine mammals, sea birds, and turtles. As the environmental movement has gained strength over the past decades it is becoming increasingly clear that there are more players to be recognized than the hunters and gatherers. Examples are the ‘dolphin-free’ tuna stamp now required to export tuna to the USA, and the turtle-excluding devices required for shrimp fisheries in order to maintain export markets. Incorporating the ecosystem perspective into management has been an arduous task that has kept the fisheries research community challenged for decades. The pioneers in the field were E. Ursin working in the North Sea and T. Laevastu in the north-east Pacific back in the 1970s. Since then the major steps in the northern temperate areas have been focused on ‘multispecies virtual population analysis’ (MSVPA), based on single species assessment methodologies but incorporating biological interaction. Building directly on traditional assessment methodologies, MSVPA has had a wide support base in the fisheries assessment circles of the North, but it has not been widely used, partly because of a heavy price tag caused by its extensive data requirements, partly because it was designed to include only the commercial fish species, and partly because its ability to address ecosystemlevel questions is limited. In the tropics the development has taken a different route. A US fisheries scientist, J. Polovina, who had the task of developing an ecosystem model (Ecopath) for a Hawaiian reef system, was inspired by Laevastu’s approach and developed a simple version of it incorporating all important functional groups of the ecosystem studied. The model required comparatively little information for parameterization and its ecological book-keeping system ensured that gross errors were likely to be recognized, hence providing a feedback to the biological sampling system. In the nearly two decades since it was first described, the Ecopath approach has been developed considerably by a growing team of scientists, and it has now reached a level where it is being used for ecosystem-based fisheries management in many countries. The approach (and the software that makes it operational, freely available from http:// www.ecopath.org) is easy to adapt to new areas, as illustrated by the fact that it is being used in more than 100 countries in almost equal measures in temperate and tropical countries. Fisheries Affect the Physical Environment

Many of traditional small-scale fisheries previously prevailing in the developing countries of the inter-



tropical belt were benign, i.e. they did not modify the habitat in which the resources they used were embedded. The introduction of industrialized fishing methods and the modernization of local techniques both impacted local ecosystems very severely. Thus trawling in the narrow coastal shelf destroys structured habitats with sponges and soft coral to an extent that was not visible in the fisheries of the higher latitudes (where benthic habitats are also destroyed). Furthermore, coral reefs, which are important in the inter-tropical belt, are now exploited by extremely destructive methods such as dynamite and cyanide fishing; the latter usually to catch live large fish for restaurants or for the ornamental trade, both of which support valuable export sectors. Here again it is the visibility of the impacts that have subsequently led in the North to reassessment of the physical impact of fisheries on habitat. An example of this is the growing concern about how trawling impacts ecosystems; for instance, how heavy beam trawls have eradicated cold-water coral reef structures. Globalization and Poverty Jointly Destroy Fisheries Resources

The bulk of the countries of the inter-tropical belt do not have the administrative structure that would make it possible to limit entry into fisheries. The result is that millions of landless farmers and other rural poor have become small-scale fishers in recent decades – the fisheries being the last resort for livelihood. These fishers have no tradition of restrained and community control of effort as they start to compete with more established fishers. Their lack of access to capital forces them to use cheap and often, unfortunately, destructive methods – most notably explosives and poisons (cyanide or insecticides). Growing populations with a need for cheap fish provide an outlet for the low quality products that emerge from these activities, while a largely insatiable world market (notably Western Europe, USA, Japan, and China) absorbs the high quality products, especially expensive invertebrates such as shrimps and sea urchins. This massive uncontrolled inflow of effort leads to what has been called ‘Malthusian overfishing’, wherein extreme poverty reduces the cost of labor (usually a limiting factor in fisheries) to virtually nothing; thus making fishing ‘profitable’ even when productivity is very low. Adult fishers are often seen operating over an entire day to land only one pound of small fish per person. In such cases poverty and other nonfishers in the family subsidize the fishery. Other subsidies impacting small-scale fisheries and leading to gross overexploitation are provided by

653

development banks that continue to encourage increasing fishing effort, and the access rights that are negotiated by powerful developed countries for operating distant water fleets in the tropics, e.g. for trawling off west Africa by EU vessels or tuna fishing in the Indian and Pacific Oceans. Such developments, which have contributed to an enormous excess capacity in the developing world, have also contributed to making the subsidy issue more visible in developing countries, where attempts to address these problems have so far been stymied by shipbuilding and other powerful lobbies. There is, however, a growing realization of the need to reduce the overcapitalization, as perhaps best exemplified by the growing reluctance of development agencies to fund ‘fisheries development’ projects due to the recognized high failure rate of such projects.

Marine Protected Areas are Part of the Solution While the foregoing presented a number of problems, which appeared or became visible first in the inter-tropical belt and only later were seen as global problems, this section presents elements of solutions to the global ills, which also have appeared first in the developing world. One of these is marine protected areas (MPAs), first demonstrated in the Philippines to have the potential both to allow resources to regenerate/rebuild themselves and to increase the yield for fishers. Such marine protection implies permanent closure of a significant part of an area traditionally exploited by fisheries, 40% in the case of the island first studied by the pioneering team of Dr Angel Alcala in the Central Philippines. These closures allow the growth of animals, in this case especially groupers and snappers, which would otherwise be caught as juveniles. Their increased biomass and the high percentage of adults within the MPA lead to migration from the MPAs to the surrounding waters, and hence to increased catches in the surrounding areas, and as has been demonstrated, to larger overall catch. As widely noted, MPAs are not a panacea; notably they require that fishing outside the MPAs must be regulated. Even more importantly the communities whose fishing grounds have been impacted by an MPA must accept the gamble that the MPA represents, and must indeed be the advocates and protectors of the MPA. This implies that they have accepted, through fruitful interaction with scientists or other communities that are patrons of an MPA, that their fishing impacts the resource, something still hotly denied by the less enlightened part of the


654


fishing community in many countries. Further, the communities must be self-organized so that they can prevent defection, i.e. fishing within the MPAs for the resources that rapidly build up as the fishing pressure is released. Third, for the MPA to be set in place, the central government for the country in question must have relinquished enough power to the communities for them to be in a position to take pertinent decisions and see them implemented. Finally, for the MPA approach to work fishing communities must have accepted the notion that the immediate gains from exploiting a resource must be put in a longer-term context. Essentially the community must take a conservation-oriented standpoint, i.e. it must overcome the huge chasm which until now has separated fisheries from conservation. Demanding as these requirements may seem, they have been met in many parts of the Philippines, the Caribbean, and other parts of the world – including New Zealand, a developed country. Given the difficulties in implementing other forms of effort reduction measures and the very destructive trends in the areas where MPAs are not implemented it is likely that the trend toward using MPAs as an integral part of sustainable fisheries management will continue. Indicative of this trend is the fact that in his last year of office US President Clinton signed an executive order calling for ‘protection of existing MPAs and the establishment of new MPAs, as appropriate.’

See also Ecosystem Effects of Fishing. Fisheries: Multispecies Dynamics. Fishery Management. Fishing

Methods and Fishing Fleets. Marine Fishery Resources, Global State of. Network Analysis of Food Webs.

Further Reading Alcala AC (1988) Effects of marine reserves on coral fish abundance and yields of Philippine coral reefs. Ambio XVII: 194--199. Christensen V (1996) Managing fisheries involving top predator and prey species components. Reviews in Fish Biology and Fisheries 6: 417--442. Christensen V and Pauly D (eds.) (1993) Trophic Models of Aquatic Ecosystems. ICLARM Conference Proceedings 26. Manila: ICLARM. Hilborn R and Walters CJ (1992) Quantitative Fisheries Stock Assessment, Choices, Dynamics and Uncertainty. New York: Chapman and Hall. Longhurst A and Pauly D (1987) Ecology of Tropical Oceans. San Diego: Academic Press. Munro JL (ed.) (1983) Caribbean Coral Reef Fishery Resources, ICLARM Study Review 7. Manila: ICLARM. Munro JL and Munro PE (eds.) (1994) The Management of Coral Reef Resource Systems. ICLARM Conference Proceedings 44. Manila: ICLARM Studies and Reviews. Pauly D (1979) Theory and Management of Tropical Multispecies Stocks. ICLARM Study Review 1. Manila: ICLARM Studies and Reviews. Pauly D, Christensen V, Dalsgaard A, Froese R, and Torres J (1998) Fishing down marine food webs. Science 279: 860--863. Polunin NVC and Roberts CM (eds.) (1996) Reef Fisheries. London: Chapman and Hall. Safina C (1998) Song for the Blue Ocean. New York: H Holt & Co.


CORAL REEF FISHES M. A. Hixon, Oregon State University, Corvallis, OR, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: Diversity, Distribution, and Conservation Coral reef fishes comprise the most speciose assemblages of vertebrates on the Earth. The variety of shapes, sizes, colors, behavior, and ecology exhibited by reef fishes is amazing. Adult body sizes range from gobies (Gobiidae) less than 1 cm in length to tiger sharks (Carcharhinidae) reportedly over 9 m long. It has been estimated that about 30% of the some 15 000 described species of marine fishes inhabit coral reefs worldwide, and hundreds of species can coexist on the same reef. Taxonomically, reef fishes are dominated by about 30 families, mostly the perciform chaetodontoids (butterflyfish and angelfish families), labroids (damselfish, wrasse, and parrotfish families), gobioids (gobies and related families), and acanthuroids (surgeonfishes and related families). The latitudinal distribution of reef fishes follows that of reef-building corals, which are usually limited to shallow tropical waters bounded by the 20 1C isotherms (roughly between the latitudes of 301 N and S). The longitudinal center of diversity is the IndoAustralasian archipelago of the Indo-Pacific region. Local patterns of diversity are correlated with those of corals, which provide shelter and harbor prey. There is a high degree of endemism in reef fishes, especially on more isolated reefs, and many species (about 9%) have highly restricted geographical ranges. The major human activities that threaten reef fishes include overfishing (especially by destructive fishing practices and live collections for restaurants and aquariums), and habitat destruction, which includes both local effects near human population centers and the ongoing worldwide decline of reefs due to coral bleaching and ocean acidification caused by anthropogenic carbon emissions and global warming. Worldwide, about 31% of coral reef fishes are now considered critically endangered and 24% threatened. The major solution for local conservation is fully protected marine reserves, which have proven effective in replenishing depleted populations.

Fisheries Where unexploited by humans, coral reef fishes typically exhibit high standing stocks, the maximum

being about 240 t km 2 (about 24 t C km 2). High standing crops reflect the high primary productivity of coral reefs, often exceeding 103 g C m 2 yr 1, much of which is consumed directly or indirectly by fishes. Correspondingly, reported fishery yields have reached 44 t km 2 yr 1, with an estimated global potential of 6 Mt yr 1. These fisheries provide food, bait, and live fish for the restaurant and aquarium trades. However, the estimated maximum sustainable yield from shallow areas of actively growing coral reefs is around 20–30 t km 2 yr 1, so many reefs are clearly overexploited. Indeed, overfishing of coral reefs occurs worldwide, due primarily to unregulated multispecies exploitation in developing nations. Few and inadequate stock assessments or other quantitative fishery analyses, susceptibility of fish at spawning aggregations (see below), and destructive fishing practices (including the use of dynamite, cyanide, and bleach) are contributing factors. In the Pacific, some 200–300 reef fish species are taken by fisheries, about 20 of which comprise some 75% of the catch by weight. As fishing intensifies in a given locality, large fishes, especially piscivores (see below), are typically depleted first, followed by less-preferred, smaller, and moreproductive planktivores and benthivores. (Note that fishing of some piscivores is naturally inhibited in some regions by ciguatera fish poisoning, caused by dinoflagellate toxins concentrated in the tissues of some species.) The indirect effects of overfishing include the demise of piscivores, perhaps enhancing local populations of prey species or causing a trophic cascade. Overfishing of urchin-eating species (such as triggerfishes, Balistidae) and various herbivorous fishes may provide sea urchins predatory and competitive release, respectively. Although urchin grazing may help maintain benthic dominance by corals, overabundant urchins may overgraze and bioerode reefs.

Morphology A typical perciform reef fish (virtually an oxymoron) is laterally compressed, with a closed swimbladder and fins positioned in a way that facilitate highly maneuverable slow-speed swimming. Compared to more generalized relatives, reef fishes have a greater proportion of musculature devoted to both locomotion and feeding. Their jaws and pharyngeal apparatus are complex and typically well developed for suction feeding of smaller invertebrate prey, with


655

656

CORAL REEF FISHES

tremendous variation reflecting a wide variety of diets. For example, most butterflyfish (Chaetodontidae) have forceps-like jaws that extract individual polyps from corals, many damselfish (Pomacentridae: e.g., genus Chromis) have highly protrusible jaws that facilitate pipette-like suction feeding of zooplankton, and parrotfishes (Scaridae) have fused beak-like jaw teeth and molar-like pharyngeal teeth enabling some species to excavate algae from dead reef surfaces. (This excavation and subsequent defecation of coral sand can bioerode up to 9 kg of calcium carbonate per square meter annually.) Tetraodontiform reef fishes typically swim relatively slowly with their dorsal and anal fins, and consequently are morphologically well defended from predation by large dorsal–ventral spines (triggerfishes, Balistidae), toxins (puffers, Tetraodontidae), or quill-like scales (porucupinefishes, Diodontidae). The latter two families have fused dentition which is well adapted for consuming hard-shelled invertebrates. Diurnal reef fishes are primarily visual predators. Visual acuity is high and retinal structure indicates color vision. At least some planktivorous damselfishes have ultraviolet-sensitive cones, which may assist in detecting zooplankton by enhancing contrast against background light. Coloration is highly variable (including ultraviolet reflectance), ranging from cryptic to dazzling. Bright ‘poster’ colors are hypothesized to serve as visual signals in aggression, courtship, and other social interactions. Sexually dimorphic coloration is associated with haremic social systems (see below). Nocturnal reef fishes are either visually oriented, having relatively large eyes (e.g., squirrelfishes, Holocentridae), or rely on olfaction (e.g., moray eels, Muraenidae).

Behavior

interaction is not always mutualistic in that cleaners occasionally bite their hosts, and some saber-tooth blennies (Blenniidae) mimic cleaner wrasse and thereby parasitize host fish. Anemonefishes (Pomacentridae, especially the genus Amphiprion) live in a mutualistic association with several genera of large anemones. By circumventing discharge of the cnidarian’s nematocysts, the fish gain protection from predators by hiding in the stinging tentacles of the anemone. In turn, the fish defend their host from butterflyfishes and other predators that attack anemones. However, some host anemones survive well without anemonefish, in which case the relationship is commensal rather than mutualistic. Finally, some gobies cohabit the burrows of digging shrimp. The shrimp provides shared shelter and the goby alerts the shrimp to the presence of predators. Territoriality

The most overt form of competition involves territoriality or defense of all or part of an individual’s home range. Many reef fishes behave aggressively toward members of both their own and other species, but the most obviously territorial species are benthicfeeding damselfishes (Pomacentridae: e.g., genus Stegastes). By pugnaciously defending areas about a meter square from herbivorous fishes, damselfish prevent overgrazing and can thus maintain dense patches of seaweeds. These algal mats serve as a food source for the damselfish as well as habitat for small juvenile fish of various species that manage to avoid eviction. At a local spatial scale, the algal mats can both smother corals as well as maintain high species diversity of seaweeds. By forming dense schools, nonterritorial herbivores (parrotfishes and surgeonfishes) can successfully invade damselfish territories. Piscivory and Defense

Overt behavioral interactions between coral reef fishes include mutualism (when both species benefit), interference competition (often manifested as territoriality), and predator–prey relationships. Mutualism

Three of the best-documented cases of marine mutualism occur in reef fishes. ‘Cleaning symbiosis’ occurs when small microcarnivorous fish consume ectoparasites or necrotic tissue off larger host fish, which often allow cleaners to feed within their mouths and gill cavities. The major cleaners are various gobies (Gobiidae) and wrasses (Labridae). Some of the cleaner wrasses are specialists that maintain fixed cleaning stations regularly visited by hosts, which assume solicitous postures. The

Predation is a major factor affecting the behavior and ecology of reef fishes. There are three main modes of piscivory. Open-water pursuers, such as reef sharks (Carcharhinidae) and jacks (Carangidae), simply overtake their prey with bursts of speed. Bottomoriented stalkers, such as grouper (Serranidae) and trumpetfishes (Aulostomidae), slowly approach their prey before a sudden attack. Bottom-sitting ambushers, such as lizardfishes (Synodontidae) and anglerfishes (Antennariidae), sit and wait cryptically for prey to approach them. The vision of piscivores is often suited for crepuscular twilight, when the vision of their prey is least acute (being adapted for either diurnal or nocturnal foraging). Hence, many prey species are inactive during dawn and dusk, resulting in crepuscular ‘quiet periods’ when both diurnal and


CORAL REEF FISHES

nocturnal species shelter in the reef framework. (Parrotfishes may further secrete mucous cocoons around themselves at night, and small wrasses may bury in the sand.) Otherwise, prey defensive behavior when foraging or resting typically involves remaining warily near structural shelter and shoaling either within or among species. Associated with day–night shifts in activity are daily migrations between safe resting areas and relatively exposed feeding areas. Caribbean grunts (Haemulidae) spend the day schooling inactively on reefs, and after dusk migrate to nearby seagrass beds and feed. Reproducing reef fishes may avoid predation by spawning (in some combination) offshore, in midwater, or at night. Spawning during ebbing spring tides that carry eggs offshore or guarding broods of demersal eggs further defends propagules from reef-based predators. Subsequent settlement of larvae back to the reef, which occurs mostly at night, is also an apparent antipredatory adaptation.

Reproduction Social Systems and Sex Reversal

The best-studied examples of highly structured social systems in reef fishes are the harems of wrasses and parrotfishes. Typically, these fish are born as females that defend individual territories or occupy a shared home range. A larger male defends a group of females from other males, thereby sequestering matings. When the male dies, the dominant (typically largest) female changes sex (protogyny) and becomes the new harem master. At high population sizes, some fish may be born as males, develop huge testes, resemble females, infiltrate harems, and sneak spawnings with the resident females. Spatially isolated at their home anemones, anemonefishes have monogamous social systems in which the largest individual is female, the second largest is male, and the remaining fish are immature. Upon the death of the female, the male changes sex (protandry) and the behaviorally dominant juvenile fish matures into a male. Simultaneous hermaphroditism occurs among a few sparids and serranine sea basses. These fish have elaborate courtship behaviors during which individuals switch male and female roles between successive pair spawnings. Regardless of the broad variety of mating systems found in reef fishes, each individual behaves in a way that tends to maximize lifetime reproductive success. Life Cycle

The typical bony reef fish has a bipartite life cycle: a pelagic egg and larval stage followed by a demersal

657

(seafloor-oriented) juvenile and adult stage. Most bony reef fishes broadcast spawn, releasing gametes directly into the water column where they are swept to the open ocean. Smaller species spawn at their home reefs and some larger species, such as some grouper (Serranidae) and snapper (Lutjanidae), migrate to traditional sites and form massive spawning aggregations. Gametes are released during a paired or group ‘spawning rush’ followed by rapid return to the seafloor. Exceptions to broadcast spawning include demersal spawners that brood eggs until they hatch, either externally (e.g., egg masses defended by damselfishes) or internally (e.g., mouthbrooding cardinalfishes, Apogonidae), and a few ovoviviparous or viviparous species that give birth to welldeveloped juveniles (including reef sharks and rays). Annual fecundity of broadcast spawners ranges from about 10 000 to over a million eggs per female. Spawning is weakly seasonal compared to temperate species, typically peaking during summer months but not strongly related to any particular environmental variable. Lunar and semilunar spawning cycles are common. These are presumably adaptations that transport larvae offshore away from reef-based predation, maximize the number of settlement-stage larvae returning during favorable conditions that vary on lunar cycles, and/or benefit spawning adults in some way. Little is known about the behavior and ecology of reef-fish larvae. Duration of the pelagic larval stage ranges from about 9 to well over 100 days, averaging about a month. Larval prey include a variety of small zooplankters. Comparisons of fecundity at spawning to subsequent larval settlement back to the reef suggest that larval mortality is both extremely high and extremely variable, apparently due mostly to predation. Patterns of endemism, settlement to isolated islands, and limited data tracking larvae directly suggest that there is considerable larval retention at the scale of large islands, yet substantial larval dispersal nonetheless. Later-stage larvae are active swimmers and may control their dispersal by selecting currents among depths. The overall reproductive strategy is apparently to disperse the larvae offshore from reef-based predators, but then to retain offspring close enough to shore for subsequent settlement in suitable habitat. Settlement, the transition from pelagic larva to life on the reef (or nearby nursery habitat), occurs at a total length of c. 8 to c. 200 mm. Larger larvae are either morphologically distinct (e.g., the acronurus of surgeonfishes) or essentially pelagic juveniles (e.g., squirrelfishes and porcupinefishes). Choice of settlement habitat is apparent in some species, and both seagrass beds and mangroves can serve as nursery


658

CORAL REEF FISHES

habitats. Some wrasse larvae bury in the sand for several days before emerging as new juveniles. There is typically weak metamorphosis during settlement involving the growth of scales and onset of pigmentation. Estimates of settlement are generally called ‘recruitment’ and are based on counts of the smallest juveniles that can be found by divers some time after settlement. Once settled, most reef fish are thought to live less than a decade, although some small damselfish live at least 15 years.

Ecology Coral reef fishes are superb model systems for studying population dynamics and community structure of demersal marine fishes because they are eminently observable and experimentally manipulable in situ. These characteristics make studies of reef fishes conceptually relevant to demersal fisheries and ecology in general. Population Dynamics

Because reefs are patchy at all spatial scales and reef fish are largely sedentary, coral reef fishes form metapopulations: groups of local populations linked by larval dispersal. Many local populations are demographically open, such that reproductive output drifts away and is thus unrelated to subsequent larval settlement originating from elsewhere. Ultimately, the degree of openness depends on the spatial scale examined. For example, anemonefish populations are completely open at the scale of each anemone, may be partially closed at the scale of an oceanic island, and mostly closed at the scale of an archipelago. It is clear that variability in population size is driven by variation in recruitment due to larval mortality (and perhaps spawning success). Input to local populations via recruitment varies considerably at virtually every spatial and temporal scale examined. Increasing evidence indicates that two mechanisms predominate in regulating reef fish populations. First, given that density-dependent growth is common and that there is a general exponential relationship between body size and egg production in fish, density-dependent fecundity is likely. Second, early postsettlement mortality is often density-dependent, and has been demonstrated experimentally to be caused by predation in a variety of species. Community Structure

Due to high local species diversity, reef fish communities are complex. There are about five major feeding guilds, each containing dozens of species

locally (with approximate percentage of total fish biomass): zooplanktivores (up to 70%), herbivores (up to 25%), and piscivores (up to 55%), with the remainder being benthic invertebrate eaters or detritivores. The benthivores can be further subdivided based on prey taxa (e.g., corallivores) or other categories (e.g., consumers of hard-shelled invertebrates). Grunts that migrate from reefs at night and feed in surrounding seagrass beds subsequently return nutrients to the reef as feces. There is also considerable consumption of fish feces (coprophagy) by other fish on the reef. Fishes thus contribute substantially to nutrient trapping (via planktivory and nocturnal migration) and recycling (via coprophagy and detritivory) on coral reefs. Within each feeding guild, there is typically resource partitioning: each species consumes a particular subset of the available prey or forages in a distinct microhabitat. Communities are also structured temporally, with a diurnal assemblage being replaced by a nocturnal assemblage (the resting assemblage sheltering in the reef framework). The diurnal assemblage is dominated by perciform and tetraodontiform fishes, whereas the nocturnal assemblage is dominated by beryciform fishes (evolutionary relicts apparently relegated to the night by more recently evolved fishes). Maintenance of Species Diversity

Four major hypotheses have been proposed to explain how many species of ecologically similar coral reef fishes can coexist locally. There are data that both corroborate and falsify each hypothesis in various systems, suggesting that no universal generalization is possible. The first two hypotheses are based on the assumption that local populations are not only saturated with settlement-stage larvae, but also regularly reach densities where resources become limiting. First, the ‘competition hypothesis’, borrowed from terrestrial vertebrate ecology, suggests that coexistence is maintained despite ongoing interspecific competition by fine-scale resource partitioning (or niche diversification) among species. Second, the ‘lottery hypothesis’, derived to explain coexistence among similar territorial damselfishes that did not appear to partition resources, is based on the assumptions that, in the long run, competing species are approximately equal in larval supply, settlement rates, habitat and other resource requirements, and competitive ability. Thus, settling larvae are likened to lottery tickets, and it becomes unpredictable which species will replace which following the random appearance of open space due to the death of a territory holder or the creation of new habitat. The relatively restrictive assumptions of this


CORAL REEF FISHES

hypothesis can be relaxed if one considers the ‘storage effect’, which is based on the multiyear life span of reef fishes and the fact that settlement varies through time. Even though a species is at times an inferior competitor, as long as adults can persist until the next substantial settlement event, that species can persist in the community indefinitely. The third hypothesis, ‘recruitment limitation’, assumes that larval supply is so low that populations seldom reach levels where competition for limiting resources occurs, so that postsettlement mortality is density-independent and coexistence among species is guaranteed (assuming a storage effect). Finally, the ‘predation hypothesis’ predicts that early postsettlement predation, rather than limited larval supply, keeps populations from reaching levels where competition occurs, thereby ensuring coexistence.

See also Coral Reef and Other Tropical Fisheries. Coral Reefs. Diversity of Marine Species. Fish Feeding

659

and Foraging. Fish Predation and Mortality. Fish Reproduction. Fish Vision.

Further Reading Bo¨hlke JE and Chaplin CCG (1993) Fishes of the Bahamas and Adjacent Tropical Waters. Austin, TX: University of Texas Press. Caley MJ (ed.) (1998) Recruitment and population dynamics of coral-reef fishes: An international workshop. Australian Journal of Ecology 23(3). Lieske E and Myers R (1996) Coral Reef Fishes: IndoPacific and Caribbean. Princeton, NJ: Princeton University Press. Randall J (1998) Shore Fishes of Hawai’i. Honolulu, HI: University of Hawai’i Press. Randall JE, Allen GR, and Steene RC (1997) Fishes of the Great Barrier Reef and Coral Sea, 2nd edn. Honolulu, HI: University of Hawai’i Press. Sale PF (ed.) (1991) The Ecology of Fishes on Coral Reefs. San Diego, CA: Academic Press. Sale PF (ed.) (2002) Coral Reef Fishes: Dynamics and Diversity in a Complex Ecosystem. San Diego, CA: Academic Press.


CORAL REEFS J. W. McManus, University of Miami, Miami, FL, USA

General


Types of Coral Reefs


The term ‘coral reef’ commonly refers to a marine ecosystem in which a prominent ecological functional role is played by scleractinian corals. A ‘structural coral reef’ differs from a ‘nonstructural coral community’ in being associated with a geomorphologically significant calcium carbonate (limestone) structure of meters to hundreds of meters height above the surrounding substrate, deposited by components of a coral reef ecosystem. The term ‘coral reef’ is often applied to both structural and nonstructural coral ecosystems or their fossil remains, although many scientists, especially geomorphologists, reserve the term for structural coral reefs and their underlying limestone. Both types of ecosystem occur within a wide range of tropical and subtropical marine environments, although structural development tends to be greater in waters of lower silt or mud concentration and oceanic salinity. Many reefs survive well amid open ocean waters with low nutrients, aided by efficient ‘combing’ of waters for plankton, high levels of nitrogen fixation and fast and thorough nutrient cycling. However, extensive coral reefs also occur in coastal waters of much higher nutrient concentrations. Scleractinian (stony) corals grow as colonies or solitary polyps on a wide variety of substrates, including fallen trees, metal wreckage, rubber tires and rocks. Rates of settlement are often enhanced by the presence of calcareous encrusting algae. Soft sand, silt and mud tend to inhibit the settlement of stony corals, and so few coral ecosystems occur in modern or ancient deltaic deposits. However, coral can grow very near the mouths of small rivers and steams, and fresh or brackish groundwater often percolates through reef structures or emerges periodically through tunnels and caves. Vertical caves are often called ‘blue holes’. Nonstructural coral communities are common on rocky outcrops in shallow seas in many tropical and subtropical regions. They can range from a few clumps of coral to very substantial communities covering many square kilometers of wave-cut shelves near deeper areas. Structural coral reefs come in many shapes and sizes, from less than a kilometer to many tens of kilometers in linear dimension. It is helpful to differentiate individual coral reefs from systems of coral reefs, such as the misnamed ‘Great Barrier Reef’ of

Introduction Coral reefs are highly diverse ecosystems that provide food, income, and coastal protection for hundreds of millions of coastal dwellers. They are found in a diverse range of geomorphologies, from small coral communities of little or no relief, to calcareous structures hundreds of kilometers across. The most diverse coral reefs occur in the waters around southeast Asia. This is primarily because of extinctions that have occurred in other regions due to the gradual restriction of global oceanic circulation associated with continental drift. However, rates of speciation of coral reef organisms in this region may also have been high within the last 50 million years. Human activities have caused the degradation of coral reefs to varying degrees in all areas of the world. A major focus of present research is on the resilience of coral reefs to disturbances such as storms, diseases of reef species, bleaching (the expulsion of endosymbiotic photosynthetic zooxanthellae) and harvesting to local extinction. Along many coastlines, a combination of increased eutrophication due to coastal runoff and the extraction of herbivorous fish and invertebrates appears to favor the replacement of corals with macroalgae following disturbances. Another important aspect of resilience is the degree to which a depleted population of a given species on one reef can be replenished from other reefs. Most coral reef organisms undergo periods of free-swimming (pelagic) life ranging from a few hours to a few months. Genetic studies show evidence of broad dispersal of the progeny of some species among reefs, but most of the replenishment of a given population from year to year is believed to be from the same or nearby coral reefs. Coral reefs are complex biophysical systems that are generally linked to similarly complex socioeconomic systems. Their proper management calls for system-level approaches such as integrated coastal management.

660


CORAL REEFS

Shore

661

Crest

Reef flat Lagoon

Waves Slope Dropoff

Mangroves

Talus slope Corals Seagrass

Microatolls

Corals and other benthos

Figure 1 Profiles of a hypothetical fringing reef showing geomorphological and ecological zonation relative to wave action.

Australia, which actually consists of thousands of densely packed coral reefs. Common types of structural coral reefs include fringing reefs, barrier reefs, knoll reefs, pinnacle reefs, platform reefs, ribbon reefs, crescent reefs, and atolls. The term ‘patch reef’ may refer either to a patch of coral and limestone a few meters across within a structural coral reef (typically in a lagoon or on a reef flat), or to a platform or knoll reef. Thus, the term is best avoided. There is a wide range of structures intermediate

between crescent reefs, platform reefs, and atolls in areas such as the Great Barrier Reef System. A fringing reef is, by definition, always found adjacent to a land mass. Most fringing reefs include a wave-breaking reef crest, one or more meters above the rest of the reef, forming a thin strip offshore (Figure 1). Between the crest and the land, there is usually a relatively level area, broken by channels, called a ‘reef flat’ (Figure 2). Fringing reefs differ from barrier reefs, in that the latter are separated

Shore Lagoon Calm, shallow water between shore and reef Coral mounds

Back reef Reef flat Reef crest Reef edge Upper reef slope: a zone thronging with many forms of marine life, especially fish Lower reef slope: here, platelike hard corals and colourful soft corals and sea fans predominate

Seabed

Figure 2 Geomorphology of a typical fringing reef. (Adapted from Holliday, 1989.)


662

CORAL REEFS

from land by a ‘navigable’ water body (lagoon). It is useful to differentiate a lagoon from a reef flat in terms of depth; a lagoon is at least 2 m deep at mean tide. Fringing reefs often include lagoons, but the separation of a reef crest and slope from land is more complete in a barrier reef (Figure 3). Structural reefs such as knoll, pinnacle, platform, ribbon, or crescent reefs, are arbitrarily labeled based on their shape. Atolls are donut-shaped structures which, although often supporting islands along the outer rim, do not have an island in the central portion (Figure 4). The large Apo Reef, east of the Philippine island of Mindoro, is a double atoll, with two lagoons within adjacent triangular rims, the whole reef being roughly diamond-shaped. Atolls can be quite large, such as the North Male Atoll, which houses the capital of the Maldives (Figure 5). Although one commonly thinks of structural coral reefs as reaching to the sea surface, most of the coral reefs of the world do not. For example, there is a system of atolls and other reefs to approximately 50 km off the north-west of Palawan Island (Philippines) that closely resembles portions of the Australian Great Barrier Reef. However, very few reefs of the Palawan ‘barrier system’ come within 10 m of the sea surface. Some estimates of coral reef area are based on reefs at or near the sea surface, partly because the larger examples of such reefs tend to be

well-charted, whereas most other coral reefs are poorly known in terms of location and characteristics. Importance

Coral reefs support the highest known biodiversity of marine life, and constitute the largest biologically generated structures on Earth. Coral reefs are of substantial social, cultural, and economic importance. Coral reef systems in Florida, Hawaii, the Philippines, and Australia each account for more than $1 billion in tourist-related income each year. Coral reefs provide food and livelihoods for several tens of millions of fishers and their families, most of whom live in developing countries on low incomes, and have limited occupational mobility. Coral reefs protect coastal developments and farm lands from erosion. In many countries, particularly among Pacific islands, coral reefs are culturally very important, as they are involved in social structuring and interaction, and in religion. Despite poor taxonomic understanding and increasingly strict controls on bioprospecting, coral reef species are yielding substantial numbers of important drugs and other products. Zonation

The ecological zonation of most coral reefs depends on physical factors, including depth, exposure to

Shore Lagoon Newly forming fringing reef Channel Leeside face Reef flat Reef crest

Upper reef slope: spur-and-groove formations may occur on more exposed reefs Lower reef slope: gradient can vary here from a gentle slope to steep cliff

Seabed

Figure 3 Geomorphology of a typical barrier reef. (Adapted from Holliday, 1989.)


CORAL REEFS

663

Outer lagoon: shallow water bounded by a fringing reef Sand banks and islets form the 'dry land' of atoll Central lagoon: may be enclosed or open to the sea in places Patch reefs: separate, small coral growths in shallow water Reef face: gradient and coral life vary on location Figure 4 Geomorphology of a typical atoll. (Adapted from Holliday, 1989.)

waves and currents, and oxygen limitation. A fringing reef is basically a large slab of limestone jutting out from land. The landward margin may support extensive mangrove forests. A sandy channel may

NORTH MALE ATOLL

MALE Figure 5 North Male Atoll of the Maldives. Large atolls often support substantial human population that will be threatened as sealevels rise. (Adapted from Holliday, 1989.)

separate these from extensive seagrass beds on a reef flat. Other channels, depressions, and basins may be predominated by sandy bottoms, studded with clumps of coral. Clumps in deeper waters, rising 2 m or more from the bottom and consisting of several species of coral are known as ‘bommies.’ Some lagoons hold large numbers of ‘pillars,’ tall shafts of limestone, similarly supporting corals. Throughout the reef flat, there are, typically, low clumps of coral. Macroalgae and seagrass are often kept away from these clumps by the feeding action of the herbivorous fish and sea urchins, particularly active at night and resting during the day in the clumps of coral. Low massive coral colonies may form ‘microatolls,’ in which central portions of the colony are dead, while the raised outer edge and sides continue to flourish. On some reef flats, branching corals form patches which extend over hectares. Other reef flats and lagoons may be packed with high densities of diverse coral colonies, and have little or no seagrass. The seaward edge of the reef flat often leads into low branching or massive corals, including microatolls, which become increasingly dense to seaward. A thin band of macroalgae, such as Turbinaria, may be present, just before the clumps of coral and hard substrate rise to form a reef crest or ‘pavement.’ In some areas, the crest may be made of living coral. In areas protected from heavy wave action, the crest may consist of fingerlike Montipora or Porites


664

CORAL REEFS

forming a band several meters wide, broken periodically by whorl-shaped colonies of leafy corals, all apparently formed to efficiently comb breaking waves for zooplankton. On higher-energy coasts, the corals may be primarily dense growths of waveadapted Pocillopora. Other crests may be covered with various species and growth forms of the brown alga Sargassum. The edible and commercially important alga Caulerpa may also be abundant. Other reef crests may be densely packed with small clumps of articulated calcareous algae such as Halimeda and Amphiroa. Still others consist of a pavement of calcareous material or sheared-off ancient corals, and are relatively devoid of all macrobenthos except tiny clumps of algal turf or encrusting algae, sometimes forming a white or pink algal crest. The height of a reef crest above the reef flat is often determined by the height of local tides. Most reef crests are broken by channels of varying width and depth – exit routes for water piled up behind the reef crest by the breaking waves. These may be studded with corals or smoothed by scouring sand and rock carried along with the exiting water. They are particularly important as breeding grounds for a variety of reef fish. Beyond the reef crest, on the upper reaches of the ‘reef slope,’ one often finds small thick clumps of branching Pocillopora coral or various species of Acropora colonies in similar, wave-resistant growth forms. Encrusting and low lump-like (submassive) colonies may be common. More of the brown algae Turbinaria and Sargassum may be present. Along a rounded or gentle upper reef slope in the Indo-Pacific, there may be table-shaped Acropora colonies of gradually larger size as one proceeds to deeper waters. On many reefs, the reef slope is the most active area of coral growth, because of the oxygen, nutrients, and plankton brought in on the waves and currents. Coral cover may exceed 100%, as colonies overgrow colonies, all competing for light. Soft corals dominate some reef slopes, Sargassum others, and on some, the profusion of corals, sponges, algae, and other benthic organisms may prevent the identification of a dominant group. The mean and median stony coral cover (the percentage of the substrate covered with coral) on a reef slope globally are both approximately 40%, with a broad variance. On most reefs, small channels on the upper slope consolidate into increasingly wider and deeper channels along lower reaches of the slope. Some may be steepened or converted into tunnels by coral growth. The channels may occur fairly regularly at distances of tens of meters, resulting in a ‘ridge and rift’ or ‘spur (or buttress) and groove’ structure resembling the toes of giant feet. The bottoms are

generally filled with a mix of sand and debris from the reefs, which makes its way downwards, particularly during storms. Sand may drop continuously from escarpments, in flows resembling waterfalls, and spectacular columns of limestone cut away from the reef proper may border deeper rifts. Many reef slopes lead to a steep ‘wall’ or ‘dropoff,’ often beginning about 10–20 m depth. On shelf areas, the drop-off may end at 20–30 m depth, followed by a ‘talus’ slope of deposited reef materials, often 301–601, leading into the more gradual shelf slope. Typically, this shelf will be interrupted by outcrops of limestone and bommies for considerable distances from the reef itself. On reefs jutting into deep waters, the drop-off may extend downwards for hundreds of thousands of meters. Corals and a myriad of other organisms generally cover the slopes, leaving little or no bare substrate. On some reefs, such as some small fringing reefs along the Sinai, the upper portion consists almost entirely of live massive, platy or occasionally thick branching corals extending from land or from a small reef flat. The corals are joined together tangentially, leaving large spaces of water between. There may be no identifiable reef crest, and the mass simply juts out over a steep dropoff to hundreds of meters depth. The zonation of atolls and other surface reefs is generally similar. However, as one proceeds across a lagoon basin from the windward to the leeward side of an atoll, one encounters a ‘backreef’ area (note that the term is also applied by some to the coraldominated area behind a reef crest). The leeward side of the atoll is subject to less energetic waves and currents. It tends to have less well-defined zonation, less of a distinct crest, and often broader, more gradual slopes. Storms are very important in determining features of the reef. It is common to find large chunks of limestone on upper slopes, crests and reef flats, representing masses of coral and substrate pulled up from the lower slope and deposited during a storm. Pieces of this jetsam may weight several tonnes each. Smaller pieces of coral and substrate, and masses of sand may pile up during storms, forming islands. Processes of calcification in intertidal areas may glue together pieces of coral and reef substrate thrown up by storms, forming ‘beach rock,’ which helps to prevent erosion at the edge of low islands. Other islands may be formed on portions of a reef uplifted by tectonic processes. Islands developed by storm action or uplift often have very little relief, but may support human communities. Entire nations, as with the Maldives or some Pacific island nations, may consist of such low islands on atoll reefs.


CORAL REEFS

665

Geology Calcification

Carbon dioxide and water combine to form carbonic acid, which can then dissociate into hydrogen ions 2 (Hþ) and (HCO 3 )bicarbonate, or carbonate (CO3 ) ions as follows: CO2 þ H2 O3Hþ þ HO3 32Hþ þ CO3 2 Colder water can hold more CO2 in solution than warmer water. Calcium carbonate reacts with CO2 and water as follows: CO2 þ H2 O þ CaCO3 3Ca2þ þ 2HCO3 The dissolution of CaCO3 depends directly on the amount of dissolved CO2 present in the water. Thus, in colder waters the higher levels of CO2 inhibit calcification. Structural coral reefs are primarily limited to a circumglobal band stretching from the Line Islands in Hawaii to Perth, Australia, with nonstructural coral communities being increasingly dominant along the fringes. The distribution band varies, such that cold waters (e.g., Peru) narrow the range of coral reefs closer to the equator, and warm currents facilitate higher latitude development (e.g., Bermuda). Calcification in stony corals is greatly enhanced by the presence of zooxanthellae in the tissues. Zooxanthellae are nonmotile stages of a dinoflagellate algae. They are contained within specialized coral cells, growing on excess nutrients and trace metals from the carnivorous corals, producing sugars for utilization by the host, and using up excess CO2 to enhance calcification by the coral. Zooxanthellae are found in other reef organisms, including giant clams, and the tiny foraminifera, amoeba-like organisms (Order Sarcodina) with calcareous skeletons that create substantial amounts of the sand on coral reefs and nearby white-sand beaches. Much of the biology and ecology of zooxanthellae is poorly known.

Figure 6 Geological time-scale (in thousand years) highlighting biogenic reef development. Arrows denote changes in scale. (Adapted from Hallock, 1989.)

Paleoecology

Structural coral reefs are forms of ‘biogenic reefs,’ distinct geomorphological structures constructed by living organisms. Biogenic reefs have existed in various forms since approximately 3.5 billion years ago, at which time cyanobacteria began building stromatolites (Figure 6). Bioherms, biogenic reefs constructed from limestone produced by shelled animals, became prominent by 570 million years ago. During the mid-Triassic, the Jurassic and early Cretaceous periods (roughly 200–100 million years ago), scleractinian corals had become significant

components of coral–algal–sponge bioherms. Some of these corals may have included zooxanthellae. However, by the mid-Cretaceous, rudist bivalves dominated shallow-water reefs, and scleractinian corals were mainly restricted to deeper shelf-slope environments. This may have been due to seawater chemistry or competition with rudists or both. The massive extinction at the end of the Cretaceous, which ended the reign of dinosaurs on land, also led to the extinction of rudists. Many scleractinian corals survived in their deeper habitats, and evolved into


666

CORAL REEFS

most of the modern genera by the Eocene. However, CO2 concentrations in the atmosphere were high at the time, apparently inhibiting reef development by aragonite-producing corals. Instead, major limestone deposits of the time were formed by calcite-producing organisms such as larger foraminifera (including the limestones from which the Egyptian pyramids were later to be built) and red coralline algae. Calcite is less stable over time than aragonite, but can form in waters of higher CO2 concentration. Major reef-building by stony corals and their associates did not occur until the middle to late Oligocene. By then, CO2 concentrations were falling and sea water was warming in tropical areas, whereas at high-latitudes it became cooler. By the late Oligocene, Caribbean coral reefs achieved their greatest development, and by the early Miocene, coral reefs globally had extended to beyond 101 north and south. During the late Eocene and early Oligocene, tropical oceans were openly connected, in a system of equatorial waterways known as the Tethys Sea. During this period, most scleractinian corals were cosmopolitan. The upward movements of Africa and India restricted circulation, particularly with the closure of the Qatar arch in the Middle East around the time of the Oligocene–Miocene boundary (Figure 7). The Central American passageway became restricted, but did not close until the middle Pliocene. However, nearly half of the Caribbean coral genera became extinct at the Oligocene–Miocene boundary, and many more disappeared during the Miocene. Larger foraminifera suffered similar extinctions. Far less dramatic extinctions occurred in the Indo-Pacific.

EARLY PLIOCENE

MIDDLE OLIGOCENE

MIDDLE PALEOCENE Figure 7 Major tectonic movements and inferred patterns of major currents since the Paleocene. Gradually, the circumglobal waterway of the Paleocene, the Tethys Sea, was closed off, resulting in regional extinction of coral reef organisms. This process explains many of the differences in biodiversity now seen among the world’s coral reefs. (Adapted from Hallock, 1997.)

(A)

Reef Geomorphology

Charles Darwin proposed that atolls were formed by a process involving the sinking of islands (Figure 8). He noted that most high islands in the Pacific were

(B)

(C)

Figure 8 Stages in the development of an atoll as envisioned by Darwin. Fringing reefs surrounding a volcano (A) become barrier reefs as the volcano sinks (B) and finally form a donut-shaped atoll (C) Sinking islands and/or rising sea levels do help to explain the structures of many reefs, but other factors have been more important in the case of other reefs. (Adapted from Slafford-Deitsch, 1991.)


CORAL REEFS

the tops of volcanic cones. A rocky, volcanic island would tend to form fringing reefs around its shores. Should the island begin to sink, the highly oxygenated outer edges would be able to maintain themselves at the sea surface, while the less actively growing reef flats would tend to sink, forming lagoons defining barrier reefs. Eventually, a sinking cone would disappear altogether, leaving a lagoon in the midst of an atoll. The feasibility of this explanation was confirmed by the mid-twentieth century when drilling on Enewetak and Bikini reefs yielded signs that the atolls were perched over islands that had gradually subsided over thousands of years. Modern researchers have identified a variety of explanations for the morphology of various reefs. Variations in relative sea level have had profound influences on most structural coral reefs. Few coral reefs are believed to have exhibited continuous growth prior to 8000 years ago, and most are considerably younger. Coral reefs located on extensive continental shelf areas, such as the Florida reef tract and the vast shelf areas off the Yucatan and peninsular south-east Asia, grow on substrates that were generally far inland during the previous ice age. The three-dimensional shapes of many coral reefs have been highly influenced by underlying topography, often an ancient reef. Wind, storms, waves, and currents have helped to shape many reefs, resulting in tear-drop shapes with broad, raised, steep edges facing predominant winds and currents. The shapes of some reefs are believed to have been particularly dependent on the erosion of much larger blocks of reef limestone during low stands of sea level. It has been demonstrated in the laboratory that small blocks of limestone subjected to rain-like erosion can form many features now identifiable on coral reefs, including lagoons, rifts and ridges, and channels. Evidence has also been found that the ridge and rift structures and some other features of coral reefs can result from collective ecological growth processes as the biota accommodates effects of waves, channeled water, and scouring by sand and other reef matter as calcifying benthic organisms compete for exposure to sunlight. It has been suggested that the double lagoon structure of the Atoll reef off Mindoro, Philippines, resulted from coral reef growth around the subsurface rim of a volcanic crater. Drilling on some reefs has demonstrated continuous growth for 30 m or more. On other reefs, little or no substantial calcification has occurred, the extreme case being nonstructural coral communities. The Hawaiian and Line Islands are part of a vast chain of islands formed as a tectonic plate has drifted north and west across the Pacific. Volcanic lava has tended to erupt in a fixed ‘hot-spot’, forming

667

volcanic islands one after the other. As the islands have drifted north, they have subsided gradually into deeper waters until coral growth has ceased. This process has had a strong influence over the structure of coral reefs along these islands and seamounts. However, at smaller scales, coral reefs along the coasts of the Hawaiian islands also show the effects of heavy storm action, exhibiting mound-like features at the sea surface often dominated by calcareous algae, and sometimes broad platforms which produce the famous surfing waves of the islands. Role in Oil Exploration

Coral reefs are generally highly porous, filled with holes, tunnels, and caves of all sizes. The most famous ‘blue-holes’ of the Bahamas are the vertical seaward entrances of huge cave and tunnel systems formed by erosion at low stands of sea level, through which sea water flows beneath the islands as water through a sponge. Much of the porosity of coral reefs results from incomplete filling of spaces between coral heads during the active growth of the reef. Much of the world’s crude oil comes from ancient coral reefs that have been subsequently overlain with terrigenous (land-based) sediments. Dense, horizontally layered sediments have often trapped oil within the mounds of porous coral reef limestone. In some areas, subsequent tectonic activity has produced faults, and much of the oil has been lost. In other areas, however, the ancient reefs have yielded vast amounts of oil. In some parts of the world, modern coral reefs have grown in areas where ancient reefs once occurred and were subsequently covered by dense sediments that trapped oil. Thus, it is common to find oil drilling platforms on modern reefs. This has frequently raised environmental concerns, centering on the damage done to modern corals by the construction and associated activities, the occasional spillages of drilling muds of various compositions and the leakage of oil.

Biology Biogeography

The highest species diversity on coral reefs occurs in the seas of south-east Asia, often referred to as the Central Indo-West Pacific region. In the Caribbean or Hawaii, there are 70–80 species of reef-building corals. In south-east Asia, there are more than 70 genera and 400 or more species of reef-building corals (Figure 9). The diversities of most other groups of reef-associated species, including fish, tend to be higher in south-east Asia than in other regions.


668

CORAL REEFS

Figure 9 Patterns of generic diversity is scleractinian corals. Contours were estimated from generic patterns and existing data on species distributions. (From Veron 1993; reprinted with permission of the author.)

Diversity begins to drop off gradually to the south below Indonesia, and north above the Philippines. The decline eastward is very dramatic, such that only a few species of corals live in Tahiti and the Galapagos. Westward, the decline is only slight; at least 50 genera of corals live along the coasts of East Africa and the Red Sea. There are only a few species of reef-building corals off the coast of West Africa. Species in south-east Asia tend to have very broad ranges, and endemism is higher in peripheral areas such as the central Pacific. Much of the global pattern of diversity, particularly at the level of genera and families, can be explained in terms of selective regional extinctions following the gradual breakup of the circumglobal Tethys Sea during the Cenozoic. This is well supported in the fossil records of corals, seagrasses and mangrove trees. Although fish fossils are uncommon, there is a highly diverse assemblage of fossil coral reef fish in Italy, a remnant from the Tethys Sea. However, there is some reason to believe that rates of speciation have been higher in south-east Asia than elsewhere. The evidence is best known for mollusks, and include unusually high ratios of species to genera and particularly highly evolved armament (such as the spines on many murex shells) in this region. There is a myriad of explanations for this potentially higher speciation rate. Two popular explanations, the Pacific island vicariance hypothesis and the basin isolation hypothesis, both involve changes in sea level. In the former hypothesis, high levels of the sea tend to isolate groups of Pacific islands, facilitating speciation. At low sea levels, the species spread to the heterogeneous refugia habitat of south-east Asia and are gradually lost on the lessheterogeneous Pacific islands due to competition from other species. In the latter hypothesis, low stands of sea level, and, more importantly, periods of mountain building, have isolated particular marine basins within south-east Asia in the past, possibly permitting rapid speciation among whole biotas. Further understanding of these processes must await increased efforts in taxonomy and systematics.

The vast majority of coral reef organisms, particularly in the Indo-West Pacific region, have never been formally identified. The ranges of known organisms are poorly known. Furthermore, recent decades have seen a rapid decline in interest in and support for taxonomy and systematics. It is likely that human impacts will result in vast changes to the distributions and abundances of coral reef species long before existing biogeography and ecology have been well understood. Coral Reef Ecosystem Health

Resilience and phase shifts Most coral reefs are subjected to periodic disturbances, such as storms. The ability of an ecosystem to maintain constancy in terms of ecological functions and in the abundances and distributions of organisms is ecological resistance. The capacity of an ecosystem to revert to a previous state (or near to it) in terms of these characteristics is ecological resilience. Terms such as ecosystem health, integrity, and stability usually infer degrees of resistance and resilience. Because profound changes often occur to a coral reef following a perturbation such as a storm, there is increasing focus on the resilience of a coral reef. Many coral reefs have been known to undergo losses of coral cover from greater than 50% to less than 10%, and then to recover to the former level within 4–10 years. Naturally, in reefs in which colonies may have been decades or centuries old, the age structure of the corals present have often been disrupted. This, in turn, may affect resilience to future perturbations, as certain corals are believed to exhibit higher fecundities in older colonies. Some coral reefs do not seem to return to high levels of coral cover after a perturbation, especially if they have been subsequently overgrown by fleshy macroalgae. Human intervention, in the form of increased eutrophication and the removal of herbivorous fish and invertebrates, are suspected to favor the growth of macroalgae following perturbations.


CORAL REEFS

Reefs around Jamaica have shown little recovery in more than a decade following coral damage by a hurricane, and both forms of intervention are suspected of being causes of this loss of resilience. Widespread losses of resilience are a concern throughout most coral reef areas. Reef interconnectivity Most coral reef organisms undergo pelagic life stages before settling into a reef community. Most corals live as planulae for a few hours to a few days before setting, although longer periods have been recorded. The average benthic invertebrate spends roughly two weeks in waterborne stages, but some survive for months. The average coral reef fish appears to require nearly a month before settling, and many require two months. The pelagic stages are by no means passive, and although the sizes are very small, the organisms may be adapted to swim into currents and eddies that facilitate their retention or return to particular reefs or groups of reefs. Their success in doing so is believed to depend on factors such as reef geomorphology and the nature and predictability of local oceanography. Analyses indicate that some fish populations regularly exchange genetic material over thousands of kilometers. Although most coral-reef populations are believed to be replenished each year from local progeny, a severely depleted reef may be restocked to some degree from other reefs. This process is crucial to the problem of resilience, and is an active area of controversy and research. The results of this work may have profound implications for the design of marine protected areas, for international agreements on the coordination of management schemes, and for the regulation of harvesting on coral reefs. Furthermore, climate change is likely to alter local oceanographic processes, and has important implications for reef management as such disruptions occur.

Coral Reef Management Disturbances

Various types of perturbations have affected coral reefs with increasing frequency within recent decades. Some of these are clearly directly related to human activities, whereas others are suspected to be the indirect results of human interventions. The majority of coral reefs are located in developing countries. In many of these, crowded lowincome human populations increasingly overfish, often reducing herbivorous fish and invertebrates, thereby decreasing reef resilience. In a process

669

known as Malthusian overfishing, social norms break down and fishers turn to destructive fishing methods such as the use of poisons and explosives to capture fish. Organic pollution from coastal habitations and agricultural fertilizing activities are common along coastlines. Runoff from deforested hillsides, mining operations, and coastal construction often contains materials that favor macroalgal growth, and silt and mud that restrict light to zooxanthellae, abrade reef benthos or bury portions of coral reefs entirely. Increasingly common outbreaks of the coral-eating crown-of-thorns starfish, Acanthaster planci, may be related to reductions in predators, such as lethrinid fishes. During 1997–98, a worldwide epidemic of coral bleaching occurred, in which the zooxanthellae of many colonies were expelled and high rates of coral mortality resulted. The cause was unusually warm patches of sea water associated with a strong El Ninõ event, which some believe to be related to increasing levels of atmospheric CO2 and global warming. A more controversial suggestion is that the increasing CO2 levels will cause acidification of the oceans sufficient to result in the net erosion of some coral reefs, especially at high latitudes. Although there have recently been major epidemics of diseases killing corals and associated organisms in western Atlantic reefs, only certain coral diseases appear to be directly linked to stress from human activities. Assessments

Efforts are underway to gather empirical information on coral reefs via the quantification of benthic organisms and fish by divers. These efforts are being supplemented by the use of remotely sensed data from satellites, space shuttles, aircraft, ships, manned and unmanned underwater vehicles. The usefulness of these approaches ranges from identifying or mapping reefs, to quantifying bleaching and disease. A global database, ReefBase, has been developed by the International Center for Living Aquatic Resources Management (ICLARM) to gather together existing information about the world’s reefs and to make it widely available via CDROM and the Internet. Integrated Coastal Management

Management is a process of modifying human behavior. Biophysical scientists can provide advice and predictions concerning factors such as levels of fishing pressure, siltation, and pollution. However, the management decisions must account for social, cultural, political, and economic considerations. Furthermore, almost all management interventions will


670

CORAL REEFS

have both positive and negative effects on various aspects of the ecosystem and the societies that impact it and depend upon it. For example, diverting fishers into forestry may lead to increased deforestation, siltation, and further reef degradation. It is increasingly recognized that effective management is achieved only through approaches that integrate biophysical considerations with socioeconomic and related factors. Balanced stakeholder involvement is generally a prerequisite for compliance with management decisions. The field of integrated coastal management is rapidly evolving, as is the set of scientific paradigms on which it is based. To a large degree, the future of the world’s coral reefs is directly linked to this evolution.

See Also Air–Sea Transfer: N2O, NO, CH4, CO. Beaches, Physical Processes Affecting. Carbon Cycle. Carbon Cycle. Carbon Dioxide (CO2) Cycle. Cenozoic Climate – Oxygen Isotope Evidence. Cenozoic Oceans – Carbon Cycle Models. Coral Reef Fishes. Diversity of Marine Species. El Ninõ Southern Oscillation (ENSO). Eutrophication. Fish Feeding and Foraging. Geomorphology. History of Ocean Sciences. Lagoons. Macrobenthos. Mangroves. Manned Submersibles, Shallow Water. Nitrogen Cycle. Past Climate from Corals. Pelagic Biogeography. Remotely Operated Vehicles (ROVs). Rocky Shores. Sandy Beaches, Biology of. Satellite Oceanography, History and Introductory Concepts. Satellite Remote Sensing of Sea Surface Temperatures. Sea Level Change.

Sea Level Variations Over Geologic Time. Ships. Storm Surges.

Further Reading Birkeland C (ed.) (1997) Life and Death of Coral Reefs. New York: Chapman and Hall. Bryant D, Burke L, McManus J, and Spalding M (1998) Reefs at Risk: a Map-based Indicator of the Threats to the World’s Coral Reefs. Washington, DC: World Resources Institute. Davidson OG (1998) The Enchanted Braid: Coming to Terms with Nature on the Coral Reef. New York: John Wiley. Holliday L (1989) Coral Reefs: a Global View by a Diver and Aquarist. London: Salamander Press. McManus JW, Ablan MCA, Vergara SG, et al. (1997) ReefBase Aquanaut Survey Manual. Manila, Philippines: International Center for Living Aquatic Resources Management. McManus JW and Vergara SG (eds.) (1998) ReefBase: a Global Database of Coral Reefs and Their Resources. Version 3.0 (Manual and CD-ROM). Manila, Philippines: International Center for Living Aquatic Resource Management. Polunin NVC and Roberts C (eds.) (1996) Reef Fisheries. New York: Chapman and Hall. Sale PF (ed.) (1991) The Ecology of Fishes on Coral Reefs. New York: Academic Press. Stafford-Deitsch J (1993) Reef: a Safari Through the Coral World. San Francisco: Sierra Club Books. Veron JEN (1993) A Biogeographic Database of Hermatypic Corals. Species of the Central Indo-Pacific, Genera of the World. Monograph Series 10. Townsville, Australia: Australian Institute of Marine Science.


CORALS AND HUMAN DISTURBANCE N. J. Pilcher, Universiti Malaysia Sarawak, Sarawak, Malaysia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 543–549, & 2001, Elsevier Ltd.

Introduction Coral reefs are the centers of marine biodiversity on the planet. Reefs are constructed by a host of hermatypic (reef-building) coral species, but also are home to ahermatypic (non-calcium-carbonate depositing) corals, such as soft corals, black corals and gorgonians. The major structural components of reefs are the scleractinian corals. Much like their terrestrial rivals the tropical rainforests, reefs combine a host of microhabitats and a diverse array of life forms that is still being discovered and described. Coral reefs are mostly distributed throughout the tropical belt, and a large fraction are located in developing countries. To understand how human activities affect coral reefs, it is necessary to briefly review their basic life history. Coral reefs are mostly made up of numerous smaller coral colonies; these colonies are in turn made up of thousands of minute polyps, which secrete a calcium carbonate skeleton. The deposition rate for individual coral species varies, but generally ranges between 0.1 mm and 10.0 cm per year. The accumulation of these skeletons over a long period of time results in massive, three-dimensional geological structures. The actual living tissue, however, is only the thin layer of living coral polyps on the surface. Corals are particularly susceptible to contaminants in sea water because the layer of tissue covering the coral skeleton is thin (B100 mm) and rich in lipids, facilitating direct uptake of chemicals. Coral polyps feed by filtering plankton using nematocyst (stinging cell)-tipped tentacles, and also receive organic matter through their symbiotic relationship with minute dinoflagellates called zooxanthellae. Zooxanthellae live within the gastrodermal tissues, and chemical communication (exchange) occurs via the translocation of metabolites. These small algal cells use sunlight to photosynthesize carbonates and water into organic matter and oxygen, both of which are used by the polyp. Coral reefs support complex food and energy webs that are interlinked with nutrient inputs from outside sources (such as those brought with ocean currents

and runoff from nearby rivers) and from the reef itself (where natural predation and die-off recirculate organic matter). These complex webs mean that any effect on one group of individuals will ultimately impact another, and single disturbances can have multiple effects on reef inhabitants. For example, the complete eradication of the giant Triton Charonia trinis through overfishing can result in outbreaks of the crown-of-thorns starfish Acanthaster planci. This can lead to massive coral mortality as the starfish reproduce and feed on the coral polyps. The mortality in turn may reduce habitats and food sources for reef fishes, which again, in turn, could lead to declines of larger predatory fishes. Similarly, the introduction of an invasive species either by accident or through ignorance (e.g., dumping of a personal aquarium contents into local habitats) might disrupt feeding processes and kill resident fishes. The death of key organisms on the reef (which then shifts from an autotrophic to a heterotrophic, suspension/detritus-feeding community) changes the dominant ecological process from calcium carbonate deposition to erosion, and ultimate loss of coral reef. Reef ecosystems may respond to environmental change by altering their physical and ecological structure, and through changes in rates of accretion and biogeochemical cycling. However, the potential for adaptation in reef organisms may be overwhelmed by today’s anthropogenic stresses. The following sections provide a review of human disturbances and their general effects on coral reefs.

Collection of Corals Corals have been mined for construction purposes in numerous Pacific Ocean islands and in South-East Asia. Usually the large massive life forms such as Porites, Platygyra, Favia, and Favites are collected and broken into manageable sizes or crushed for cement and lime manufacture. Similarly, the shells of the giant clam Tridacna and the conch shell Strombus are collected. Coral blocks and shells are used for construction of houses, roads, and numerous other projects. Corals are also collected for use in the ornamental trade, either as curios and souvenirs or as jewelry. Entire, small colonies of branching species such as Acropora and Seriatophora are used in the souvenir trade and for decoration, while black corals Antipathes and blue coral Heliopora are used for jewelry. The aquarium industry is also responsible for coral collection either for direct sale as live


671

672

CORALS AND HUMAN DISTURBANCE

colonies or through the process of fish collecting. In many cases entire colonies full of fish are brought to the surface and are then smashed and discarded. The removal of coral colonies decreases the shelter and niche areas available to numerous other reef inhabitants. Juvenile stages of fish that seek shelter among the branching species of corals, and worms and ascidians that take up residence on massive life forms, are deprived of protection and may become prey to other reef organisms. Removal of adult colonies also results in a reduction of overall reproductive output, as the corals no longer serve as a source of replenishment larvae. Further, removal of entire colonies reduces the overall structural stability of the reef, and increases rates of erosion through wave and surge damage.

Destructive Fishing Destructive fishing pressures are taking their toll on coral reefs, particularly in developing countries in South-East Asia. The use of military explosives and dynamite was common shortly after the Second World War, but today this has shifted to the use of home-made explosives of fertilizer, fuel, and fuse caps inserted into empty beer bottles. Bombs weigh approximately 1 kg and have a destructive diameter of 4–5 m. Blast fishers hunt for schooling fish such as sweetlips and fusiliers, which aggregate in groups in the open or hide under large coral heads. Parrot fish and surgeon fish schools grazing on the reef crest are also actively sought. The bombs are usually set on five-second fuses and are dropped into the center of an area judged to have many fish. After the bomb has exploded, the fishers use dip nets, either from the boats or from underwater, to collect the stunned and dying fish. Many larger boats collect the fish using ‘hookah’ compressors and long air hoses to support divers working underwater. The pressure wave from the explosion kills or stuns fish, but also damages corals. Natural disturbances may also fragment stony reef corals, and there are few quantitative data on the impacts of skeletal fragmentation on the biology of these corals. Lightly bombed reefs are usually pockmarked with blast craters, while many reefs in developing countries comprise a continuous band of coral rubble instead of a reef crest and upper reef slope. The lower reef slope is a mix of rubble, sand, and overturned coral heads. Typically at the base of the reef slope is a mound of coral boulders that have been dislodged by a blast and then rolled down the slope in an underwater avalanche. The reef slopes are mostly dead coral, loose sand, rubble, or rock and occasionally

have overturned clams or coral heads with small patches of living tissue protruding from the rubble. The blasts also change the three-dimensional structure of reefs, and blasted areas no longer provide food or shelter to reef inhabitants. Further, once the reef structure has been weakened or destroyed by blast fishing, it is much more susceptible to wave action and the reef is unable to maintain its role in coastline protection. Larvae do not settle on rubble and thus replenishment and rehabilitation is minimal. Additionally, the destruction of adult colonies also results in a reduction of overall reproductive output, and reefs no longer serve as a source of replenishment larvae. Experimental findings, for instance, indicate that fragmentation reduces sexual reproductive output in the reef-building coral Pocillopora damicornis. The recovery of such areas has been measured in decades, and only then with complete protection and cessation of fishery pressure of any kind. Another type of destructive fishing is ‘Muro Ami’, in which a large semicircular net is placed around a reef. Fish are driven into the net by a long line of fishermen armed with weighted lines. The weights are repeatedly crashed onto the reef to scare fish in the direction of the net, reducing coral colonies to rubble. The resulting effects are similar to the effects of blast fishing, spread over a larger area. Cyanide fishing is also among the most destructive fishing methods, in which an aqueous solution of sodium cyanide is squirted at fish to stun them, after which they are collected and sold to the live-fish trade. Other chemicals are also used, including rotenone, plant extracts, fertilizers, and quinaldine. These chemicals all narcotize fish, rendering them inactive enough for collection. The fish are then held in clean water for a short period to allow them to recover, before being hauled aboard boats with live fish holds. In the process of stunning fish, the cyanide affects corals and small fish and invertebrates. The narcotizing solution for large fish is often lethal to smaller ones. Cyanide has been shown to limit coral growth and cause diseases and bleaching, and ultimately death in many coral species. Among other destructive aspects of fishing are lost fishing gear, and normal trawl and purse fishing operations, when these take place near and over reefs. Trawlers operate close to reefs to take advantage of the higher levels of fish aggregated around them, only to have the trawls caught on the reefs. Many of these have to be cut away and discarded, becoming further entangled on the reefs, breaking corals and smothering others. Similarly, fishing with fine-mesh nets that get entangled in coral structures also results in coral breakage and loss. In South-East Asia,



fishing companies have been reported to pull a chain across the bottom using two boats to clear off corals, making it accessible to trawlers. Spearfishing also damages corals as fishermen trample and break coral to get at fish that disappear into crevices, and crowbars are frequently used to break coral. The collection of reef invertebrates along the reef crest results in breakage of corals that have particular erosion control functions, reducing the reef’s potential to act as a coastal barrier.

Discharges Mankind also effects corals through the uncontrolled and often unregulated discharge of a number of industrial and domestic effluents. Many of these are ‘point-source’ discharges that affect local reef areas, rather than causing broad-scale reef mortality. Sources of chemical contamination include terrestrial runoff from rivers and streams, urban and agricultural areas, sewage outfalls near coral reefs, desalination plants, and chemical inputs from recreational uses and industries (boat manufacturing, boating, fueling, etc.). Landfills can also leach directly or indirectly into shallow water tables. Industrial inputs from coastal mining and smelting operations are sources of heavy metals. Untreated and partially treated sewage is discharged over reefs in areas where fringing reefs are located close to shore, such as the reefs that fringe the entire length of the Red Sea. Raw sewage can result in tumors on fish, and erosion of fins as a result of high concentrations of bacteria. The resulting smothering sludge produces anaerobic conditions under which all benthic organisms perish, including corals. In enclosed and semi-enclosed areas the sewage causes eutrophication of the coral habitats. For instance, in Kaneohe bay in Hawaii, which had luxuriant reefs, sewage was dumped straight into the bay and green algae grew in plague quantities, smothering and killing the reefs. Evidence indicates that branching species might be more susceptible to some chemical contaminants than are massive corals. Abattoir refuse is another localized source of excessive nutrients and other wastes that can lead to large grease mats smothering the seabed, local eutrophication, red tides, jellyfish outbreaks, an increase in biological oxygen demand (BOD), and algal blooms. Similarly, pumping/dumping of organic compounds such as sugar cane wastes also results in oxygen depletion. The oil industry is a major source of polluting discharges. Petroleum hydrocarbons and their derivatives and associated compounds have caused

673

widespread damage to coastal ecosystems, many of which include coral reefs. The effects of these discharges are often more noticeable onshore than offshore where reefs are generally located, but nonetheless have resulted in the loss of reef areas, particularly near major exploration and drilling areas, and along major shipping routes. Although buoyant eggs and developing larvae are sometimes affected, reef flats are more vulnerable to direct contamination by oil. Oiling can lead to the increased incidence of mortality of coral colonies. In the narrow Red Sea, where many millions of tonnes per annum pass through the region, there have been more than 20 oil spills along the Egyptian coast since 1982, which have smothered and poisoned corals and other organisms. Medium spills from ballast and bilgewater discharges, and leakages from terminals, cause localized damage and smothering of intertidal habitats. Oil leakage is a regular occurrence from the oil terminal and tankers in Port Sudan harbor. Seismic blasts during oil exploration are also a threat to coral reefs. Refineries discharge oil and petroleum-related compounds, resulting in an increase in diatoms and a decrease in marine fauna closer to the refineries. Throughout many parts of the world there is inadequate control and monitoring of procedures, equipment, and training of personnel at refineries and shipping operations. Drilling activities frequently take place near reef areas, such as the Saudi Arabian shoreline in the Arabian Gulf. Drilling muds smother reefs and contain compounds that disrupt growth and cause diseases in coral colonies. Field assessment of a reef several years after drilling indicated a 70–90% reduction in abundance of foliose, branching, and platelike corals within 85–115 m of a drilling site. Research indicates that exposure to ferrochrome lignosulfate (FCLS) can decrease growth rates in Montastrea annularis, and growth rates and extension of calices (skeleton supporting the polyps) decrease in response to exposure to 100 mg l 1 of drilling mud. Oil spills affect coral reefs through smothering, resulting in a lack of further colonization, such as occurred in the Gulf of Aqaba in 1970 when the coral Stylophora pistillata did not recolonize oilcontaminated areas after a large spill. Effects of oil on individual coral colonies range from tissue death to impaired reproduction to loss of symbiotic algae (bleaching). Larvae of many broadcast spawners pass through sensitive early stages of development at the sea surface, where they can be exposed to contaminants and surface slicks. Oiling affects not only coral growth and tissue maintenance but also reproduction. Other effects from oil pollution include


674


degeneration of tissues, impairment of growth and reproduction (there can be impaired gonadal development in both brooding and broadcasting species, decreased egg size and decreased fecundity), and decreased photosynthetic rates in zooxanthellae. In developing countries, virtually no ports have reception facilities to collect these wastes and the problem will continue mostly through a lack of enforcement of existing regulations. The potential exists for large oil spills and disasters from oil tank ruptures and collisions at sea, and there are no mechanisms to contain and clean such spills. The levels of oil and its derivatives (persistent carcinogens) were correlated with coral disease in the Red Sea, where there were significant levels of diseases, especially Black Band Disease. In addition to the impacts of oils themselves are the impacts of dispersants used to combat spills. These chemicals are also toxic and promote the breakup of heavier molecules, allowing toxic fractions of the oil to reach the benthos. They also promote erosion through limiting adhesion among sand particles. The full effect of oil on corals is not fully understood or studied, and much more work is needed to understand the full impact. Although natural degradation by bacteria occurs, it is slow and, by the time bacteria consume the heavy, sinking components, these have already smothered coral colonies. Industrial effluents, from a variety of sources, also impact coral reefs and their associated fauna and habitats. Heavy metal discharges lead to elevated levels of lead, mercury, and copper in bivalves and fish, and to elevated levels of cadmium, vanadium, and zinc in sediments. Larval stages of crustaceans and fish are particularly affected, and effluents often inhibit growth in phytoplankton, resulting in a lack of zooplankton, a major food source for corals. Industrial discharges can increase the susceptibility of fish to diseases, and many coral colonies end up with swollen tissues, excessive production of mucus, or areas without tissue. Reproduction and feeding in surviving polyps is affected, and such coral colonies rarely contribute to recolonization of reef areas. Organisms in low-nutrient tropical waters are particularly sensitive to pollutants that can be metabolically substituted for essential elements (such as manganese). Metals enter coral tissues or skeleton by several pathways. Exposed skeletal spines (in response to environmental stress), can take up metals directly from the surrounding sea water. In Thailand, massive species such as Porites tended to be smaller in areas exposed to copper, zinc, and tin, there was a reduced growth rate in branching corals, and calcium carbonate accretion was significantly reduced.

Symbiotic algae have been shown to accumulate higher concentrations of metals than do host tissues in corals. Such sequestering in the algae might diminish possible toxic effects to the host. In addition, the symbiotic algae of corals can influence the skeletal concentrations of metals through enhancement of calcification rates. There is evidence, however, that corals might be able to regulate the concentrations of metals in their own tissues. For example, elevated iron in Thai waters resulted in loss of symbiotic algae in corals from pristine areas, but this response was lower in corals that has been exposed to daily runoff from an enriched iron effluent, suggesting that the corals could develop a tolerance to the metal. Cooling brine is another industrial effluent that affects shoreline-fringing reefs, often originating from industrial installations or as the outflow from desalination plants. These effluents are typically up to 5–101C higher in temperature and up to 3–10 ppt higher in salinity. Discharges into the marine environment from desalination plants in Jeddah include chlorine and antiscalant chemicals and 1.73 billion m3 d 1 of brine at a salinity of 51 ppt and 411C. The higher temperatures decrease the water’s ability to dissolve oxygen, slowing reef processes. Increases in temperature are particularly threatening to coral reefs distributed throughout the tropics, where reefbuilding species generally survive just below their natural thermal thresholds. Higher-temperature effluents usually result in localized bleaching of coral colonies. The higher-salinity discharges increase coral mucus production and result in the expulsion of zooxanthellae and eventual bleaching and algal overgrowth in coral colonies. Often these waters are chlorinated to limit growth of fouling organisms, which increases the effects of the effluents on reef areas. The chlorinated effluents contain compounds that are not biodegradable and can circulate in the environment for years, bringing about a reduction in photosynthesis, with blooms of blue/green and red algae. Chlorinated hydrocarbon compounds include aldrin, lindrane, dieldrin, and even the banned DDT. These oxidating compounds are absorbed by phytoplankton and in turn by filter-feeding corals. Through the complex reef food webs these compounds concentrate in carnivorous fishes, which are often poisonous to humans. Many airborne particles are also deposited over coral reefs, such as fertilizer dust, dust from construction activities and cement dust. At Ras Baridi, on the Red Sea coast of Saudi Arabia, a cement plant that operates without filtered chimneys discharges over 100 t d 1 of partially processed cement over the nearby coral reefs, which are now smothered by over 10 cm of fine silt.



Solid Waste Dumping The widespread dumping of waste into the seas has continued for decades, if not centuries. Plastics, metal, wood, rubber, and glass can all be found littering coral reefs. These wastes are often not biodegradable, and those that are can persist over long periods. Damage to reefs through solid waste dumping is primarily physical. Solid wastes damage coral colonies at the time of dumping, and thereafter through natural tidal and surge action. Sometimes the well-intentioned practice of developing artificial reefs backfires and the artificial materials are thrown around by violent storms, wrecking nearby reefs in the process.

Construction Construction activities have had a major effect on reef habitats. Such activity includes coastal reclamation works, port development, dredging, and urban and industrial development. A causeway across Abu Ali bay in the northern Arabian Gulf was developed right over coral reefs, which today no longer exist. Commercial and residential property developments in Jeddah, on the Red Sea, have filled in reef lagoon areas out to reef crest and bulldozed rocks over reef crest for protection against erosion and wave action. Activities of this type result in increased levels of sedimentation as soils are nearly always dumped without the benefit of screens or silt barriers. Siltation is invariably the consequence of poorly planned and poorly implemented construction and coastal development, which can result in removal of shoreline vegetation and sedimentation. Coral polyps, although able to withstand moderate sediment loading, cannot displace the heavier loads and perish through suffocation. Partial smothering also limits photosynthesis by zooxanthellae in corals, reducing feeding, growth, and reproductive rates. The development of ports and marinas involves dredging deep channels through reef areas for safe navigation and berthing. Damage to reefs comes through the direct removal of coral colonies, sediment fallout, churning of water by dredger propellers, which increases sediment loads, and disruption of normal current patterns on which reefs depend for nutrients. Landfilling is one of the most disruptive activities for coastal and marine resources, and has caused severe and permanent destruction of coastal habitats and changed sedimentation patterns that damage adjacent coral resources. Changes in water circulation caused by landfilling can alter the distribution

675

of coral communities through redistribution of nutrients or increased sediment loads.

Recreation The recreation industry can cause significant damage to coral reefs. Flipper damage by scuba divers is widespread. Some will argue that today’s divers are more environmentally conscious and avoid damaging reefs, but certain activities, such as irresponsible underwater photography finds divers breaking corals to get at subjects and trampling reef habitats in order to get the ‘perfect shot’. In areas where divers walk over a reef lagoon and crest to reach the deeper waters, there is a degree of reef trampling, heightened in cases where entry and exit points are limited. Anchor damage from boats is a common problem at tourist destinations. In South-East Asia many diving operations are switching to nonanchored boat operations, but many others continue the practice unabated. Large tracts of reef can be found in Malaysia that have been scoured by dragging anchors, breaking corals and reducing reef crests to rubble. Experiments have shown that repeated break-age of corals, such as is caused by intensive diving tourism, may lead to substantially reduced sexual reproduction in corals, and eventually to lower rates of recolonization. In the northern Red Sea, another popular diving destination, and in the Caribbean, efforts are underway to install permanent moorings to minimize the damage to reefs from anchors.

Shipping and Port Activities Congested and high-use maritime areas such as narrow straits, ports, and anchorage zones often lead to physical damage and/or pollution of coral reef areas. Ship groundings and collisions with reefs occur in areas where major shipping routes traverse coral reef areas, such as the Spratley Island complex in the South China Sea, the Red Sea, the Straits of Bab al Mandab and Hormuz, and the Gulf of Suez, to name only a few. Major groundings have occurred off the coast of Florida in the United States, such as the one off Key Largo in State park waters in the 1980s, causing extensive damage to coral reefs. Often these physical blows are severe and destroy decades, if not centuries, of growth. Fish and other invertebrates lose their refuges and foraging habitats, while settlement of new colonies is restricted by the broken-up nature of the substrate. Seismic cables towed during seabed surveys and exploration activities may damage the seabed. Cable damage from towing of


676


vessels (e.g., a tug and barge) has been reported snagging on shallow reefs in the Gulf of Mexico, causing acute damage to sensitive reefs. Discharges from vessels include untreated sewage, solid wastes, oily bilge, and ballast water. On the high seas these do not have a major noticeable effect on marine ecosystems, but close to shore, particularly at anchorages and near ports, the effects become more obvious. At low tides, oily residues may coat exposed coral colonies, and sewage may cause localized eutrophication and algal blooms. Algal blooms in turn deplete dissolved oxygen levels and prevent penetration of sunlight. Port activities can have adverse effects on nearby reefs through spills of bulk cargoes and petrochemicals. Fertilizers, phosphates, manganese, and bauxite, for instance, are often shipped in bulk, granular form. These are loaded and offloaded using massive mechanical grabs that spill a little of their contents on each haul. In Jordan, the death of corals was up to four times higher near a port that suffered frequent phosphate spills when compared to control sites. The input of these nutrients often reduces light penetration, inhibits calcification, and increases sedimentation, resulting in slower feeding and growth rates, and limited settlement of new larva.

War-related Activities The effects of war-related activities on coral reef health and development are often overlooked. Nuclear testing by the United States in Bimini in the early 1960s obliterated complete atolls, which only in recent years have returned to anything like their original form. This redevelopment is nothing like the original geologic structure that had been built by the reefs over millenia. The effects of the nuclear fallout at such sites is poorly understood, and possibly has long-term effects that are not appreciable on a human timescale. The slow growth rate of coral reefs means that those blast areas are still on the path to recovery. Target practice is another destructive impact on reefs, such as occurred in 1999 in Puerto Rico, where reefs were threatened by aerial bombing practice operations. In Saudi Arabia, offshore islands were used for target practice prior to the Gulf war in 1991. The bombs do not always impact reefs, but those that do cause acute damage that takes long periods to recover. In the Spratley islands, the development of military structures to support and defend overlapping claims to reefs and islands has brought about the destruction of large tracts of coral reefs. Man-made

islands, aircraft landing strips, military bases, and housing units have all used landfilling to one extent or another, smothering complete reefs and resulting in high sediment loads over nearby reefs. Dredging to create channels into reef atolls has also wiped out extensive reef areas.

Indirect Effects Most anthropogenic effects and disturbances to coral reefs are easily identifiable. Blast fishing debris and discarded fishing nets can be seen. Pollutant levels and sediment loads can be measured. However, many other man-made changes can have indirect impacts on coral reefs that are more difficult to link directly to coral mortality. Global warming is generally accepted as an ongoing phenomenon, resulting from the greenhouse effect and the buildup of carbon dioxide in the atmosphere. Temperatures generally have risen by 1–21C across the planet, bringing about secondary effects that have had noticeable consequences for coral reefs. The extensive coral beaching event that took place in 1998, which was particularly severe in the Indian Ocean region, is accepted as having been the result of surface sea temperature rise. Bleaching of coral colonies occurs through the expulsion of zooxanthellae, or reductions in chlorophyll content of the zooxanthellae, as coral polyps become stressed by adverse thermal gradients. Some corals are able to survive the bleaching event if nutrients are still available, or if the period of warm water is short. Coupled with global warming is change of sea level, which is predicted to rise by 25 cm by the year 2050. This sea level rise, if not matched by coral growth, will mean corals will be submerged deeper and will not receive the levels of sunlight required for zooxanthellae photosynthesis. Additionally, the present control of erosion by coral reefs will be lost if waves are able to wash over submerged reefs. Coral reef calcification depends on the saturation state of carbonate minerals in surface waters, and this rate of calcification may decrease significantly in the future as a result of the decrease in the saturation level due to anthropogenic release of CO2 into the atmosphere. The concentration of CO2 in the atmosphere is projected to reach twice the preindustrial level by the middle of the twenty-first century, which will reduce the calcium carbonate saturation state of the surface ocean by 30%. Carbonate saturation, through changes in calcium concentration, has a highly significant short-term effect on coral calcification. Coral reef organisms do not seem to be able to acclimate to the changing



saturation state, and, as calcification rates drop, coral reefs will be less able to cope with rising sea level and other anthropogenic stresses.

The Future Mankind has contributed to the widespread destruction of corals, reef areas, and their associated fauna through a number of acute and chronic pollutant discharges, through destructive processes, and through uncontrolled and unregulated development. These effects are more noticeable in developing countries, where social and traditional practices have changed without development of infrastructure, finances, and educational resources. Destructive fishing pressures are destroying large tracts of reefs in South-East Asia, while the development of industry affects reefs throughout their range. If mankind is to be the keeper of coral reefs into the coming millennium, there is going to have to be a shift in fishing practices, and adherence to development and shipping guidelines and regulations, along with integrated coastal management programs that take into account the socioeconomic status of people, the environment, and developmental needs.

Glossary Ahermatypic Non-reef-building corals that do not secrete a calcium carbonate skeletal structure. DDT Dichlorodiphenyltrichloroethane. Dinoflagellates One of the most important groups of unicellular plankton organisms, characterized by the possession of two unequal flagella and a set of brownish photosynthetic pigments. Eutrophication Pollution by excessive nutrient enrichment. Gastrodemal The epithelial (skin) lining of the gastric cavity. Hermatypic Reef-building corals that secrete a calcium carbonate skeletal structure. Quinaldine A registered trademark fish narcotizing agent. Scleractinians Anthozoa that secrete a calcareous skeleton and are true or stony corals (Order Scleractinia). Zooxanthellae Symbiotic algae living within coral polyps.

677

See also Anthropogenic Trace Elements in the Ocean. Carbon Cycle. Eutrophication. Gas Exchange in Estuaries. Metal Pollution. Oil Pollution. Past Climate from Corals. Pollution: Effects on Marine Communities. Seabird Population Dynamics.

Further Reading Birkland C (1997) Life and Death of Coral Reefs. New York: Chapman and Hall. Connel DW and Hawker DW (1991) Pollution in Tropical Aquatic Systems. Boca Raton, FL: CRC Press. Ginsburg RN (ed.) (1994) Global Aspects of Coral Reefs: Health, Hazards and History, 7–11 June 1993, p. 420. Miami: University of Miami. Hatziolos ME, Hooten AJ, and Fodor F (1998) Coral reefs: challenges and opportunities for sustainable management. In: Proceedings of an Associated Event of the Fifth Annual World Bank Conference on Environmentally and Socially Sustainable Development. Washington, DC: World Bank. Peters EC, Glassman NJ, Firman JC, Richmonds RH, and Power EA (1997) Ecotoxicology of tropical marine ecosystems. Environmental Toxicology and Chemistry 16(1): 12--40. Salvat B (ed.) (1987) Human Impacts on Coral Reefs: Facts and Recommendations, p. 253. French Polynesia: Antenne Museum E.P.H.E. Wachenfeld D, Oliver J, and Morrisey JI (1998) State of the Great Barrier Reef World Heritage Area 1998. Townsville: Great Barrier Reef Marine Park Authority. Wilkinson CR (1993) Coral reefs of the world are facing widespread devastation: can we prevent this through sustainable management practices. In: Proceedings of the Seventh International Coral Reef Symposium Guam, Micronesia. Mangilao: University of Guam Marine Laboratory. Wilkinson CR and Buddemeier RW (1994) Global Climate Change and Coral Reefs: Implications for People and Reefs. Report of the UNEO-IOC-ASPEI-IUCN Task Team on Coral Reefs. Gland: IUCN. Wilkinson CR, Sudara S, and Chou LM (1994) Living coastal resources of Southeast Asia: Status and review. In: Proceedings of the Third ASEAN-Australia Symposium on Living Coastal Resources, vol. 1. Townsville: ASEAN-Australia Marine Science Project, Living Coastal Resources.


COSMOGENIC ISOTOPES D. Lal, Scripps Institute of Oceanography, University of California San Diego, La Jolla, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 550–560, & 2001, Elsevier Ltd.

Introduction In different settings, spanning from the extraterrestrial to the terrestrial, naturally occurring nuclides offer unique possibilities for being deployed as tracers for studying a great variety of physical, chemical, and biological processes, occurring over a wide range of timescales. This article discusses the continuous production of several stable and radioactive isotopes as a result of nuclear reactions of cosmic ray particles in the Earth’s atmosphere and the hydrosphere, and their potential usefulness as tracers for studying oceanic processes. The great merit of cosmic ray produced (cosmogenic) isotopes as tracers lies in the fact that their source functions in the different geospheres can be determined, and that several nuclides with a wide range of half-lives and chemical properties are available. Cosmic radiation, which consists of energetic H, He and heavier nuclei, with kinetic energies much greater than tens of mega electronvolts (MeV), (with particles of energies much beyond 1010 MeV), produce a great variety of nuclides by their interactions with target nuclei in the atmosphere, hydrosphere and the lithosphere. The predominant cosmic ray interaction is fragmentation of the target nuclei by primary and secondary particles of the cosmic radiation. Some nuclides are produced following the capture of thermal (very slowly moving) neutrons by target nuclei, which are abundant in the secondary cosmic radiation as a result of slowing down of fast neutrons emitted in energetic cosmic ray-produced nuclear reactions. Radiocarbon, 14C, was the first cosmic ray-produced isotope to be discovered in 1947 in sewage methane. Soon thereafter it was applied for archaeological/anthropological dating. This discovery was a milestone in the use of cosmic ray-produced (cosmogenic) isotopic changes as a tool for learning about planetary sciences. It laid the foundations of the field of cosmic ray geophysics/geochemistry. Subsequently, in the early 1960s, about 25 cosmogenic radionuclides produced in the earth’s atmosphere, with half-lives ranging from B1 h to millions

678

of years were detected. The driving force for the studies of cosmic ray-produced nuclides was the realization that if they could be detected in different dynamic reservoirs of the geospheres, they could be used as tracers to obtain important information about the timescales involved in the transport of materials through the atmosphere to the hydrosphere, oceans, and the cryosphere, and that in some cases they could be used as clocks to introduce timescales into the diverse proxy records of earth’s climate. Oceans are central to the dynamic interplay between the dynamic reservoirs, and therefore considerable emphasis has been placed on understanding the nature of the principal mixing/transfer processes, of the marine biogeochemical cycles, and of the large-scale ocean circulation. All oceanic investigations, in one way or the other, are linked to the central question of what processes control the earth’s climate. Geochemical tracers serve as tools to understand these processes and their rates. As mentioned above, the field of cosmic ray produced (cosmogenic) tracers caught roots in 1947, with the discovery of 14C. It grew rapidly thereafter in the late 1950s/early 1960s, and to date it is still one of the frontier areas in modern geochemistry. There are two reasons for this sustained hold and value of the cosmogenic tracers: continued development of new and powerful techniques for measuring their distribution in natural settings at very low concentrations, and the emergence of new biogeochemical questions which crop up as our understanding of the terrestrial climate system improves. However, there are often no other (suitable) tracers available to study the short- and long-term behavior of oceans on large space scales. This article considers the essentials of the cosmogenic tracers, their potentials, and how new advances continue to keep this field growing.

Terrestrial Cosmogenic Isotopes and their Production Rates Most of the cosmic ray energy (498%) is dissipated in the earth’s atmosphere in the nuclear reactions they produce. The atmospheric column represents about 13 mean free paths for nuclear interactions of fast protons and neutrons. After traversal through the atmosphere, the secondary particles of the cosmic radiation continue to produce nuclear reactions with the surficial terrestrial reservoirs: the hydrosphere,


COSMOGENIC ISOTOPES

the cryosphere, and the lithosphere, but at a much lower rate. The techniques of the 1950s for studying the cosmogenic nuclides were barely adequate to study the isotopes produced in the atmosphere. However, because of their value in understanding geophysical and geochemical processes, they were studied fairly extensively until the 1970s. A number of technical developments in the 1970s have made it easier to study the ‘atmospheric’ cosmogenic isotopes in the ocean environs, and even the cosmogenic isotopes produced in situ in terrestrial materials, including the hydrosphere and the cryosphere. The cosmogenic tracer-based information obtained from studies of the lithosphere and the cryosphere is of great value in interpreting the oceanic records. Table 1 lists a suite of isotopes, which serve (or should serve) as useful tracers in geophysical and geochemical studies. This list includes nominally potentially useful 20 nuclides, having half-lives exceeding two weeks, which are produced in the atmosphere, hydrosphere and in the lithosphere. The target elements from which they are produced in the earth’s atmosphere, and from principal elements in surficial matter are also listed in Table 1. This article is concerned primarily with nuclides that are useful as tracers in oceanography, and therefore shorter-lived nuclides have been excluded from Table 1. Since cosmic ray intensity is appreciably reduced at sea level due to nuclear interactions in the atmosphere isotope production rates (per gram target element per second), in surficial materials, are appreciably smaller than in the atmosphere. Therefore, nuclides, which can be produced in nuclear interactions with the major elements present in the atmosphere, N and O, have their principal source in the atmosphere. The next most abundant element, Ar, in the atmosphere occurs at an abundance of only 0.93% (v/v) in the atmosphere. Nuclear interactions with surficial materials can therefore be an important source of some of the nuclides produced from Ar, and for those isotopes which have mass numbers greater than 40, since 40Ar is the most abundant nuclear isotope in the atmosphere. Permanent constituent gases heavier than Ar have very low abundances in the atmosphere. The abundances of the next heavier gases, Kr and Xe, are B1 and 0.1 ppm (v/v), respectively. The production rates of several cosmogenic isotopes in the earth’s atmosphere are given in Table 2, along with their estimated global inventories. Some of the cosmogenic isotopes are also produced directly, in situ, in the upper layers of the oceans. The source strengths of cosmic ray-produced nuclei in the oceans, due to their production in the atmosphere, and direct production in the ocean water, are given in

679

Table 1 Potentially useful cosmogenic nuclides (arranged in order of mass numbers) with half-lives exceeding 2 weeks, produced in the Earth’s atmosphere and in surficial materials Nuclide(s)

Half-life

Main targets a

3

12.3 y S 53 d 1.5 106 y 5730 y S 2.6 y 7.1 105 y B150 y 14.3 d 25.3 d 87 d 3.0 105 y 35 d 268 y 1.0 105 y 3.7 106 y 312 d 7.6 104 y 1.5 106 y 2.3 105 y 1.6 107 y

O, Mg, Si, Fe (N, O) O, Mg, Si, Fe (N, O) O, Mg, Si, Fe (N, O) O, Mg, Si, Fe (N, O) O, Mg, Si, Fe (N) Mg, Al, Si, Fe Mg, Al, Si, Fe (Ar) Si, Al, Fe (Ar) (Ar) (Ar) (Ar) Fe, Ca, K, Cl (Ar) Fe, Ca, K, Cl (Ar) Fe, Ca, K (Ar) Fe, Ca, K (Ar) Ca, Fe (Kr) Fe (Kr) Fe (Kr) Ni, Fe (Kr) Ni (Kr) Rb, Sr, Zr (Kr) Te, Ba, La, Ce (Xe)

H He, 4He 7 Be 10 Be 14 C 20 Ne, 21Ne, 22 Na 26 Al 32 Si 32 P 33 P 35 S 36 Cl 37 Ar 39 Ar 41 Ca 53 Mn 54 Mn 59 Ni 60 Fe 81 Kr 129 I 3

22

Ne

a Elements from which most production occurs; those in parentheses give the main targets from which they are produced in the Earth’s atmosphere. s, stable.

Table 3. Atmospheric production is the dominant source of all nuclides in Table 3, except for 36Cl, where its in situ oceanic production exceeds the atmospheric production by about 50%. In the case of 37 Ar the two source strengths are comparable, and in the cases of 32P and 33P in situ production in the oceans is an order of magnitude lower. Isotopes produced in the earth’s atmosphere (Table 2), are introduced in the upper ocean: 1. in wet precipitation, in the case of isotopes which are removed directly (3H), or as attached to aerosols (7Be, 10Be, 22Na, 26Al, 32,33P, 32Si, 35S, 36 Cl); 2. by air–sea exchange of 14C (as 14CO2), and of isotopes of rare gases (3He, 37Ar, 39Ar and 81Kr). Besides direct in situ production of isotopes in the ocean waters (Table 2), some isotopes are also introduced to the oceans with river runoff as a result of weathering of the crustal materials in which they are produced, e.g., 10Be, 26Al, 41Ca, and 53Mn (Table 2). To date, those introduced by weathering of crustal materials have not been either studied or identified as important, and estimates of the strength of this source, are not presented here.


680

COSMOGENIC ISOTOPES

Table 2

Production rates of several isotopes in the Earth’s atmosphere; arranged in order of decreasing half-livesa

Isotope

Half-life

Production rate (atoms cm Troposphere

3

He Be 26 Al 81 c Kr 36 Cl 14 C 39 d Ar 32 Si 3 H 22 Na 35 S 7 Be 37 Ar 33 P 32 P

Stable 1.5 106 7.1 105 2.3 105 3.0 105 5730 y 268 y B150 y 12.3 y 2.6 y 87 d 53 d 35 d 25.3 d 14.3 d

10

6.7 10 1.5 10 3.8 10 5.2 10 4 10 4 1.1 4.5 10 5.4 10 8.4 10 2.4 10 4.9 10 2.7 10 2.8 10 2.2 10 2.7 10

y y y y

2 2 5 7

3 5 2 5 4 2 4 4 4

2

s

1

)

Global inventory

Total atmosphere 0.2 4.5 10 1.4 10 1.2 10 1.1 10 2.5 1.3 10 1.6 10 0.25 8.6 10 1.4 10 8.1 10 8.3 10 6.8 10 8.1 10

3.2 103 tonsb 260 tons 1.1 tons 8.5 kg 15 tonse 75 tons 52 kg 0.3 kg 3.5 kg 1.9 g 4.5 g 3.2 g 1.1 g 0.6 g 0.4 g

2 4 6 3

2 4

5 3 2 4 4 4

a

Based on Lal and Peters (1967). The inventory of this stable nuclide is based on its atmospheric inventory, which includes an appreciable contribution from crustal degassing of 3He. c Based on atmospheric 81Kr/Kr ratio of (5.2 7 0.4) 10 13. d Based on atmospheric 39Ar/Ar ratio of (0.107 7 0.004) d.p.m. l 1 Ar (STP). e Includes a rough estimate of 36Cl produced by the capture of neutrons at the earth’s surface. b

Table 3 Source functions of cosmogenic nuclides in the oceans, for nuclides of half-lives 410 days, arranged in order of increasing half-lives Nuclide

32

P P 37 Ar 7 Be 35 S 22 Na 3 H 32 Si 39 Ar 14 C 41 Caa 81 Kr 36 b Cl 26 Al 10 Be 33

Half-life

14.3 d 25.3 d 35.0 d 53.3 d 87.4 d 2.6 y 12.3 y B150 y 269 y 5,730 y 1.0 105 2.1 105 3.0 105 7.2 105 1.6 106

y y y y y

Principal target element (s)

Global average surface injection ratea

Atmosphere

In ocean water

(atoms cm

Ar Ar Ar N,O Ar Ar N,O Ar Ar N,O — Kr Ar Ar N,O

Cl, S, K Cl, S, K K, Ca O Cl, Ca, K Na O, 2H S, Ca K, Ca O Ca Sr Cl S, K, Ca O

5.82 10 3 6.93 10 3 9.10 10 6 1.27 2.84 10 2 3.75 10 3 1.39 101 9.60 10 3 2.00 10 1 1.20 102 — 2.30 10 5 6.60 10 2 8.40 10 3 2.70

2

min

1

)

Integrated in situ oceanic production rate (atoms cm

2

min

1

)

7.6 10 4 2.9 10 4 8.1 10 6 6.0 10 3 5.1 10 4 3.9 10 4 1.2 10 2 2.5 10 5 1.2 10 5 9.0 10 3 2.4 10 5 (n)a 1.9 10 8 (n)c 1.06 10 1 (n)c 6.8 10 6 1.8 10 3

Flux to oceans from rivers should be included to take into account production in rocks and soil by 40Ca(n, g) 41Ca reaction; this estimate is not given here because of large uncertainties in these calculations. b As above, due to 35Cl(n,g) 36Cl reaction. c (n), The in situ production of 41Ca, 81Kr and 36Cl in the oceans is primarily due to the relevant thermal neutron capture reaction. Note the 50% greater in situ production of 36Cl in the oceans compared to its atmospheric source.

a


681

0.84 4.7 10 0.11 10 4 0 14.3 d

Pathways of Isotopes to the Oceans and their Approximate Inventories/ Concentrations in the Atmosphere, Hydrosphere, and Sediments

0.80 5.6 10 0.13 7 10 4 0 25.3 d

b

a

2

3 2

Approximate calculations based on Lal and Peters (1967). Values given as zero imply very small fractional inventories. Part of the inventory may in fact be carried as silt or dust to the oceans before decay.

0.27 0.21 0.44 8 10 0 2.6 2

7.2 x 10 0.27 0.35 0.3 0 12.3 Atmosphere Land surface Mixed oceanic layer Deep oceanic layer Oceanic sediments Half-life (y)

2.3 10 3 0.29b 5.7 10 6 10 4 0.71 1.5 106

1.4 10 6 0.29b 1.4 10 5 7 10 5 0.71 7.1 105

1.1 10 6 0.29b 1.4 10 2 0.69 0 3.0 105

0.96 0 6 10 4 3.5 10 2 0 2.3 105

1.9 10 4 10 2 2.2 10 0.92 4 10 3 5730

2

2.0 10 0.29b 3.5 10 0.68 2.8 10 B150

3

0.99 0 0 0.01 0 268

H

3

Ar 39

Si

32

C

14

Kr 81

Cl

36

Al

26

Be

10

Radioisotope Exchange reservoir

Table 4

Approximate steady-state fractional inventories of cosmic ray produced radioisotopes in exchange reservoirsa

22

Na

2

0.65 0.1 0.24 4 10 0 87 d

35

S

3

0.71 0.08 0.20 2 10 0 53 d

7

Be

3

0.99 0 0 0 0 35 d

37

Ar

33

P

2

32

P

2

COSMOGENIC ISOTOPES

The applications of cosmogenic isotopes as tracers depend on three principal factors: (1) their source function; (2) their half-lives; and (3) their chemical properties. These considerations decide how the fractional inventories of different tracers are distributed on the earth in the atmosphere, hydrosphere, and the sediments. The work of Lal and Peters, using simplified models for the pathways of the isotopes considering six mixing/exchange reservoirs is still quite instructive. These estimates of fractional inventories of 14 isotopes amongst these reservoirs are shown in Table 4. Table 4 shows that most of the global inventory of the long-lived isotopes, 10Be and 26Al is in the oceanic sediments whereas that of another longlived isotope, 36Cl is in the oceans; this is a manifestation of their chemical properties. Analogous to 36 Cl, the chemical behavior determines the large fractional inventories of 14C and 32Si in the oceans. The inventories of the long-lived 81Kr, and also of 39 Ar (B270 y half-life), are primarily in the atmosphere primarily because of the low abundances of Kr and Ar in the atmosphere. It should be noted that generally the applications of an isotope are favored in the reservoir where its inventory is the largest. However, this is not always true. For example, in the cases of 39Ar and 33,32P, in spite of their low inventories in the oceans, they have valuable applications in studies of oceanic processes. Approximate theoretical estimates of isotope concentrations in the oceans (disintegrations per minute (d.p.m.) per tonne of sea water) are presented in Table 5. The values are in the range of 10 5–250 d.p.m. t 1. The corresponding specific elemental concentrations are very low, with isotope/element ratios lying in range of 10 19–10 10. The concentrations of a large number of naturally occurring radioactive and stable nuclides, those produced in nuclear reactions caused by cosmic radiation, and those produced by energetic particles in radioactive disintegrations and in nuclear decays of naturally occurring long-lived nuclides, have been measured in the past five decades in the marine environment. Dramatic improvements in the radiometric techniques in the past two decades have allowed their measurements to be done fairly reliably. The database on the distribution of the cosmogenic and other tracers in the oceans is therefore growing steadily. Their


682

COSMOGENIC ISOTOPES

Table 5 Approximate specific radioactivities of cosmic ray produced isotopes in the ocean Radioisotope

10

Be Al 81 a Kr 36 Cl 14 C 32 Si 39 b Ar 3 H 26

Half-life (y)

1.5 106 7.1 105 2.3 105 3.0 105 5730 B150 268 12.3

Average specific radio activity in oceans d.p.m. per tonne water

d.p.m. per g element

10 3 1.2 10 7 10 6 0.55 260 1.5 10 2.9 10 36

1.6 103 1.2 10 3 2.1 10 2 3 10 5 10 2.3 10 2 5 10 3 3.3 10 4

5

2 3

Based on atmospheric 81Kr/Kr ratio of (5.2 7 0.4) 10 13. Based on atmospheric 39Ar/Ar ratio of (0.107 7 0.004) d.p.m. per liter Ar. (Based on Lal and Peters (1967)). a

b

measurements to date are in good agreement with the theoretically predicted values of their distribution in the oceans (cf. Tables 3 and 4). In many cases these nuclides serve as tracers for the study of physical, chemical, and biological processes in the oceans. Several radiotracers successfully provide chronology of sediments, corals, and manganese nodules, but learning about large-scale ocean circulation is another matter. The ability merely to make measurements of a tracer in the marine environment is not sufficient to use it as an effective tracer for delineating important oceanic variables. Tracer data must be examined in terms of ocean models. Constructing ocean models is an iterative process between data acquisition and model building, forcing model outputs, to become compatible with the observations. The oceanic processes are very complex, exhibiting significant spatial and temporal variability on a wide range of scales. For the tracer data to be useful in developing meaningful coupled atmosphere–ocean circulation models, which may be considered as the goal of tracer studies, one would require three-dimensional tracer data with sufficient resolution in the horizontal direction. The latter are not available, except for 14C, where a considerable database is growing as a result of recent WOCE (World Ocean Circulation Experiment) expeditions.

Tracers in Oceanography: Why We Need Them and What We Learn From Them The oceans represent a large mass of water endowed with a large amount of diverse substances and heat.

The dissolved and particulate oceanic ‘complex’ is in continuous exchange with the land surface and the atmosphere. An appreciable part of the dissolved phases is recycled within the oceans through biogeochemical cycles, which are maintained by the large-scale oceanic circulation. The latter is a manifestation of the continuous exchange of heat between the atmosphere and the ocean. Large-scale oceanic circulation replenishes nutrients in the surface waters, which are rapidly removed by biological productivity. Biological recycling changes the chemical makeup of ocean waters at all depths. Thus there is a complex cause–effect relationship with significant feedbacks between oceanic circulation, biogeochemical cycling within the oceans, and composition of sea water. Understanding these processes is essential for understanding oceanographic processes, earth’s climate, terrestrial biogeochemistry, and the proxy records contained in the oceanic sediments. Success in achieving this goal requires sensitive multidisciplinary techniques in which tracers play an important part. Chemical and isotopic tracers have been used successfully for the past five decades. Oceanic water masses are conventionally characterized by their chemical and isotopic composition, and temperature. A central problem in oceanography is to understand the origins and the processes which determine the evolution of different water masses. Radioactive isotope tracers provide additional information on timescales, specifically on the rate constants of different processes. The most attractive feature of radioisotopes is that they provide time integrals of evolution of water masses through space, influenced by exchange/mixing processes, and radioactive decay of the tracer, which introduces the element of time in the model(s). In steadystate situations, all losses and gains balance out. By combining with information on stable isotopes, one can then determine effective time required for the water mass to reach equilibrium between gain and loss terms, i.e. get an estimate of the effective equilibration time of the water mass as it evolves. Tracers fall into two broad categories: 1. Transient tracers which are introduced sporadically in a system, e.g., radionuclides introduced by testing of nuclear weapons, and from discharges from nuclear reactors. 2. Steady tracers which are introduced continuously in a system, e.g., those produced by nuclear interactions of cosmic rays on the earth, and by radioactive decay of dissolved uranium in the oceans.


COSMOGENIC ISOTOPES

It is convenient to further designate the tracers according to their chemical behavior in the system: 1. conservative tracers, which follow the motion of the ‘fluid’ in the system; 2. nonconservative tracers, which do not follow the motion of the ‘fluid’ in the system. The first naturally occurring radiotracer to be used in oceanography was 226Ra. The first successful tracer measurements of the cosmogenic 14C with a view to understanding timescales in large-scale water circulation were made in 1960 and demonstrated the great value of this tracer in oceanography. The discovery of cosmogenic 32Si in marine siliceous sponges opened up the possibility of using this as a tracer for studying biogenic silica fluxes to the deep sea, and the nutrient cycle of silicon. As the techniques for the measurement of weak activities of the nuclides became available, additional nuclides were measured in the oceans. To date 12 cosmogenic nuclides have been studied in the oceans, some during the 1960s, several during the 1980s. It is important to realize that all tracers are important because of their particular unique attributes (cf. Table 4). Table 6 lists tracers, which are studied in oceanographic research, together with 36Cl (which is included for its potential usefulness for determining the average source strength of cosmic ray neutrons in the past 0.5–0.7 My (million years)). The usefulness of cosmogenic tracers depends on their half-lives,

Table 6

683

chemical properties, and their source functions, which can be appreciated from their expected distribution in the geospheres (Table 4), and by the ease with which they can be measured. (Table 4 does not include 4He, 20Ne, 21Ne, 22Ne, 22Na, 35S, 36Cl, 37Ar, 41 Ca, 53Mn, 52Ni, 60Fe and 81Kr, which are either not useful as oceanic tracers because of their short halflives, or very long half-lives, or have very low cosmogenic production rates. However, with technical developments these nuclides may eventually become useful.) The long-lived cosmogenic radionuclide, 129I is not included in Tables 3 and 4 because (1) its halflife is rather long (15.7 My) to be useful for studying oceanic processes, (2) it is continuously produced in the oceans, and in ocean sediments in the spontaneous fission of 238U, and (3) it has been added to the oceans in appreciable amounts in the last five decades by human activities; such as nuclear weapons’ testing and processing of nuclear plants (which have raised the prenuclear age inventory of 129I in the oceans of about 100 kg by more than an order of magnitude. Several cosmogenic tracers also qualify as transient tracers at the present time, because of an appreciable contribution from anthropogenic sources (Table 3). Thus, the nuclides 3H, 14C (produced in appreciable amounts in nuclear weapons testing), tritugenic 3He and 129I (which has also been produced in large amounts by nuclear weapons tests and operation of nuclear power plants), serve as (useful) transient tracers in some geophysical reservoirs.

Important characteristics and principal applications of selected cosmogenic tracers

Isotope

Half-life

Principal applications

Isotopes which do not form compounds He Stable 37 Ar 35 d 39 Ar 268 y 81 Kr 2.3 105 y 3

Air–sea exchange; escape of helium from the atmosphere Air–sea exchange; tropospheric circulation Air–sea exchange; vertical mixing in oceans Ground water ages, and constancy of cosmic radiation

Isotopes which label constituent molecules in the atmosphere and the hydrosphere 3 H (H2O) 12.3 y Characterizing water molecules in the atmosphere, hydrosphere and cryosphere 14 C (CO2, CO3, HCO3) 5730 y Characterization of the carbon cycle reservoirs 32 Si (HSiO3, SiO2) B150 y Biogeochemical cycle of silicon 32 P, 33P (DIP, DOP) 14.3, 25.3 d Biogeochemical cycle of phosphorus Isotopes which attach to aerosols/particles 7 Be 53 d 10 Be 1.5 106 y 26

7.1 105 y

32

B150 y 14.3, 25.3 d

Al Si (HSiO3, SiO2) P, 33P

32

Note: Not included here are

36

Cl and

Atmospheric circulation, vertical mixing in surface ocean waters Role of particle scavenging in the coastal and open oceans; dating of sediments and accretions Role of particle scavenging in the coastal and open oceans; dating of marine sediments and accretions Labeling the dissolved oceanic silicon pool; atmospheric circulation Labeling the dissolved oceanic phosphorus pool; tropospheric circulation

129

I for reasons discussed in the text.


684

COSMOGENIC ISOTOPES

An important consideration in the use of a transient tracer is knowledge of its source strength. Anthropogenic sources are generally not well defined; however, in the case of 14C, an important advantage is that its contribution to the atmospheric CO2 reservoir is well known (i.e., the excess 14C amount relative to 12C), and this precisely defines its source function. In the case of anthropogenic 3H, this is not the case, but a great advantage is that one can measure both the 3H and tritugenic 3He in a water sample and defines an ‘age’ of the water mass. These measurements are by no means easy, however, but a large database of information has been produced which has yielded very useful insights into large-scale ocean circulation in the upper ocean. There are also two important nonnuclear transient tracers, chlorofluorocarbons CFC-11 and CFC-12, which have also proven very useful in view of their known (changing) relative concentrations in the atmosphere. These behave essentially as conservative tracers; any CFC losses would not be expected to change their ratio in the oceanic water mass. The use of new tracers (F113, CCl4) has extended the timescales of CFC in both directions. In practice, one has to work with tracers of different properties, and each of its properties can be taken advantage of, as its special attribute. Even nonconservative tracers, e.g., 14C and 32Si, have their own significance and merit. In fact, in the oceans the only conservative tracers of natural origin are 3H (half-life 12.3 y) and 3He (stable).

New Techniques for Measurements of Tracers in the Oceans in the 1980s and 1990s By the end of the 1970s, the field of cosmogenic tracers had clearly recognized the usefulness of most of the cosmogenic tracers, with sufficient measurements at hand in each case. After isolated studies of a few tracers, e.g., 14C, 10Be, in individual water samples, it became apparent that oceans can yield their secrets only with multiple tracer attack. In early multiple tracer investigations, detailed information regarding the nature and rate of processes responsible for the formation of the Antarctic Bottom water was obtained by including the tracers 3He and 14C. The field was expanded in the 1980s and 1990s with larger-scale exploitation of several tracers for answering specific questions. This came about due to a fruitful combination of events and discoveries, which gave a tremendous fillip to both chemical oceanography and tracer studies including nuclides belonging to U-Th series. Foremost was the decision to

study oceans in a systematic manner, along geochemical sections (GEOSECS), using a suite of tracers. GEOSECS expeditions were successfully carried out to the principal oceans in 1972–78 and resulted in fairly accurate tracer data. The GEOSECS concept was very successful; it rested on the necessity for making more precise measurements of several tracers and ocean properties in addition this integrated study resulted in information about temporal changes in the property profiles at the same stations after an elapse of one to two decades since the site was occupied in the GEOSECS expedition. and finally, it was an artful and timely combination of theory and experiment, which gave a tremendous boost to the field of learning about oceanographic processes. The 1980s also marked an era of dramatic advance in the techniques of measurements of long-lived cosmogenic radionuclides 14C, 10Be, and 26Al in the oceans, in sediments, and in manganese nodules using AMS (accelerator mass spectrometry). This opened up new windows for observing in detail a host of physical, chemical, and biological processes. The ease with which these nuclides can be measured allowed long series of measurements in space and time to be obtained. Examples of this development are the direct measurements of 10Be and 26Al concentration profiles in sea water in the principal oceans; and profiles of 10 Be concentrations in marine sediments and in manganese nodules which opened up a new field of investigation in marine beryllium geochemistry. A new field, the study of P-biodynamics in surface waters using cosmogenic 32P and 33P surfaced in the late 1980s. This tracer application was not held up for want of a technique. In this case, it was not realized that these short-lived nuclides (half-lives, 14.3 and 25.3 days) in fact had about the appropriate half-lives for studying timescales of exchange of phosphorus between dissolved inorganic P, organic P and plankton. Concurrently, technical advances were also being made to measure short-lived radionuclides in ocean waters, where the AMS technique does not offer any gain in detection sensitivity, e.g. 32P (half-life, 14.3 days), 33P (half-life, 25.3 days) and 32Si (half-life, 150 years). By using a standard liquid scintillation counting system to simultaneously measure both 32P and 33P activities, much higher sensitivity is attainable than by using low-level counters, especially for samples of low specific radioactivity.

Examples of Oceanic Data Derived Using Cosmogenic Tracers With this foreground what has been learnt about the oceanic processes using cosmogenic tracers is now


COSMOGENIC ISOTOPES

briefly examined. We are obviously interested in learning about cycles of principal nutrient elements, and processes by which the ocean chemistry evolves (as regulated by aeolian and fluvial fluxes from the land, biogeochemical cycles within the oceans, and large-scale oceanic circulation). This problem can be approached in bits and pieces only, and then the interconnections and feedbacks examined. A comprehensive mosaic of all the interactions and controls may or may not be achieved. The records of present day ocean biogeochemical processes are recorded in the sediments. It is therefore important to study the chronology and the makeup of ocean sediments to get a comprehensive picture of the temporal evolution of ocean chemistry and climate through aeons. The suite of tracers listed in Table 3 has provided sufficient information on oceanic processes in four broad fields. 1. Biogeochemical cycling of nutrients and trace elements 2. Chronology of marine sediments and manganese nodules 3. Principal features of large-scale oceanic circulation 4. Biogeochemical and ocean circulation controls on climate. In each of these studies, the task is complemented by the availability of radiotracers belonging to the UTh series. It should be stressed that tracers each have some particular unique features for studying critical problems in oceanography. All tracers are not created equal: some are more equal than others. This social expression also finds a rightful place in the realm of oceans. This can be illustrated by citing unique features of two of the cosmogenic nuclides, 14 C (half-lifem, 5730 years) and 10Be (half-life, 1.5 My), which have a special status among all natural tracers. The great ‘virtues’ or attributes of 14C are that (1) it is a carbon isotope, and is introduced in to the carbon cycle reservoirs as carbon dioxide in the earth’s atmosphere, and (2) its half-life is well suited to study late Quaternary events and processes, including dating of sediments and timing of deep and bottom water formation. In the oceans, 14C does not behave as a conservative tracer, since carbon (and its compounds) is not distributed uniformly in the oceans. But this does not present any problems; rather its studies allow determination of carbon fluxes within the ocean. In fact, if 14C was a conservative tracer, it could not have been used to date marine sediments. It has been added in significant amounts to the atmosphere as a result of nuclear weapons tests during the 1950s and

685

1960s. Consequently, 14C can be used both as a ‘steady-state’ and as a ‘transient’ tracer. The second most attractive ocean tracer is the cosmogenic 10Be, which serves to delineate pathways of particle active elements through the water column, and is useful for dating sediments and in manganese nodules to about 10 My BP. The particle active nature of 10Be leads to its preferential deposition in the coastal regions of the oceans. Recent studies have demonstrated that using special chemical techniques, the activity of cosmogenic 26Al can be measured in the oceanic environment, and it has been suggested that it should be a useful tracer for studying changes in the past biological productivity of the oceans. This application arises from the higher chemical reactivity of 26Al, compared to 10Be. If this suggestion is borne out from future studies, 26Al would constitute an invaluable tracer for studying temporal and spatial variations in biological productivity. Its studies would complement the information obtained using the cosmogenic 10Be. Recent measurements of cosmogenic 32P (half-life, 14.3 days) and 33P (half-life, 25.3 days) in surface ocean waters have opened up new possibilities of quantifying P-biodynamics with complementary information on eddy diffusivity in the waters, based on the cosmogenic 7Be (Table 3). A wealth of new 32P, 33 P data have been added on the distribution of cosmogenic 32P and 33P in the surface ocean waters, and in plankton.

Epilogue A large number of cosmogenic tracers are available for oceanic studies, and the use of these tracers has steadily increased to date. But of course, tracers are not the complete answer to the mysteries of the ocean. It is necessary to learn how to use tracers, how to model them, how to combine them with other tracers, singly and multiply, etc. The usefulness and application of a tracer cannot be discussed on an absolute basis, because such an approach would result in a largely academic discussion. A tracer may have the appropriate physical and chemical attributes, but its source strength may be too weak, or its source function may not be known at the present time. Tracer suitability has therefore to be evaluated periodically as frontiers of knowledge expand. With the freedom in thinking about what type of tracer measurements can be made, a great deal of valuable information will probably be derived from it; however, one has to think about applying it within practical constraints. An important constraint is the number of measurements,


686

COSMOGENIC ISOTOPES

which can be made within a reasonable period, and within available financial resources. Today, the principal constraint is the inadequacy of spatial and temporal coverage of the tracer data. With more synoptic data, tracer modeling could also be refined. Any shortcomings in these foil the goal of these studies. We are now in the mode of understanding details of ocean circulation and chemistry; not just the integrated overall features. In the earliest applications of cosmogenic tracers just a few measurements of radiocarbon in the oceans were sufficient to show that the upper ocean had a turnover time of about 600 years. In the case of the atmosphere, the mean time for removal of aerosols was similarly learned very quickly from the observed fallout of nuclear weapons tests-injected radioactivities (90Sr, 3 H, etc.) and from the fallout of cosmogenic 22 Na, 7Be and 32P. With continued applications of these tracers, it has been shown that oceanic processes are indeed complex, especially at the interfaces (air–sea, mixed layer–intermediate waters– deep waters, and water–sediment), and in the polar oceans, which determine formation of bottom–deep waters. Thus, our hasty perception of these tracers being a panacea quickly changed, even back in the 1970s. What is the basic nature of large-scale oceanic circulation? The radiochemists used the Kw model in its simple one-dimensional form; K denotes the eddy diffusivity, and w the advection velocity. At this time a stimulating discussion of essential mathematical approaches for treating the tracer data was presented. Simple material balance calculations by Lal showed that an appreciable amount of carbon was added to the dissolved carbon inventory (J-flux) at depths by sinking biogenic particles. This important aspect of resetting the 14C clock in the deep sea by Jflux is now being examined on a global basis under WOCE experiments, and constitutes a critical parameter in climatic feedback processes. Subsequently, tracer data showed that an important transfer of oceanic properties occurred across pycnoclines. Modeling of tracer data in fact reveals model inadequacies and fosters development of more appropriate physical models. There are several very basic issues that are recognized, but not well understand. For example, what are the roles of tides and internal waves in large-scale ocean circulation? These questions have been asked several times, but not yet attacked properly due to our present limitations. Today, we are far from a synthesis of comprehensive models which are capable of providing an interactive atmosphere–ocean coupled model which can respond to changes in climate, or predict climatic changes as the model is run.

There are several academic and technical issues which we are confronting today: 1. complexity of the ocean system; variable response at different time and space scales; 2. lack of three-dimensional tracer data; 3. lack of information on temporal and spatial changes in tracer distribution; 4. lack of understanding of physical, chemical, and biological processes. However, there are proven methods based on 14C, Be, 26Al, 39Ar, 32Si, 7Be, 33P and 32P. Improved tracer modeling will emerge only with further advances in techniques for their measurement, and with a better understanding of the atmospheric and oceanic circulation, mixing, and biogeochemical processes. On the question of ease of measurement, the radionuclide 39Ar is an important case in point here. It is a conservative tracer, ideally suited for studying vertical mixing in the oceans, but to date very few measurements have been made, since they are very time consuming. Another example is that of 32Si. Although it has been measured at several stations in the Atlantic, Pacific and Indian oceans, these measurements are not currently precise enough to make detailed mixing and transport models to define the silica cycle in the oceans. They are, however, useful to determine vertical J-fluxes, one-diemnsional K/o ratios, and the latitudinal inventories of 32Si in the oceans. Physical oceanography provides the theoretical basis for oceanic mixing and circulation, but the experimental data necessary to understand the nature of this circulation must be based on present day and proxy observations of chemical composition of sea water in space and time. Directed global scale coordination between scientists to study important oceanic processes, such as physical and biological controls on biological production, and export of carbon, are rapidly providing new insights and accelerated developments of realistic models. An example is the coordinated US Joint Global Ocean Flux Studies (JGOFS) in the Equatorial Pacific Ocean in 1992, during a four-month period which coincided with the maximum intensity of the warm El Ninõ event, and another three-month period during welldeveloped cool surface-water conditions. The combined physical/chemical and biological data produced, which included 234Th concentrations of sinking particulates, led to new insights about the roles of dissolved organic carbon, microzooplankton grazing, nutrient and CO2 fluxes, and highlighted the importance of physical, in contrast to biological, processes in this region, where net carbon fluxes out of the system are very small as a result of highly efficient biological cycling. 10


COSMOGENIC ISOTOPES

This discussion would not be complete without mention of the ‘purposeful tracer experiment’ carried out as a component of WOCE, using SF6 released in the open ocean. The results confirmed the earlier estimates of very low diapycnal diffusivity, of the order of 0.1 cm2 s 1, implying that heat, salt, and tracers must penetrate the thermocline primary by transport along, rather than across, density surfaces. Clearly, understanding of the large-scale ocean circulation will come from multiple approaches, with directed research to simultaneously understand transfer and mixing of ‘properties’ and ‘substances’ and their relationship to climatic changes. It has taken five decades to develop techniques to make relevant oceanographic measurements, and one should expect rapid advances in our knowledge of oceanic processes as a result of innovative research, and international observational programs such as Tropical Oceans and the Global Atmosphere (TOGA) and WOCE. It must be realized that a state must be reached where experiments and theory go hand in hand, leading to the development of better (more realistic) models, and acquisition of critical tracer data. In the absence of a knowledge of the processes involved, models employed often yield very erroneous results. Thus, whereas even a few tracer data are quite informative (since a few data points can be treated only with zero order models), any attempts to understand oceanic processes in detail pose a serious challenge. A few examples are considered here, where tracer data have contributed to the development of realistic models. As mentioned earlier, simple one-dimensional models were developed earlier on using two parameters K and w, to consider vertical transfer of tracers through an oceanic column. Even today these are used, in the absence of better alternatives, and in reality, because of a lack of tracer data in the threedimensional space. The result is that as yet the general validity of the K-w models in space is not known or their dependence on climate. The latter arises because there are experimental tracer data for ocean waters only during the Holocene. The recent significant developments in oceanic general circulation are a result of transient tracer experiments, and order of magnitude improvements in a number of fields, including orbit dynamics, gravity field estimation and atmospheric variability. High accuracy data on ocean surface elevation by satellite altimetry is leading to hopes of complete theee-dimensional time-evolving estimates of ocean circulation, which would also improve estimates of oceanic heat, and several property fluxes. There has been prolific growth in the field of tracer oceanography within the last two decades, but there are acute limitations in providing a consistent picture

687

of the interrelationships between physical, chemical, and biological processes, which are needed to develop a coupled atmosphere–ocean model that responds to climate in an interactive manner. It is not clear how this will be achieved in the near future. Finally, it is gratifying to see that the cosmogenic radiotracer field has evolved highly from an academic curiosity in the 1950s and 1960s to its presentday form, wherein it aims to become an integral part of realistic atmosphere–ocean global atmosphere and ocean circulation models.

See also Carbon Cycle. Carbon Dioxide (CO2) Cycle. Phosphorus Cycle. Radiocarbon. Stable Carbon Isotope Variations in the Ocean.

Further Reading Broecker WS (1981) Geochemical tracers and ocean circulation. In: Warren BA and Wunsch C (eds.) Evolution of Physical Oceanography, pp. 434--460. Cambridge, MA: MIT Press. Broecker WS and Peng TH (1982) Tracers in the Sea. New York: Lamont-Doherty Geological Observatory. Jenkins WJ (1992) Tracer based inferences of new primary production in the sea. In: Falkowski PG and Woodhead AD (eds.) Primary Productivity and Biogeochemical Cycles in the Sea. New York: Plenum Press. Lal D (1962) Cosmic ray produced radionuclides in the sea. J. Ocean. Soc. Japan: 20th Anniv. Vol. 600–614. Lal D and Peters B (1967) Cosmic ray produced radioactivity on the earth. Handbuch der Physik 46(2): 551--612. Lal D (1999) An overview of five decades of studies of cosmic ray produced nuclides in the oceans. Science of the Total Environment 237/238: 3--13. Ledwell JR, Watson AJ, and Law CS (1993) Evidence for slow mixing across the pycnocline from an open ocean tracer-release experiment. Nature 364: 701--703. Libby WF, Anderson EC, and Arnold JR (1949) Age determination by radiocarbon content: world-wide assay of natural radiocarbon. Science 109: 227--228. Measures CI and Edmond JM (1982) Beryllium in the water column of the Central Pacific. Nature 297: 51--53. Murray JW, Barber RT, Roman MR, Bacon MP, and Feely RA (1994) Physical and biological controls on carbon cycling in the equatorial Pacific. Science 266: 58--65. Raisbeck GM and Yiou F (1999) 129I in the oceans: origins and applications. Science of the Total Environment 237/ 23831--41. Schlosser P and Smethie WM Jr (1995) Transient tracers as a tool to study variability of ocean circulation. In: Natural Climate Variability on Decade-to-century Time Scales, pp. 274--289. Washington: National Research Council.


COUPLED SEA ICE–OCEAN MODELS A. Beckmann and G. Birnbaum, Alfred-WegenerInstitut fu¨r Polar und Meeresforschung, Bremerhaven, Germany Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 560–570, & 2001, Elsevier Ltd.

A number of important feedback processes between the components of the coupled system can be identified which need to be adequately represented (either resolved or parameterized) in coupled sea ice–ocean models (see Figure 1):

• •

Introduction Oceans and marginal seas in high latitudes are seasonally or permanently covered by sea ice. The understanding of its growth, movement, and decay is of utmost importance scientifically and logistically, because it affects the physical conditions for air–sea interaction, the large-scale circulation of atmosphere and oceans and ultimately the global climate (e.g., the deep and bottom water formation) as well as human activities in these areas (e.g., ship traffic, offshore technology). Coupled sea ice–ocean models have become valuable tools in the study of individual processes and the consequences of ice–ocean interaction on regional to global scales. The sea ice component predicts the temporal evolution of the ice cover, thus interactively providing the boundary conditions for the ocean circulation model which computes the resulting water mass distribution and circulation.

• • •

ice growth through freezing of sea water, the related brine release and water mass modification; polynya maintenance by continuous oceanic upwelling; lead generation by lateral surface current shear and divergence; surface buoyancy loss causing oceanic convection; and pycnocline stabilization in melting regions.

Not all of these are equally important everywhere, and it is not surprising that numerous variants of coupled sea ice–ocean models exist, which differ in physical detail, parameterizational sophistication, and numerical formulation. Models for studies with higher resolution usually require a higher level of complexity. The main regions for applying coupled sea ice– ocean models are the Arctic Ocean, the waters surrounding Antarctica, and marginal seas of the Northern Hemisphere (e.g., Baltic Sea, Hudson Bay). A universally applicable model needs to include (either explicitly or by adequate parameterization) the

Freezing Ridged sea ice

Coastal polynya Tides

Melting

Open ocean Leads polynya Brine release

Slope convection

Drift Heat transfer Freshwater flux Restratification

Momentum flux Open ocean convection

Dense water formation Large-scale overturning

Figure 1 Cartoon of coupled sea ice–ocean processes. Effects in the ocean are in italics.

688


COUPLED SEA ICE–OCEAN MODELS

specific mechanisms of each region, e.g., mainly seasonal variations of the ice cover or the presence of thick, ridged multiyear ice. This article describes the philosophy and design of large- and mesoscale prognostic dynamic–thermodynamic sea ice models which are coupled to primitive equation ocean circulation models (see General Circulation Models). The conservation principles, the most widely used parameterizations, several numerical and coupling aspects, and model evaluation issues are addressed.

Basics Sea water and sea ice as geophysical media are quite different; whereas the liquid phase is continuous, three-dimensional, and largely incompressible, the solid phase can be best characterized as granular, two-dimensional and compressible. Both share a high degree of nonlinearity, and many direct feedbacks between oceanic and sea ice processes exist. Today’s coupled sea ice–ocean models are Eulerian; granular Lagrangian models, which consider the floe–floe interaction explicitly, exist but have so far not been fully coupled to ocean circulation models. Originally designed for use with large-scale coarse resolution ocean models, sea ice in state-of-the-art models is treated as a continuous medium. Following the continuum hypothesis only the average effect of a large number of ice floes is considered, assuming that averaged ice volume and velocities are continuous and differentiable functions of space and time. Thus, very similar numerical methods are being applied to both sea water and sea ice, which greatly facilitates the coupling of models of these two components of the climate system. Conceptually, the continuum approach limits the applicability of sea ice models to grid spacings that are much larger than typical floes, i.e., several to several tens of kilometers. Typical values are 100–300 km for global climate studies, 10–100 km for regional climate simulations and about 2 km for process studies and quasi-operational forecasts. The latter cases are stretching the continuum concept for sea ice quite a bit, but still give reasonable results. State-of-the-art coupled sea ice–ocean models are based on two principles for the description of sea ice, the conservation of mass and momentum, covering its thermodynamics and dynamics. Mass conservation for snow is also taken into account. A snow layer modifies the thermal properties of the ice cover through an increased albedo and reduced conductivity. This leads to delayed surface melting and lower basal freezing rates. In the following description sea

689

ice is frozen sea water, and ice is the sum of sea ice and snow. The temporal change of sea ice and snow volume due to local sources/sinks and drift is described by the mass conservation equation @hði;sÞ thdyn þ r ui hði;sÞ ¼ Shði;sÞ @t

½1

where h(i, s) is the sea ice and snow volume per unit area (with h ¼ hi þ hs), respectively, vi is the twothdyn dimensional ice velocity vector and Shði;sÞ denotes the thermodynamic sources and sinks for sea ice and snow. The corresponding ice velocities are obtained from the momentum equation ðri hi þ rs hs Þ

- @ ui - þ ui rui þ f k ui þ grH @t -

-

-

½2

¼ tai tiw þ Fi where the local time rate of change, advection of momentum and the accelerations due to Coriolis are included. The wind stress tai is external to the coupled sea ice–ocean model; the ocean surface current stress tiw and the sea surface height H is part of the coupling to the ocean. The so-called ice stress term Fi summarizes all internal forces generated by floe– floe interactions.

Subgridscale Parameterizations The granular nature of the medium, combined with the strong sensitivity of both thermodynamics and dynamics on the number, size, and thickness of individual ice floes requires the inclusion of a subgridscale structure of the modeled ice. Ice Classes

An obvious assumption is that of a subgridscale ice thickness distribution. In this widely used approach the predicted ice volume h is thought to be the average of several compartments, the thermodynamic ice classes, which represent both thinner and thicker ice and possibly include open water. The relative contribution of an ice class is fixed (e.g., uniform between 0 and twice the average ice thickness). For each ice class, a separate thermodynamic balance is computed; the resulting fluxes are then averaged according to their relative areal coverage. Ice Categories

As the subgridscale distribution of ice thickness will change with time and location, an even more


690


sophisticated approach considers time-evolution of the volume of each compartment, which is then called an ice category. The form of these prognostic equations follows the conservation eqn [1]. Different subgridscale ice velocities are not taken into account; the advection of all compartments takes place with the resolved velocity field. Models with several ice categories are in use. A minimum requirement, however, has been identified in the discrimination between ice-covered areas and open water. Then the prognostic equation for ice volume [1] is accompanied by a formally similar equation for ice concentration, i.e., the percentage of ice-covered area per unit area A, @A þ r ðui AÞ ¼ Sthdyn þ Sthdyn A A @t

½3

Thermodynamic sources and sinks Sthdyn for ice A concentration are chosen empirically and involve parameterizations of subgridscale thermodynamic melting and freezing. The conceptual ansatz: 1A @h A @h thdyn ¼ max 0; min þ ; 0 ½4 SA hcls @t hopn @t describes the formation of new ice between the ice floes with the first term on the right hand side; here hcls is the so-called lead closing parameter. The second term parameterizes basal melting of sea ice with a similar approach involving hopn. The coefficients are often derived from the assumed internal structure of the ice-covered portion of the grid cell, the ice follows classes. The dynamic source/sink term Sthdyn A from the constitutive law (see section on dynamics below). Conceptually, the ice concentration (or compactness) A has to lie between 0 and 1, which has to be enforced separately. With the introduction of an ice concentration, the ice volume h has to be replaced by the actual ice thickness h ¼

h A

½5

i.e., the mean value of individual floe height, such that the total ice volume per grid box is not affected by this approach. An approach using two categories, open water and ice-covered areas with an internal structure (ice classes), has been proven highly adequate for a large number of situations.

Thermodynamics Mass conservation for ice is closely tied to the heat balance at its surfaces. Sea ice forms, if the freezing

temperature of sea water is reached. The surface freezing point Tf (in K) is a function of salinity, estimated by the polynomial approximation 3=2

Tf ¼ 273:15 0:0575Sw þ 1:710523 103 Sw 2:154996 104 S2w

½6

where Sw is the sea surface salinity. The majority of today’s sea ice thermodynamics models is based on a one-dimensional (vertical) heat diffusion equation, which for sea ice (without snow cover) reads ri cpi

@Ti @ @Ti ¼ þ Ki Ioi exp½Ki z ki @t @z @z

½7

Here, Ti, ri, cpi and ki are the sea ice temperature, density, specific heat, and thermal conductivity, respectively. The net short-wave radiation at the sea ice surface is Ioi and Ki is the bulk extinction coefficient. If a snow cover is present, the penetrating short wave radiation is neglected and a second prognostic equation for the snow is solved: rs cps

@Ts @ @Ts ¼ ks @t @z @z

½8

At the snow–sea ice interface, the temperatures and fluxes have to match. Assuming that ice exists, the local time rate of change of ice thickness due to freezing of sea water or melting of ice is the result of the energy fluxes at the surface and the base of the ice. At the surface, the ice temperature and thickness change is determined from the energy balance equation: Qaði;sÞ ¼ 1 aði;sÞ ð1 Ioi ÞRkSW þ RkLW eði;sÞ s0 T04ði;sÞ þ Ql þ Qs þ Qc ( ¼

0

if Toði;sÞ oTmði;sÞ

@h rði;sÞ Lði;sÞ @t

if Toði;sÞ ¼Tmði;sÞ

½9

where To(i,s) and Tm(i,s) are the surface and melting temperatures, respectively, r(i,s) and L(i,s) are the density and heat of fusion for sea ice and snow. Besides the conductive heat flux in the ice Qc ¼ kði;sÞ @Tði;sÞ =@t the following atmospheric fluxes are considered: downward short-wave radiation RkSW ðf; l; Acl , net long-wave radiation RkLW ðTa ; Acl Þ 4 , as well as sensible Qs ðva ; Ta ; Toði;sÞ Þ eði;sÞ so Toði;sÞ and latent Ql ðva ; qa ; qoði;sÞ Þ heat fluxes. The albedos aði;sÞ and emissivities Aði;sÞ are dependent on the surface structure of the medium (sea ice, snow). The atmospheric forcing data are:



• • • • •

-

near-surface wind velocity va ; near-surface atmospheric temperature Ta; near-surface atmospheric dew point temperature Td, or specific humidity qa; cloudiness Acl; precipitation P and evaporation E (needed for the sea ice and snow mass balance).

Note that the wind velocity is also needed for the atmospheric forcing in the momentum [2]. At the base of the ice (the sea surface), an imbalance of the conductive heat flux in the sea ice (Qc) and the turbulent heat flux from the ocean Qsw ¼ rw cpw ch u% Tf Tml

½10

leads to a change in ice thickness: Qiw ¼ Qow Qc ¼ ri Li

@hi @t

½11

Here, Tml is the ocean surface and mixed layer temperature, ch is the heat transfer coefficient and u% is the friction velocity. In general, the main sink for ice volume is basal melting due to above-freezing temperatures in the oceanic mixed layer. The source for snow is a positive rate of P–E, if the air temperature is below the freezing point of fresh water. The main source for sea ice is basal freezing. However, the formation of additional sea ice on the upper ice surface is possible through a process called flooding. This conversion of snow into sea ice takes place when the weight of the snow exceeds the buoyancy of the ice and sea water intrudes laterally. Often, the vertical structure of temperature is approximated by simple zero-, one- or two-layer formulations, with the resulting internal temperature profile being piecewise linear. The most simple approach, the zero-layer model, eliminates the capacity of the ice to store heat. However, it has been used successfully in areas where sea ice is mostly seasonal and thus relatively thin (o1 m). The specifics of the brine-related processes in the sea ice are difficult to implement in models. As a consequence, sea ice models usually assume constant sea ice salinity Si of about 5 PSU (practical salinity units) to calculate the heat of fusion and the vertical heat transfer coefficient. The errors arising from this assumption are largest during the early freezing processes, when salt concentrations are considerably higher. The open water part of each grid cell, where the atmosphere is in direct contact with the ocean, is treated like any other air–sea interface. The thermodynamic eqns [9] and [11] are modified to the radiative and heat fluxes between ocean and atmosphere. In the case of heat loss resulting in an ocean

691

temperature below the freezing point Tf , new ice is formed: 4 Qaw ¼ ð1 aÞð1 Iow ÞRkSW þ RkLW ew s0 T0w

þ Ql þ Qs þ Qsw ( ¼

0

@h ri Li @t

if Tow 4 Tf

½12

if Tow ¼ Tf

In the case of above-freezing ocean surface temperatures, Qsw follows from [10] with Tf replaced by Tow. An illustration summarizing the fluxes is given in Figure 2. The solution of eqns [7], [8], [9] and [11] is conceptually straightforward but algebraically complicated in that it involves iterative solution of the energy balance equation to obtain the surface temperature.

Dynamics Driven by wind and surface ocean currents, sea ice grown locally is advected horizontally. Free drift (the absence of internal ice stresses) is a good approximation for individual ice floes. In a compact ice cover, however, internal stresses will resist further compression and react to shearing stresses. These internal ice forces are expressed as the divergence of the isotropic two-dimensional internal stress tensor -

Fi ¼ r s

½13

which depends on the stress–strain relationship, where the deformation rates are proportional to the spatial derivatives of ice velocities. A general form of the constitutive law is 1 0 @ui @ui @ui @ui @ui @ui þ þ þz 1 @x @y Z @y @x C Pi B Z @x @y B C 2 s¼B C ½14 @ @ui @ui @ui @ui @ui @ui 1 A @y @x @y @x @x þ @y Z Z þz Pi 2 here, z and Z are nonlinear viscosities for compression and shear. Pi is the ice strength. The most widely used sea ice rheology is based on the viscous–plastic approach. Introduced as the result of the ice dynamics experiment AIDJEX, it has proven to be a universally applicable rheology. It treats the ice as a linear viscous fluid for small deformation rates and as a rigid plastic medium for larger deformation rates. Simpler rheologies have been tested but could not reproduce the observed ice distributions and thicknesses nearly as well as the viscous–plastic approach. The viscosities are then


z ¼ e2 Z ¼

Pi 2D

½15

692


Multiyear ice

New ice

Thin first-year ice

Thick first-year ice

Sea water

> s RSW

R net LW P–E

>

QS

ai

Ts =Ti Qcs =Qci _

aw

R net LW w RSW Q L QS

Water

IOW

P–E

QSW

Qc QSW

Ice

RSW

Qc

Ioi Ice

P–E

>

QC

R net LW i RSW Q L QS

>

RSW

QL

as Snow

>

>

RSW

iw

Qc QSW

iw

Figure 2 Schematic of the concept of ice classes/categories, surface energy balance and the ice–ocean flux coupling. The upper panel shows a grid cell covered with several classes/categories of ice, including open water. The lower panel is a detailed view at the fluxes between atmosphere, ice, and ocean, in the three cases: snow-covered sea ice, pure sea ice, and open water.

where an elliptic yield curve of ellipticity e is assumed, with the deformation rate given by "( ) @ui 2 @ui 2 @ui @ui 2 2 2 þ e D¼ þ 1þe þ @x @y @y @x 1=2 @ui @ui 2 1e þ2 ½16 @x @y

In particular, internal forces are only important for densely packed ice floe fields, i.e., for ice concentrations exceeding 0.8. Most sea ice models take this into account by assuming that the ice strength is Pi ¼ P hexp½C ð1 AÞ

½17

where P* and C* are empirical parameters. The same functional dependence is also used successfully to



describe the generation of open water areas through shear deformation, which is parameterized by

-

½18 Sdyn A ¼ 0:5 D r u i exp½ C ð1 AÞ Subgridscale processes like ridging and rafting can be successfully parameterized this way.

Coupling Numerical ocean circulation models are described in article General Circulation Models. A schematic illustration of the interactions in a coupled sea ice–ocean model is given in Figure 3. The coupling between the sea ice and ocean components is done via fluxes of heat, salt, and momentum. They enter the ocean model through the surface boundary

Radiation, heat and moisture fluxes

693

conditions to the vertically diffusive/viscous terms. Given the relative (to the depth of the ocean) small draught of the ice, it is assumed not to deform the sea surface, all boundary conditions are applied at the air–sea interface. Due to the presence of subgridscale ice categories, the fluxes have to be weighed with the areal coverage of open water, and ice of different thickness. The resulting boundary conditions for the simplest twocategory (ice and open water) case are: -

M @ uw

¼ Atiw þ ð1 AÞtaw ½19 Au @z z¼0

@Tw

¼ ATu @z z¼0

1 ðAQiw þ ð1 AÞQaw Þ rw cp

½20

Surface wind stresses

ATMOSPHERE

SEA ICE * Volume Mass conservation Growth decay

Drift Deformation

Thermodynamics Surface energy balances Vertical diffusion equation

Dynamics Momentum conservation

Ice classes/categories Subgridscale structure (including open water)

Mixed layer heat and salt, water masses

OCEAN

Circulation, sea surface height

Figure 3 Concept of a coupled sea ice–ocean numerical model with prescribed atmospheric forcing. Thick broken arrows represent the atmospheric forcing, thin dashed arrows represent the coupling pathways between ice and ocean, and thin arrows indicate how the ice volume is computed from thermodynamical and dynamical principles (mass and momentum conservation), with the assumption of a subgridscale ice distribution. * Includes both sea ice and snow.


694


Asu

@Sw

r @hi ¼ ð Sw Si Þ i @z z¼0 rw @t

8 PE > > > >
> > > : rs @hs rw @t

if h ¼ 0 if h > 0; Ta 40 if h40; Ta o0

½21

if hs 40; Ta 40

where the freshwater flux is converted into a salt flux, and the momentum exchange is parameterized (like the atmospheric wind forcing taw and tai ) in the form of the usual quadratic drag law:

- -

tiw ¼ rw Cw ui uw

h i ½22 ðui uw Þcosyiw þ k ðui uw Þsinyiw Here, Cw is the drag coefficient and yiw the rotation angle. The vertical viscosities and diffusivities T S (AM n ; An ; An ) are ocean model parameters. The sea surface height required to compute the ice momentum balance is either taken directly from the ocean model (for free sea surface models) or computed diagnostically from the upper ocean velocities using the geostrophic relationship. The described coupling approach can be used between any sea ice and ocean model, irrespective of vertical resolution, or the use of a special surface mixed layer model. However, the results may suffer from a grid spacing that does not resolve the boundary layer sufficiently well. A technical complication results from the different timescales in ocean and ice dynamics. With the implicit solution of the ice momentum equations, time steps of several hours are possible for the evolution of the ice. General ocean circulation models, on the other hand, require much smaller time steps, and an asynchronous time-stepping scheme may be most efficient. In that case, the fluxes to the ocean model remain constant over the long ice model step, whereas the velocities of the ocean model enter the fluxes only in a time-averaged form, mainly to avoid aliasing of inertial waves.

Numerical Aspects Equations [1]–[3] are integrated as an initial boundary problem, usually on the same finite difference grid as the ocean model. A curvilinear coordinate system may be used to conform the ice model grid to an irregular coastline or to locally increase the resolution. The horizontal grid is usually staggered, either of the ‘B’ or ‘C’ type. Both have advantages and disadvantages: the ‘B’ grid has been favored because of the more convenient formulation of the stress terms and the better representation of

the Coriolis term; the ‘C’ grid avoids averaging for the advection and pressure gradient terms. The treatment of coastal boundaries and the representation of flow through passages is also different. Due to the large nonlinear viscosities in the viscous plastic approach, an explicit integration of the momentum equations would require time steps of the order of seconds, whereas the thickness equations can be integrated with time steps of the order of hours. Therefore, the momentum equations are usually solved implicitly. This leads to a nonlinear elliptic problem, which is solved iteratively. An explicit alternative has been developed for elastic–viscous– plastic rheology. Other general requirements for numerical fluid dynamics models also apply: a positive definite and monotonic advection scheme is desired to avoid negative ice volume and concentration with the numerical implementation and algorithms depending on the computer architecture (serial, vector, parallel). Finally, the implementation needs to observe the singularities of the system of ice equations, which occur when h, A and D approach zero. Minimum values have to be specified to avoid vanishing ice volumes, concentrations, and deformation rates.

Model Evaluation Modeling systems need to be validated against either analytical solutions or observational data. The various simplifications and parameterizations, as well as the specifics of the numerical implementation of both components’ thermodynamics and dynamics and their interplay make this quite an extensive task. Analytical solutions of the fully coupled sea ice– ocean system are not known, and so model validation and optimization has to rely on geophysical observations. Since in situ measurements in high latitude icecovered regions are sparse, remote sensing products are being used increasingly to improve the spatial and temporal coverage of the observational database. Data sets of sea ice concentration, thickness and drift, ocean sea surface temperature, salinity, and height, are currently available, as well as hydrography and transport estimates. Ice Variables

The most widely used validation variable for sea ice models is the ice concentration, i.e., the percentage of ice-covered area per unit area, which can be obtained from satellite observations. From these observations, maps of monthly mean sea ice extent are constructed, and compared to model output (see Figure 4). It



695

GM

90˚W

90˚E

180˚ Figure 4 September 1987 simulated (blue) and remotely observed (red) sea ice edge around Antarctica. After Timmermann et al., 2001.

3.5 Observations Model

3.0

Ice thickness [m]

should be noted though that the modeled ice concentrations represent a subgridscale parameterization with an empirically determined source/sink relationship such that an optimization of a model with respect to this quantity may be misleading. A more rigorous model evaluation focuses on sea ice thickness or drift, which is a conserved quantity and more representative of model performance. Unfortunately, ‘ground truth’ values of these variables are available in few locations and over relatively short periods only (e.g., upward looking sonar, ice buoys) and comparison of point measurements with coarse resolution model output is always problematic. A successful example is shown in Figure 5. The routine derivation of ice thickness estimates from satellite observations will be a major step in the validation attempts. The evaluation of ice motion is done through comparison between satellite-tracked and modeled ice buoys. Thickness and motion of first-year ice in free drift is usually represented well in the models, as long as atmospheric fields resolving synoptic weather systems are used to drive the system.

2.5 2.0 1.5 1.0 0.5 0

Jan 1991 Jul 1991 Jan 1992 Jul 1992 Janl 1993

Figure 5 Time series of simulated (blue) and ULS (upward looking sonar) measured (red) sea ice thickness at 151W, 701S in the Weddell Sea. After Timmermann et al., 2001.

Ocean Variables

The success of the coupled system also depends on the representation of oceanic quantities. The model’s temperature and salinity distribution as well as the corresponding circulation need to be consistent with


696


prior knowledge from observations. The representation of water masses (characteristics, volumes, formation locations) can be validated against the existing hydrographic observational database, but a rigorous quantification of water mass formation products has been done only in few cases. Parameter Sensitivities

Systematic evaluations of coupled model results have shown that a few parameter and conceptual choices are most crucial for model performance. For the sea ice component, these are the empirical source/sink terms for ice concentration, as well as the rheology. The most important oceanic processes are the vertical mixing (parameterizations), especially in the case of convection (see Open Ocean Convection). in general, the formulation of the heat transfer between the oceanic mixed layer and the ice is central to the coupled system. Finally, the performance of a coupled sea ice– ocean model will depend on the quality of the atmospheric forcing data; products from the weather centers (European Centre for Medium Range Weather Forecasts, National Centers for Environmental Prediction/National Center for Atmospheric Research) provide consistent, but still rather coarsely resolved atmospheric fields, which have their lowest overall quality in high latitudes, especially in areas of highly irregular terrain and for the P–E and cloudiness fields. Some errors, even systematic ones, are presently unavoidable. All these data products are available with different temporal resolution. Unlike for stand-alone ocean models, which can be successfully run with climatological monthly mean forcing data, winds, sampled daily or 6-hourly, have been found necessary to produce the observed amount of ridging and lead formation in sea ice models.

needs to be captured by the model. For operational forecast purposes, the ice thickness distribution in itself is most important; here atmospheric data quality and assimilation methods become crucial. Obvious next steps may be the inclusion of tides, icebergs, and ice shelves. Ultimately, however, a fully coupled atmosphere– ice–ocean model is required for the simulation of phenomena that depend on feedback between the three climate system components.

See also Arctic Ocean Circulation. Bottom Water Formation. Current Systems in the Southern Ocean. Forward Problem in Numerical Models. General Circulation Models. Ice–ocean interaction. Open Ocean Convection. Polynyas. Satellite Remote Sensing SAR. Sea Ice: Overview. Under-Ice Boundary Layer. Upper Ocean Heat and Freshwater Budgets. Upper Ocean Mixing Processes. Weddell Sea Circulation.

Glossary ch cp(i,s,w) e f g h h(i,s) h* * h(i,s) hcls, hopn k(i,s)

Conclusions A large amount of empirical information is needed for coupled sea ice–ocean models and the often strong sensitivity to variations of these makes the optimization of such modeling systems a difficult task. Yet, several examples of successful simulation of fully coupled ice–ocean interaction exist, which qualitatively and quantitatively compare well with the available observations. Coupled sea ice–ocean modeling is an evolving field, and much needs to be done to improve parameterizations of vertical (and lateral) fluxes at the ice–ocean interfaces. For climate studies, water mass variability on seasonal and interannual timescales

! k qa qo(i,s) t ! vi ¼ ðui ; vi Þ u% ! va ¼ ðua ; va Þ ! v w ¼ ðuw1 ; vw Þ x, y, z A Acl T S AM n ; An ; An


heat transfer coefficient specific heat of sea ice/snow/water (J kg 1 K 1) ellipticity Coriolis parameter (s 1) gravitational acceleration (m s 2) ice (sea ice plus snow) volume per unit area (m) sea ice/snow volume per unit area (m) actual ice thickness (m) actual sea ice/snow thickness (m) lead closing/opening parameter (m) thermal conductivity of sea ice/ snow (W m 1 K 1) vertical unit vector atmospheric specific humidity surface specific humidity of sea ice/ snow time (s) horizontal ice velocity (m s 1) friction velocity (m s 1) wind velocity (m s 1) ocean surface velocity (m s 1) spatial directions (m) ice concentration cloudiness oceanic vertical mixing coefficients (m2 s 1)


C* Cw Io(i,w) E Ki L(i,s) P Pi P* Qa(i,s,w)

Qc Ql, Qs Qiw Qsw RkSW ; RkLW S(i,w) Ta, Ti, Tw Td Tf Tml Tm(i,s) To(i,s,w) FH ðthdyn;dynÞ SA Sthdyn hði;sÞ aði;swÞ D eði;swÞ Z; z @ @ ; @yÞ r ¼ ð@x l; f rða;i;s;wÞ s so yiw ! tiw ; ! taw tai ; !

empirical parameter oceanic drag coefficient short wave radiation penetrating sea ice/water (W m 2) evaporation (m s 1) bulk extinction coefficient (m 1) heat of fusion (J kg 1) precipitation (m s 1) ice strength (N m 1) ice strength parameter (N m 2) net energy flux between atmosphere and sea ice/snow/ water (W m 2) conductive heat flux in the ice (W m 2) atmospheric latent/sensible heat flux (W m 2) turbulent heat flux at the ocean surface (W m 2) oceanic sensible heat flux (W m 2) downward short/long wave radiation (W m 2) sea ice/sea water salinity (PSU) air/ice/water temperature (K) dew point temperature (K) freezing temperature of sea water (K) oceanic mixed layer temperature (K) melting temperature of sea ice/ snow at the surface (K) sea ice/snow/water surface temperature (K) internal ice forces (N m 2) sea surface elevation (m) source/sink terms for ice concentration (s 1) source/sink terms for sea ice/snow volume per unit area (m s 1) sea ice/snow/sea water albedo ice deformation rate (s 1) sea ice/snow/sea water emissivity nonlinear viscosities (kg s 1) horizontal gradient operator geographical longitude/latitude (deg) densities of air/sea ice/snow/water (kg m 3) two-dimensional stress tensor (N m 1) Stefan-Boltzmann constant (W m 2 K 4) turning angle (deg) air-ice/ice-water/air-water stress (N m 2)

697

Appendix A typical parameter set for simulations with coupled dynamic–thermodynamic ice-–ocean models (e.g., Timmermann et al., 2001), as shown in Figures 4 and 5, are

ra ¼ 1.3 kg m 3 ri ¼ 910 kg m 3 rs ¼ 290 kg m 3 rw ¼ 1027 kg m 3 e¼2 C* ¼ 20 P* ¼ 2000 N m 2 hcls ¼ 1 m hopn ¼ 2 h* Cw ¼ 3 10 3 ch ¼ 1.2 10 3 aw ¼ 0.1 ai ¼ 0.75 ai ¼ 0.65 (melting) as ¼ 0.8 aw ¼ 0.7 (melting) Ki ¼ 0.04 m 1 Si ¼ 5 PSU y ¼ 10 degrees cpi ¼ 2000 J K 1 kg 1 cpw ¼ 4000 J K 1 kg 1 cpa ¼ 1004 J K 1 kg 1 Li ¼ 3.34 105 J kg 1 Ls ¼ 1.06 105 J kg 1 ki ¼ 2.1656 W m 1 K 1 ks ¼ 0.31 W m 1 K 1

Further Reading Curry JA and Webster PJ (1999) Thermodynamics of Atmospheres and Oceans. London: Academic Press. International Geophysics Series.. Fichefet T, Goosse H, and Morales Maqueda M (1998) On the large-scale modeling of sea ice and sea ice–ocean interaction. In: Chassignet EP and Verron J (eds.) Ocean Modeling and Parameterization, pp. 399--422. Dordrecht: Kluwer Academic. Haidvogel DB and Beckmann A (1999) Numerical Ocean Circulation Modeling. London: Imperial College Press. Hibler WD III (1979) A dynamic-thermodynamic sea ice model. Journal of Physical Oceanography 9: 815--846. Kantha LH and Clayson CA (2000) Numerical Models of Oceans and Oceanic Processes. San Diego: Academic Press. Leppa¨ranta M (1998) The dynamics of sea ice. In: Leppa¨ranta M (ed.) Physics of Ice-Covered Seas, vol. 1, 305--342.


698


Maykut GA and Untersteiner N (1971) Some results from a time-dependent thermodynamic model of sea ice. Journal of Geophysical Research 76: 1550--1575. Mellor GL and Ha¨kkinen S (1994) A review of coupled ice–ocean models. In: Johannessen OM, Muench RD and Overland JE (eds) The Polar Oceans and Their Role in Shaping the Global Environment. AGU Geophysical Monograph, 85, 21–31.

Parkinson CL and Washington WM (1979) A large-scale numerical model of sea ice. Journal of Geophysical Research 84: 311--337. Timmermann R, Beckmann A and Hellmer HH (2001) Simulation of ice–ocean dynamics in the Weddell Sea. Part I: Model description and validation. Journal of Geophysical Research (in press).


CRUSTACEAN FISHERIES J. W. Penn, N. Caputi, and R. Melville-Smith, Fisheries WA Research Division, North Beach, WA, Australia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 570–578, & 2001, Elsevier Ltd.

Introduction The Crustacea are one of the most diverse groups of aquatic animals, occupying a wide variety of habitats from the shore to the deep ocean and the tropics to Arctic waters, and extending into fresh water and in some cases on to land for part of their life history. Crustacean species contribute in the order of 7 million tonnes annually, or about 6–8% of the total world supply of fish, according to FAO statistics. Approximately 75% of this volume is from harvesting wild stocks, with the remainder from aquaculture, dominated by the tropical marine and freshwater shrimps, crayfish, and crab species. Owing to their high market value as a sought-after high-protein food, crustaceans make up a disproportionate share of the value of the world’s seafoods. As a result of their high value, crustacean fisheries are generally heavily exploited and require active management to be sustained. Research to underpin management of these resources has been undertaken in many parts of the world, and particularly Australia where, unusually, lobsters and shrimps are the dominant fisheries. Crustacean fisheries are focused on the more abundant species, particularly those in relatively shallow, accessible areas. Shrimps are the most important wild fishery products, followed by the crabs, lobsters, and krill.

Biology and Life History Fisheries research on crustacean stocks is significantly influenced by their unusual life history and biology. A unique feature of crustaceans is that they all must undergo a regular process of molting (casting off their outer shell or exoskeleton) to grow. Once the old shell is cast off, the animal absorbs water to swell or ‘grow’ to a larger size before the shell hardens. Volume increase at a molt varies between species, but typically results in a gain in the range of 10–60%. The molt also serves as an opportunity to regenerate damaged limbs, such as legs

or antennae, although regeneration results in lower or even negative growth increments. This molting process occurs throughout all stages of the life history and is often correlated with environmental factors such as temperature, moon phase, or tidal cycles. Because of this mechanism, growth occurs as a series of discrete ‘steps’ rather than a ‘smooth’ increase over time and is complicated to measure. Growth rates are highly dependent on water temperature, with tropical and shallow-water crustaceans tending generally to grow much faster than those in cooler and deeper waters. As a result of the molting process, it is not possible to ‘age’ crustaceans using any of the usual methods (growth rings on bones or shells) applied to other fished species. The molting process also significantly influences feeding activity and hence catch rates for all crustaceans. Prior to a molt, feeding activity is generally reduced, then ceases in the lead-up to the actual molt process. Following the molt the animals are particularly hungry, begin active feeding, and are more easily caught in baited traps or trawls, but contain relatively little meat for their shell size and are of lower market value. These cyclic catches in crustacean fisheries are well known to fishers and are also crucial knowledge for fisheries stock assessment and industry management. Significant short-term gains in value of the catch can be achieved where the fishery management arrangements take them into account. The second very important feature of crustaceans is that exploited marine species are generally highly fecund, producing large numbers of eggs (millions in some species) which hatch into pelagic larvae which in turn can be widely distributed by ocean currents. Typically, the early larval stages are of a different form to the adults, but after a series of larval stages the individual molts into a form resembling the adults. Larval stages generally have limited swimming ability but are able to migrate up and down within the water column, and often have behaviors which, in combination with tides and currents, result in active dispersal into ‘nursery’ areas suitable for the later juvenile and adult stages. The large numbers of eggs and larvae produced, together with widespread dispersal mechanisms common in the major marine crustacean species, make them relatively resilient to fishing pressure compared with the freshwater species. The typical marine larval life history does not generally apply to the freshwater species, where


699

700

CRUSTACEAN FISHERIES

some or all larval development stages occur within a much larger egg and live young are often produced to minimize downstream losses due to river flow. These alternative larval strategies adopted by freshwater crustacea are efficient but, owing to the relatively low numbers produced, make such species more susceptible to overfishing than their marine counterparts.

Marine Shrimps and Prawns This group contains by far the most important crustacean fisheries. In marine waters two major families, the Penaeidae in tropical waters and the Caridea in cold waters, support most of the significant export fisheries. In tropical fresh waters, paleomonid shrimps of the genus Macrobrachium support the major commercial production. The terms ‘shrimp’ and ‘prawn’ have no scientific basis and are used interchangeably in different parts of the world. For the purposes of this chapter, the more commonly applied term ‘shrimp’ will be used for simplicity. A wide array of penaeid species are harvested from tropical to subtropical waters. These species have a complex life cycle where mated females spawn generally in coastal marine waters where eggs (hundreds of thousands per spawning) are broadcast freely and hatch as free-swimming planktonic nauplius larvae. After a series of larval molts, postlarvae actively move into estuaries and coastal embayments where they develop into the juvenile stage. Juveniles and subadults then actively migrate offshore using tidal flows to further develop, mate, and spawn at 6– 12 months of age. Coupled with these relatively short life cycles is rapid growth, but also high levels of natural mortality such that few individuals survive to more than 12 months of age, although some species may live to 2 or 3 years without fishing. High natural mortality does, however, allow for high but sustainable exploitation rates for most of this group of commercially important species, although there are some exceptions noted later. Fishing for these species is generally by means of fixed nets in estuary mouths during the offshore migration of subadults, or by vessels otter trawling in waters offshore from the estuarine nursery areas. Penaeid shrimps are generally not catchable in traps. Powered otter trawling for shrimp, which evolved in the Gulf of Mexico, has now been adopted worldwide as the main method for industrial-scale catching of the more valuable export market-sized adult shrimps. Otter trawling can only occur on

smooth bottoms, usually sand or mud adjacent to nursery areas, and typically harvests approximately 20–50% of the shrimps in the path of the net. The remainder are generally buried in the sediments, particularly during the day. The exception to this is where some shrimp species (e.g. Penaeus merguiensis) form dense schools and generate turbid mud ‘boils’ as a defense against predators. This behavior, including mid-water swimming, allows very high exploitation rates and catches on some occasions, but has been noted to break down at high levels of exploitation and in areas where river/estuarine habitats and adjacent waters have become increasingly turbid. Major fisheries for penaeid shrimps occur through the Gulf of Mexico and Central/South American coasts (P. aztecus, P. setiferus, P. duorarum, P. braziliensis, P. californiensis and P. vannamei), off the Chinese river deltas (P. orientalis), through southeast Asia (various Metapenaeus and Penaeus species), Indonesia–Papua New Guinea (P. merguiensis), Australia (P. latisulcatus, P. esculentus/semi-sulcatus, P. merguiensis), and the African coasts (P. indicus, P. notialis). In addition to the large or more valuable penaeids, very large quantities of very small Acetes and sergestid shrimps are harvested, particularly in Asian coastal waters, by small-scale coastal fisheries. The second commercially important group of shrimps comprises the caridean species, which occur predominantly in the Northern Hemisphere, in temperate to Arctic waters. Where they extend into more tropical waters, they do so only at greater depths with cold temperatures corresponding to Arctic waters. This group of shrimps is relatively long-lived (up to 4–6 years), spawning at several years of age. These species are also typically protandric hermaphrodites, growing into functional males before undergoing a series of molts to become female for the remainder of their life. Females produce larger but fewer eggs than the penaeid species, and carry them after spawning attached under their tail. The eggs remain attached for an extended period, undergoing some developmental stages within the egg before hatching into pelagic larvae which grow for several months before settling onto a wide range of habitat types. Pandalus borealis is a typical caridean shrimp for which the life cycle has been well studied and represents the general life history pattern for this important group. Fishing occurs by both otter trawling and trapping with baited traps which are particularly effective for these species, unlike penaeids which do not trap easily. Major fisheries for these pandalid species occur in the northern Atlantic and north Pacific.



In the Antarctic zone, the major equivalent crustacean fishery is for euphausiids or krill (see Krill). These krill species are particularly abundant in the nutrient-rich Southern Ocean, where it is estimated that a biomass of 20 or more million tonnes occurs. Krill are small pelagic species which swim by way of modified walking legs (swimmerets), and generally undertake a diurnal migration between the surface and significant depths. They form dense schools on the surface, particularly at night, where they are a major source of food for Antarctic whales, seals, and fish stocks. Estimates of potential sustainable yield range into millions of tonnes per year, but the catch has been limited by processing difficulties to about 100 000 tonnes.

Crabs Most commercially significant crabs belong to the Brachyara (true crabs) or Anomura (hermit crabs and king crabs) within the order Decapoda. They are generally characterized by a pair of claws, three pairs of walking legs, and a wide, flattened body. Crabs are probably the most highly developed, successful, and diverse of the crustaceans. They occupy a wide range of environments, from shallow tropical seas to deep ocean trenches, estuarine and fresh waters, and some species spend the majority of their life on land, only returning to the water to reproduce. Reproductive patterns in crabs are diverse and often involve intricate courtship behaviors where the male protects the female before mating. Following copulation, crabs retain spermatozoa until egg laying, at which time fertilization takes place as the eggs are extruded. Spermatozoa can be retained by the female in a viable condition for considerable periods of time – more than a year in some species. The important swimming crab species are generally resilient to heavy fishing pressure due to their

Table 1

701

often complex but efficient reproductive behavior and high levels of fecundity. Some species carry multiple broods of eggs, which are extruded, fertilized, and attached to the underside of the female during the early development stages. Numbers of eggs produced per year are frequently in the order of 50 000–500 000 per female, and over a million eggs are achieved by some species. Crab larval stages are known as zoea and most marine species have four or five zoeal stages before molting into a megalopa, which generally settles out of its planktonic existence. In a number of cold-water crab fisheries (e.g. the important snow, tanner, king, and Dungeness fisheries) where breeding is more restricted, managers have elected to allow harvesting of males only, thereby giving complete protection to the brood stock. This precautionary approach, whilst useful for these species, imposes unusual constraints on research due to the inability to monitor female crabs in the commercial catch. Most crab fishing worldwide is by use of traps. This method is preferred to most others because traps are simple to use (particularly in deep water) and labor-efficient, and the crabs are less likely to be injured. This latter fact is particularly important because it allows the product to be sold live, guaranteeing a better market price than frozen forms. Many other methods are used to catch crabs, including trawling, tangle netting, dredges, trotlines, and drop nets. Interestingly, the majority of the large crab fisheries operate in tropical and Northern Hemisphere temperate and Arctic waters. Table 1 shows that the three most important commercial crab species are all fast-growing ‘swimming crabs’, found in shallow tropical or temperate waters and embayments. These are a family of crabs which have a flattened, paddle-like hindmost leg used to burrow in sand and mud, or to propel them

World landings in order of quantity reported by FAO catch statistics for 1996

Common name

Species name

1996 catch (tonnes)

Gazami crab Blue crab Blue swimmer crab Snow and tanner crab King crab Dungeness crab Edible crab Red crab

Portunus trituberculatus Callinectes sapidus Portunus pelagicus Chionoecetes spp. Paralithodes spp. Cancer magister Cancer pagurus Geryon/Chaceon spp.

303 000 116 000 112 000 100 000 81 000 34 000 29 000 7000

The table excludes landings of mud crab (Scylla spp.), as the majority of that is produced by aquaculture.


702


through the water during infrequent occasions when they ‘swim’ over short distances. They reach maturity and are harvested between 1 and 3 years of age. The largest crab catches landed worldwide are those of gazami crab; however, a very substantial, but unspecified portion of these reported landings are from aquaculture operations. China alone produced 80 000 tonnes of gazami crab by aquaculture in 1997. The species has a wide distribution through the western Pacific and lives in shallow inshore waters in sheltered embayments. Stocking of waters with juvenile gazami crab has become widespread, particularly off the Japanese coast, and is considered to be economically effective. Blue crabs occur in the western and central western Atlantic. The vast majority of the landings are made off the US coastline from states in the Gulf of Mexico and mid-Atlantic. The commercial fishery targets both hard crabs and peeler/soft crabs, softshelled crabs being considered a delicacy in the USA. Soft crabs have very recently molted and have a shell that has yet to become hard. While some of the peeler/soft crab product is taken with crab scrapes and other specialized methods capable of taking nonfeeding animals, the majority of the product is produced in operations which hold peelers in shedding tanks until molting occurs. The snow, tanner, and king crabs, which are high on the list of important species in Table 1, are examples of moderately deep-water species occurring in cold water conditions. The distributional range of these species encompasses water o400 m deep (and, particularly for king and snow crabs, usually o200 m and colder than 101C). These species are very slow growing when compared with the inshore warmer water species mentioned earlier. Their age at maturity is generally upward of 5 years and in most cases they enter into the commercial fishery over 8 years after settlement. Over the long history of crab production in the north Pacific, large catches of king, tanner, and snow crabs have been made. Despite stock collapses of some species, this area is still important for its crab production and for the research efforts that have been made to understand the biology and management of these important stocks. Crabs belonging to the Geryon and Chaceon genus (Table 1) are commercially important deep-water crabs. They have a wide depth range, but most of the commercially exploited populations tend to be in the 500–1000 m depth range. Water temperatures at these depths are typically less than 101C and these animals are therefore slow growing. In Namibia, where the largest and one of the longest-standing

Chaceon fisheries exists, the crabs take approximately 8 years to reach maturity.

Lobsters and Crayfish Lobster and crayfish species support significant and high-value fisheries. The major commercial lobster species are marine and taken from tropical to cold temperate waters, while freshwater crayfish are mostly taken from tropical and subtropical regions. Most of the marine species have similar life history patterns, where females carry fertilized eggs externally under their abdomens. Following hatching, the larvae undergo a series of molts before taking up a benthic habitat and growing to adulthood, a process which can take many years. Freshwater crayfish species generally have a reduced larval life and hatch as small juveniles. Fisheries are dominated by three groups, the Homarus species (large-clawed lobsters), the Nephrops species (small-clawed lobsters), and the palinurid group (spiny or rock lobsters, without claws). Catches of each of these groups are in the order of 60 000–80 000 tonnes annually. Fishing is generally by baited traps (Homarus and most palinurid species), although some are taken by trawl (Nephrops), and diving (tropical Panulirus species). Freshwater crayfish are also taken by baited traps. The major clawed lobster fishery is for Homarus americanus off eastern Canada and the USA. A similar-sized fishery for Nephrops norvegicus occurs off the European Atlantic coasts and through the Mediterranean. Spiny lobster fisheries for the Panulirus species occur through the tropics, with major fisheries in the Caribbean (P. argus) and Western Australia (P. cygnus). Smaller but significant fisheries for Jasus species occur in the temperate waters off southern Australia, New Zealand, and South Africa. The major freshwater crayfish fishery occurs in the southern states of the USA. Stocks of these species have generally been resilient to fishing, with the exception of some Jasus species off Africa which have been significantly reduced over time.

Fishery Assessment Research There are two fundamental biological issues to be addressed in the management of fish stocks (including crustaceans). The most important problem is to control the level of fishing such that there is sufficient breeding stock to continue to provide adequate supply of new recruits to the fishery. This is



generally tackled using the relationship between breeding stock and recruitment. The second issue is to maximize the overall catch (and value). This problem is traditionally examined using a yield-perrecruit model which examines the trade-off between the increase in biomass through growth over time and the decrease in survival through natural and fishing mortality. The other biological studies undertaken, such as growth, migration, reproduction, and mortality, are generally the building blocks to enable the assessment of these two key issues. There are three main differences in the biological assessment of crustacean fisheries compared with many finfish fisheries; they are growth, migration, and catchability. Growth by molting is probably the key difference between crustaceans and other marine species. Thus stock assessment needs to take into account the timing of the growth and the size increment at the molt. The frequency and size increment of molting usually decreases with age, especially after reaching maturity. Crustacean growth contrasts with that of finfish populations, which can generally be modeled using a continuous growth model. Because crustaceans totally replace their outer shell with each molt they cannot be aged in this way, creating a major problem for stock assessment. This process also makes tag recapture less reliable for these species. As a consequence, age is usually estimated by following length frequencies of particular year-classes, although this is often possible for only the younger year-classes. Some recent work on measuring the age pigment, lipofuscin, which generally increases linearly with age and is not lost at the molt, may provide an opportunity in the future to regularly utilize age information in the stock assessment of crustaceans. The second feature of crustaceans which sets them apart from finfish and affects their stock assessment is migration. While generally poor in swimming ability, crustacean species often undergo significant migrations linked to specific stages in their life cycle. For example, tropical shrimps and swimming crabs have specific behaviors which enable them to actively migrate offshore, utilizing ebb tidal flows, as they approach sexual maturity. Many spiny lobster species also undergo extensive directional migrations at a particular age, usually following a coordinated molt, marching in columns from shallow nursery areas to offshore spawning areas before reaching sexual maturity. These migration ‘events’ are often of short duration and usually unidirectional. Such rapid, shortterm interruptions to the normal, relatively sedentary

703

behavior of crustaceans pose special constraints on stock assessment. That is, crustacean migration typically causes erratic changes in stock distribution and catches, contrasting with most finfish fisheries where the regular, more consistent swimming movements of the fish result in continuous redistribution of the stock. Because molting is often synchronized and related to growth and migration, the catchability of many crustaceans is also typically inconsistent and often cyclic. For this reason, crustacean catch rates do not directly reflect the abundance of the stock and must be corrected for in the data sets utilized in stock assessments. To assess the status of exploited crustacean stocks which present these particular problems, long data series are extremely valuable, especially where they can be used to refine the catch rate–abundance relationship and in establishing the linkage between different life history stages. Such data can be used to assess the relationship between spawning stock, environmental factors, and recruitment to the fishery for managing the critical effects of fishing on the spawning stocks. For example, the long time-series of data (Figure 1) was fundamental in assessing the cause of the collapse of tiger prawn stocks in Shark Bay, Western Australia. These data were used to evaluate the impact of fishing effort on the spawning stock and assess the reduction in fishing effort required for the fishery to recover to its optimal level. Figure 2, derived from the historical data, shows the relationship between spawning stock levels and subsequent recruitment to the Shark Bay tiger prawn stock. This relationship, together with the reverse relationship between recruitment and surviving spawner abundance (in the same year) relative to variations in fishing effort targeting the stock, has been used to construct a simple model (Figure 3) to determine optimal levels of tiger prawn fishing effort. Management changes to redirect effort away from the species based on this modeling have resulted in a recovery of the tiger prawn stock (Figure 1). Similarly, the use of catch predictions in the western rock lobster (Panulirus cygnus) fishery up to 4 years ahead using an index of abundance of settling puerulus (first post-larval stage) and juveniles entering the fishery has enabled fisheries management to be proactive rather than reactive to changes in stock abundance. This relationship, presented in Figure 4, shows that catch is determined by variations in puerulus settlement 3 and 4 years previously, and fishing effort during the year of recruitment to the fishery. Such predictive relationships have been used in Western Australia to adjust fishing levels in


704


Shark Bay annual prawn catch and effort 2500

80 King 70

Tiger Effort

2000

50

1500

40 1000

30

Effort (hours x 1000)

Landings (tonnes)

60

20 500 10 0

0 62 64 66 68 70 72 74 76 78 80 82 84 86 88 Year

90 92 94 96 98

Figure 1 The time-series data on catch and fishing effort (hours trawled) for tiger prawns (Penaeus esculentus) and western king prawns (P. latisulcatus) since the inception of the Shark Bay (Western Australia) prawn fishery in 1962.

900 67 800

76 95

700

70

79 96

Recruitment index

97

83

98

68

69

91 72

400

92

84

200

71

77

94 500

66

74

73

600

300

75

78

86 87 88

89 81 90 80

93

85 82

100

0 0

5

10

15 Spawning stock index

20

25

99(2.9)

Figure 2 The relationship between spawner abundance and recruitment (1 year later) for the tiger prawn (P. esculentus) stock in Shark Bay (Western Australia). Year of recruitment is shown against each data point.

advance to ensure that breeding stock levels are maintained. This development of predictive relationships using long-run data sets also enables environmental factors

which may influence survival of larval stages, and catchability in crustacean stocks, to be examined. Figure 5, showing the relationship between rock lobster puerulus settlement and Fremantle sea level



80

900

60

705

40

800 X Ricker SRR 700 Y

Recruitment index

600 0

500

400

300 200

100

0 0

5

10 Spawning index

15

20

25

Figure 3 A model combining the spawner–recruit relationship SRR and recruitment to spawner (as affected by fishing effort) for the Shark Bay tiger prawn stock, which has been utilized to estimate optimal fishing effort levels. Trajectory ‘X’ shows the expected annual decline in recruitment and spawning stock at an unsustainable level of fishing effort (80 000 trawling hours). Trajectory ‘Y’ shows the converse stock recovery from low levels when fishing occurs at optimal levels of about 40 000 hours of trawling effort.

as an index of flow of the Leeuwin Current along the WA coastline, is an example of this type of analysis. The availability of this type of relationship is particularly valuable to researchers attempting to distinguish between the effects of fishing and shortterm ‘natural’ variations in recruitment to the fishery caused by environmental influences. This is an important distinction, as a poor year-class due to environmental factors, coupled with high fishing effort, can combine to produce a very low breeding stock and trigger a long-term stock decline. The second most important fisheries problem, optimizing yield per recruit to the fishery, is also particularly difficult to assess in crustaceans owing to the molting process. The resulting inability to age or reliably tag these species makes estimation of natural mortality and growth of pre-recruit year-classes relatively unreliable. This has led to an ‘adaptive’ management approach using adjustments to sizes at first capture over a number of years to directly assess the resulting impact on catch. The alternative approach has been to develop complex simulation models based on length rather than age. These model-based

assessments have improved significantly the ability to manage crustacean fisheries, but again where successful have relied heavily on long-run, detailed fishery databases for their testing and validation.

Crustacean Management Techniques Management techniques applied to the significant tropical shrimp fisheries focus mainly on minimum trawl mesh sizes, accompanied by area and seasonal closures to optimize the quantity and size of shrimps caught. Many of these trawl fisheries also involve specific gear regulations to minimize unwanted bycatch. Owing to the highly variable annual recruitment to these fisheries, the most common and successful management approaches have involved fishing effort controls and, more recently, transferable effort quotas. For the longer-lived, more consistent cold-water pandalid shrimp and krill fisheries, catch quotas are more applicable and often utilized. Because of their larger individual size (and value) and the dominant method of capture (trapping) which facilitates live discarding of unwanted catch, the


706


5.0

4.5 4.5

82(4.0) 87(4.7)

91(4.5)

4.0

89(4.5)

4.0

98(3.7)

83(4.0) 90(4.3)

Millions of kg

88(4.4)

84(4.0)

96(3.9)

3.5

78(3.6) 3.5

95(3.8) 85(4.2)

77(3.5) 97(3.7)

81(3.6)

80(3.6)

93(3.6)

79(3.6)

76(3.6)

3.0

92(3.7)

94(3.8)

86(4.2) 75(3.4) 74(3.4)

2.5 72(3.3)

73(3.4) 2.0

02

01

00

1.5 0

30

60

90

120

150

180

Puerulus (Dongara) _ 3, 4 years before Figure 4 Catch forecast for western rock lobster in the northern part of its range, based on the level of puerulus settlement at Dongara 3–4 years earlier and the number of pot lifts. The catch year is shown with millions of pot lifts in brackets.

common management focus in crab fisheries is on legal minimum size regulations and protection for spawning females. Gear design rules specifying ‘escape gaps’ to reduce the capture of undersize crabs have become a common management tool for these fisheries. Overall management of the longer-lived temperate or deep-water crabs often involves catch quotas to ensure maintenance of breeding stocks and economic performance of fisheries. This methodology is less relevant to the faster-growing, more variable tropical crab fisheries, which lack the predictable recruitment and longevity necessary for effective catch quota management. In these fisheries, effort controls through limited entry are more useful and common. The management techniques for high-value lobster stocks are generally similar to those for crabs, focusing on legal minimum sizes, associated gear controls, and female protection in the trap fisheries

which dominate this crustacean sector. Historically, limited entry arrangements have been the most common overall management strategy for sustaining lobster fisheries, with ‘individually transferable trap quotas’ first applied in the 1960s to the Australian spiny lobster fisheries. Total catch quotas, applied more recently through ‘individually transferable quotas’, have also been utilized to control fishing, particularly for the more consistent, longer-lived cold-water lobster fisheries.

Conclusions The high value of crustaceans has led to increasing exploitation pressures and the need for improved research assessments to underpin management. Stock assessment methods for crustacean fisheries, however,



SOI

0

75

Sea level

Puerulus

See also

80

−10 200

70

150

65

Sea level

SOI

been critical to the more recent improvements in crustacean fisheries assessment for management.

ENSO

10

707

Demersal Species Fisheries. Dynamics of Exploited Marine Fish Populations. Fish Feeding and Foraging. Fish Locomotion. Fish Predation and Mortality. Population Dynamics Models.

100 Puerulus

Further Reading

50

2000

1996

1992

1988

1984

1980

1976

1972

1968

0

Year Figure 5 Annual mean values of Southern Oscillation Index (SOI), Fremantle sea level, and puerulus settlement at Dongara. ENSO (El Ninõ/Southern Oscillation) periods are indicated with arrows.

provide significant scientific challenges owing to the unique crustacean method of growth through molting. This mode of growth prevents the use of the longestablished age-based methods applied to the more generic finfish and some molluscan fisheries. The stock assessment approach adopted has therefore focused on direct measurement of recruitment and spawning stocks and the use of long-run fishery databases. The importance of using long-run data sets to validate and test length-based models has

Caddy JF (1989) Marine Invertebrate Fisheries: Their Assessment and Management. New York: John Wiley & Sons. Caputi N, Penn JW, Joll LM, and Chubb CF (1998) Stockrecruitment-environment relationships for invertebrate species of Western Australia. In: Jamieson GS and Campbell A (eds.) Proceedings of the North Pacific Symposium on Invertebrate Stock Assessment and Management. Canadian Special Publication on Fisheries and Aquatic Science 125: 247–255. Cobb JS and Phillips BF (1980) The Biology and Management of Lobsters, vol. II: Ecology and Management. New York: Academic Press. Debelius H (1999) Crustacea Guide of the World. Frankfurt: IKAN-Unterwasserarchiv. Gulland JA and Rothschild BJ (eds.) (1984) Penaeid Shrimps: Their Biology and Management. Farnham, Surrey: Fishing News Books. Phillips BF and Kittaka J (eds.) (2000) Spiny Lobsters: Fisheries and Culture. Oxford: Fishing News Books. Provenzano AJ (1985) The Biology of Crustacea, vol. 10: Economic Aspects: Fisheries and Culture. Orlando, FL: Academic Press.


CTD (CONDUCTIVITY, TEMPERATURE, DEPTH) PROFILER A. J. Williams, III, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: The In Situ Measurement of Salinity, Temperature, and Density in the Ocean One of the most useful instruments developed for determining seawater properties during the last four decades has been the CTD (conductivity, temperature, depth). This device has supplanted the traditional hydrocast using Nansen bottles and reversing thermometers that was standard physical oceanographic practice from about 1910 to 1970. The CTD, although an electronic instrument, has its origin in the older technology. The computations of properties such as depth, salinity, density, speed of sound, and potential temperature have been greatly facilitated by having the measurements of conductivity, temperature, and pressure in digital format for direct entry into standard formulas, originally in FORTRAN but now in Matlab and other computational engines.

surface along the wire releases a latch, which causes the thermometer to invert, breaking the column of mercury in the capillary tube, thus capturing the volume of mercury expanded into the capillary tube (Figure 1). When the reversing thermometer is returned to the surface, this length of mercury in the capillary tube can be measured, the change in length due to the change in temperature from the sample depth to the surface corrected for, and the in situ temperature of the seawater at the sampled depth computed. The technique is sensitive and reliable with well-characterized reversing thermometers

Temperature The temperature of seawater is directly important because many physical properties depend upon temperature and indirectly important because calculations of salinity from conductivity measurements are dominated by the temperature dependence of conductivity. Temperature is a conservative property of seawater. It is generally modified only when the fluid is at the surface where it can exchange heat with the atmosphere or rarely when it is in contact with another body of water where exchanges of heat may occur by mixing. Temperature has been sampled in hydrocasts (hydrographic stations) with reversing thermometers since the late nineteenth century. In a reversing thermometer, mercury in a glass bulb expands or contracts filling a capillary tube to a greater or less extent much as occurs in a normal fever thermometer. However, when the reversing thermometer has equilibrated with the seawater at its depth along the hydrowire, a messenger falling from the

708

Figure 1 The reversing thermometer has a constriction in the capillary tubing that causes the mercury column to break when the thermometer is returned to its upright orientation after being deployed upside down. This allows the temperature to be measured at depth with the thermometer in the normal, connected column, but then when the thermometer is inverted by the agency of a messenger sent down the hydrographic wire, the mercury beyond the constriction is captured and can be read upon recovery to the surface. An auxiliary thermometer allows the surface reading temperature to be applied to correct for the expansion of the trapped mercury column. General Oceanics, Inc.


CTD (CONDUCTIVITY, TEMPERATURE, DEPTH) PROFILER

where a long history of their calibration has been kept. Resolution of temperature in deep expandedrange thermometers is typically 1 or 2 millidegree. However, hydrostatic pressure at depth would compress the glass and cause the mercury to move along the capillary a greater distance than if the pressure were kept at 1 atm, so in situ temperature is measured with a pressure-protected reversing thermometer inside a pressure-resisting glass tube. The depth of the measurement can be determined even if the hydrowire is not a straight vertical line the length of which can be measured. Current shear in the water column bends the hydrowire into a curve and causes the sample depth to be less than that determined from the meter wheel reading: the distance that was paid out from that when the reversing thermometer entered the water until the lowering was stopped. So for more than a century it has been possible to determine the temperature profile from the surface to the bottom but only at discrete intervals of depth. A heavily instrumented hydrographic cast may have had a dozen reversing thermometers on it but these only gave the temperature at a dozen depths. In some stations, several casts are made with instruments only in, say, the top 1000 m for higher resolution on one and deeper instruments with more widely spaced depths on another (Figure 2).

Figure 2 The photograph shows a hydrographic cast being taken with a scientist on the hero platform attaching a bottle to the hydrographic wire. Woods Hole Oceanographic Institution.

709

A continuous profile of temperature became possible first with the bathythermograph (BT) and subsequently by the STD (salinity, temperature, depth) profiler. The BT, developed by Athelstan Spillhaus in the 1930s, was a valuable tool during World War II for US destroyers and submarines in determining the acoustic refraction in the upper water column with consequences for acoustic submarine detection. Sharp thermal interfaces were revealed at places where sound was refracted by the difference in speed of sound of the water on the two sides of the interface. In the BT, a glass slide plated with carbon (smoked) or a thin gold layer was scratched by a stylus that was moved along one arc by expansion of a fluid-filled coiled tube responding to temperature and along another arc by a second coiled tube (Bourdon tube) empty of fluid and acted upon by pressure. The fine scraped line on the slide could be read in a viewer that was calibrated for the BT with which it was used. In practice, temperatures at regular depths were read off and transcribed to a paper chart, thus missing the promise of a continuous profile of temperature. In fact, there were frequently apparent jiggles in the traces that were initially ignored and only much later were discovered to be real indications of physical phenomena and termed fine structure. Electronic instrumentation, accepted grudgingly at first, made the measurement of temperature with a platinum resistance thermometer a practical option at sea. This sensor in which a temperature-sensitive element of fine platinum wire is placed inside a pressure-protecting metal sleeve was stable and could be calibrated to a few millidegrees or even to sub-millidegree precision. The near absence of work hardening of the platinum wire and its freedom from corrosion made the platinum resistance thermometer a standard in temperature measurement in the lab and in the sea. It did, however, require some assumptions to define the specific resistance at temperatures other than the critical points of boiling water (100 1C) at standard pressure (1013 millibar) and normal isotopic composition, and the melting point of ice (0 1C) under the same conditions. The ice melting point is sensitive to trace contamination with electrolytes and to the state of the water in thermal equilibrium with the ice, so a more reliable critical point has been used to pin down the low-temperature end of the platinum resistance thermometer scale, the triple point of water. When standard isotopic composition water is in equilibrium with all three phases, liquid (water), vapor (steam), and solid (ice), and there are no other fluids or vapors present, the temperature is 0.010 1C, a very well characterized temperature. Sealed triple point cells have been


710


Thermometer Chilled liquid

Thermometer Water vapor

Thermometer well Metal bushing

Ice and water Ice mantle Inner melt

Thermometer well Ice mantle

Pure water Pure water Ice and water Container

Triple point of water cell Figure 3 The triple point cell is made with a film of ice in a sealed finger of glass containing only pure degassed water. At the point where the ice surrounding the internal cavity starts to melt, the temperature is 0.010 1C. Sea-Bird Electronics, Inc.

constructed to establish ice/water/vapor equilibrium for precise temperature calibration. A triple point cell has an internal cavity into which a sensor to be tested can be placed (Figure 3). Before its use, an ice layer is grown on the outside of the cavity in contact with the water by dropping dry ice into the cavity and freezing the water to a thickness somewhat less than the clearance of the cavity from the outer wall of the triple point cell. Then a little warming is permitted until the ice covering can be spun around the cavity without sticking. This is the point where the cavity wall is at the triple point, 0.010 1C. Two points may define a linear relation between resistance and temperature but in reality, this relationship is not exactly linear. The temperature scale is in practice defined thermodynamically as that which an ideal gas would obey: PV ¼ nRT, where P is absolute pressure, V is specific volume, n is the number of moles of gas, R is the Boltzman constant, and T is the absolute temperature. Even though there is no ideal gas that obeys this relation due to van der Waals forces between molecules of gas at low temperatures, helium approximates it better than any other gas. A practical temperature scale based upon resistance of a platinum resistance thermometer was defined in 1978 but careful work led to an improved temperature scale being established in 1983 that retained the critical endpoints but adjusted the nonlinear shape between the endpoints with a difference of several millidegrees (about 5 millidegree) near 35 1C. This is a serious error to introduce into long-term measurements of, for

example, global warming, so the exact temperature scale used for calibration and interpretation is critical. Thermistors are semiconductors where the resistance changes much more radically with a change in temperature than the platinum resistance thermometer but they are not as stable. Thermistors are heterogeneous ceramic beads exhibiting a negative or a positive temperature coefficient of conductivity; the latter are generally used in this application as they become more conductive as they get warmer. The beads, however, frequently contain defects that grow or heal and change the thermistor resistance at a fixed temperature, so ultrastability is not a characteristic of thermistors unlike platinum resistance thermometers. Thermistors are useful as a secondary sensor of temperature but are calibrated against a platinum resistance thermometer which in turn is calibrated against a triple point cell. The XBT or expendable bathythermograph uses a thermistor to profile temperature from a sinking body that is temporarily connected to the launcher by a fine electrical wire (Figure 4). There is one other enhancement of the calibration of temperature in the oceanographic temperature range. A gallium triple point cell can be used to establish a critical point, 29.764 6 1C, closer to the oceanic temperature range than the boiling point of water. In any case, the calibration of thermometers is critical for long-duration monitoring of oceanic water masses to discern warming trends but is less important for gradient measurements.



711

Deck-mounted launcher model LM 2A

Hand-held launcher model LM 3A Launcher cable − supplied with launcher Ship’s hull

MK12 mating cable − connects to J1 on MK12 board

Probe

Thru-hull launcher model LM4A

PC with MK12 interface board

Connector box − supplied wired to mating cable

Figure 4 XBT or expendable bathythermograph is a continuous recording temperature profiler that can be dropped from a moving ship or an aircraft. A thermistor in the nose of the probe is sensed through a fine wire that is streamed by both the falling probe and the moving vessel. Depth is assumed from time elapsed after probe contact with the water and is calibrated for fall rate. Lockheed Martin Sippican, Inc.

As electronic temperature measurements were installed upon profilers, in particular the STD profiler (Bissett-Berman, c. 1968), observations of fine structure in temperature came to be accepted as real, which led to an understanding of internal thermal mixing and stratification. Microstructure, as the layering at scales of meters first became known, turned out to result from double-diffusive convection in many cases and produced layers on the order of tens of meters thick, which became known as fine structure and which are now known to be nearly ubiquitous in the ocean (Figure 5).

Salinity The equation of state of seawater, which relates the density to temperature, salinity, and pressure, has

been determined with great care by laboratory methods. Of almost as great interest as the density (and more for water watchers) is the salinity. Salinity is, if anything, more conservative than temperature and only changes upon exposure of the water to the atmosphere or another body of water at a different salinity where mixing may occur. Hydrothermal vents on the seafloor have been suspected of also changing the salinity of deep water masses by injection of dissolved materials. The present theory even suggests that hydrothermal vent circulation establishes the overall salinity of the sea, not river runoff followed by evaporation. So determination of salinity is important, not just for the equation of state where salinity affects density. The direct determination of salinity is awkward. It is defined as the weight of solids in grams in 1 kg of


712


38.50

Salinity (%°)

38.60

600

Depth (m)

S T 700

800 13.4 13.6 Temperature (°C)

13.8

Figure 5 The profile of sheets and layers from an instrument named self-contained imaging microprofiler (SCIMP) shows at very high resolution the microstructure in the Tyrrhenian Sea. The sensor in SCIMP was the first CTD to use internal recording. Freedom from a cable allowed very smooth sinking and high spatial resolution of temperature and salinity. Molcard R and Williams AJ, III (1975) Deep stepped structure in the Tyrrhenian Sea. Memoires Societe Royale des Sciences de Liege VII: 191–210.

seawater when the water has been evaporated and all the carbonates have been converted to oxides, bromine and iodine converted to chlorine, and all organic matter completely oxidized. Direct evaporation does not work because chlorides are lost. But a simpler indirect measure can be based on the almost constant composition of seawater (same ratios of major ions nearly everywhere, only the water content varies). This involves titrating the chloride (and other halogens) with silver nitrate and indicating with potassium chromate. The relation is salinity ¼ 0:03 þ 1:805 chlorinity Titrating is slow and awkward, so an attempt was made to determine the salinity by electrical conductivity measurements. The comparison was made between conductivity of diluted standard seawater and full-strength standard seawater at a common temperature. The relation was fairly linear even though seawater is more than a very dilute solution. Once the relations were worked out from laboratory measurements, it became possible to measure salinity by putting the unknown sample in a temperature bath and measuring the ratio of its conductivity to that of a known sample in the same temperature bath. Schleicher and Bradshaw at Woods Hole Oceanographic Institution did this work in the 1960s.

The resulting instrument was the lab salinometer and it allowed faster, more precise determinations of salinity than the older titration method. But it did require a supply of standard seawater for comparison. Standard seawater is a commodity required by the case for oceanographic cruises where many salinity samples are expected. On a typical hydrographic cast, each reversing thermometer was attached to a salinity sample bottle. Fridtjof Nansen perfected metal sample bottles used on a hydrowire at the turn of the twentieth century and these, which were triggered with the same messenger that caused the reversing thermometers to reverse, closed valves at the top and bottom capturing a water sample that could later be titrated or compared to standard seawater in a salinometer (Figure 6). Reversing thermometers were commonly attached to Nansen bottles so that a single triggering action might capture a temperature and a salt sample at one instant. Standard seawater is collected at an open ocean site of which several were used initially. Currently the IAPSO standard seawater is of North Atlantic origin. After being filtered through a 200-nm filter, this water is evaporated and diluted to a specific conductivity to serve as a standard, the principle being that for many thousands of samples the ionic composition is essentially the same. By bringing its conductivity to a standard value with the addition or subtraction of pure water, all samples are assumed to be interchangeable. An earlier standard, Copenhagen standard seawater, was unfortunately not representative of all seawater of the world being somewhat anomalous in having a different ionic composition, so that the constant 0.03 had to be added to the chlorinity determined by titration. However, as a conductivity standard rather than as a chlorinity standard, Copenhagen water served well. The temperature bath-controlled salinometer was a boon to laboratory analyses of salinity measurements from bottles but it was not sufficient for an ocean-going salinity-profiling instrument. Two more steps were required. The first was to determine the temperature coefficient of conductivity and this permitted correction of the conductivity measurement without a thermostatic bath. The principal variable responsible for conductivity changes in seawater is temperature, not salinity, so the temperature had to be measured very accurately and the lab work done very carefully. Actually, the conductivity ratio could still be used as long as the standard seawater was at the same temperature as the unknown sample. But, it was also possible to just calibrate the salinometer occasionally with standard seawater and to calculate the difference in conductivity expected from the temperature of the sample, which was different from the standard seawater calibration temperature.



713

Figure 6 A case of standard seawater flasks is shown as carried on research cruises where many samples will be taken and analyzed at sea. The standard seawater is used to calibrate the salinometer.

Salinometers permitted salinities to be run at sea from Nansen bottles, which improved accuracy somewhat because saltwater samples can sometimes get spoilt if kept too long. But the observations were made from only a few points in the profile. Then Neil Brown in 1961 combined pressure measurements with conductivity and temperature measurements to make an in situ sampler, the STD profiler. Schleicher and Bradshaw determined the pressure effect on conductivity (by now a three-variable problem) and Brown and Allentoft extended the conductivity ratio measurements. The STD opened a new window on the ocean and immediately presented problems for physical oceanographers by showing fine structure in a way that could no longer be ignored. The STD converted the conductivity measurement to salinity with in situ analog circuitry using temperature and pressure. However, only a few years later, computers began to go to sea and Brown realized a better algorithm could be applied to raw digital conductivity, temperature, and pressure measurements by shipboard-based computers than by using the analog conversions in the in situ instrumentation. Furthermore, if recorded digitally, the original data could always be reprocessed at a later time upon the improvement of the algorithm. Finally, the precision and accuracy of the measurement could be improved and the size of the sensors reduced to push the scale of observations from the meter to the centimeter scale. It was the latter that drew Brown to Woods Hole Oceanographic Institution in 1969 to develop the microprofiler. This was the first CTD.

CTD (Conductivity, Temperature, Depth) Sensors Temperature can be measured to about 2 millidegree with reversing thermometers and salinity can be relied upon to a few parts per million. To improve on this, Brown aimed for resolution of salinity to 1 ppm which required resolution of temperature to 0.5 millidegree Celsius. Stability had to be very good to make calibrations to this standard meaningful. For standards work, the platinum thermometer is used and Brown chose that for the CTD. To minimize size and retain high stability with the conductivity measurement, Brown designed a ceramic, platinum, and glass conductivity cell. For pressure, he used a strain gauge bridge on a hollow cylinder. Using temperature, conductivity, and pressure, salinity can be calculated and from temperature, salinity, and pressure, density can be calculated. Depth can then be determined from pressure, the integral of density to the surface, and a local value for gravity. The correction from pressure to depth is small and for many purposes inconsequential so that profiles of temperature and salinity against pressure are used in place of profiles against calculated depth in most cases.

Pressure Originally Brown planned to build each sensor himself but technology in the commercial world provided him with an adequate pressure sensor initially in a pressure transducer produced by Paine


714


Signal + Excitation

Signal −

Excitation −

Figure 7 This drawing of a strain gauge pressure sensor shows four sets of wires around the body of the sensor. Internal pressure from the port at the left stretches the center of the cylinder more than the heavy ends but the thermal expansion is similar in all four coils. Driven as a resistance bridge, the sensing leads experience a doubling of the effect of a single coil but cancel the thermal effect on stretching of the wires.

Instruments, Inc., subsequently improved with temperature correction for reducing hysteresis during a profile. Typical lowering speeds for deep CTD profiles are between 30 and 100 m min 1, limited by the need to prevent the cable from going slack and jumping the sheave at the top or getting a loop in the cable near the bottom. But this speed causes the temperature in the instrument housing to vary rapidly, especially while transiting the thermocline, and temperature gradients inside the instrument are a problem for sensors that were designed for constant temperature. The hysteresis in pressure has remained a problem up to the present time and has only been tolerated because the errors are not serious for the computation of salinity. They are principally of concern when trying to measure motion of a surface of constant salinity or temperature between profiles or between the down profile (clean because the instrument is at the leading edge of the insertion) and the up profile a few minutes to an hour later (Figure 7).

Conductivity Direct measurement of conductivity presents problems because of polarization of seawater at the electrodes, so the salinity measurements of the STD made electronically used an inductive cell without electrodes. In this cell, made with dimensionally stable materials to fix the geometry, a toroidal transformer was constructed in which seawater formed a single shorted secondary turn through the hole. Electric current was induced in this shorted turn and the conductivity of the seawater in this

path measured as a transformed conductivity in the toroidal primary winding of the cell. While reasonably stable, this technique did not offer the high spatial resolution desired in the CTD nor was it the only route to removing the difficulties with electrodes, so Brown elected to design a miniature stable conductivity cell for the CTD. The conductivity cell of Brown’s CTD had four electrodes, two for current and two to measure voltage, to minimize electrode effects with a symmetry that made it insensitive to local contamination of the electrodes. It was only 3 mm in diameter and 8-mm long, so it was hoped that this small size would be able to resolve centimeter-scale structure. Measurements were made at 10 kHz to circumvent polarization at the current electrodes. Other geometries for electrode-type conductivity cells have been used, a three-electrode cell from Sea-Bird Electronics, Inc., for example (Figure 8). Conductivity cell design is a present occupation of sensor technologists. For example, long-duration, fast cells are being developed by Ray Schmitt of Woods Hole Oceanographic Institution for deployment on gliders (autonomous underwater vehicles that profile along oblique paths by gliding between two depths). Original plans by Brown to make his own thermometer, in a helium-filled ceramic capillary tube, were discarded when it was discovered how hard it was to work with ceramics. Endless difficulties in glass to ceramic and glass to metal seals developed and overcoming these in the conductivity cell which had no voids was hard enough. A commercial platinum resistance thermometer was chosen from Rosemont Inc., with a time constant of 300 ms and a guaranteed stability of 10 millidegree in a year but in practice somewhat better. The 300-ms response time of the thermometer meant that for 1-cm resolution, his original target, descent rates of 0.3 m min 1 would be the limit. This was a bitter result; however, Brown added a fast response thermistor to correct the temperature measurement at faster descent rates. The correction technique added the derivative of the thermistor temperature measurement in an analog circuit to the stable platinum resistance thermometer temperature measurement to replace the high-frequency variations that were lost, without affecting the overall accuracy of the platinum resistance temperature measurement in less dynamic regions. Later, he increased the size of the conductivity cell to facilitate manufacture. This potentially degraded spatial resolution. Flushing of the conductivity cell has been an issue and the original cell, although only 8-mm long, had a flushing length at speeds above 10 cm s 1 of about 3.5 cm. The new cell flushed in about 8 cm of descent during lowering



Platinum electrodes Borosilicate (3 places) glass cell

715

Current field between electrodes

Seawater flow in

Seawater flow out Cell terminals

Figure 8 This drawing of a conductivity cell shows three electrodes, reducing the external electric field from the current between the outer electrodes and the central electrode. Sea-Bird Electronics, Inc.

on a hydrographic station. With a thermistor response time of 30 ms, a 10-cm vertical resolution was possible at descent speeds of 50 cm s 1 or 30 m min 1, a reasonable winch speed. The requirement of resolving structure to 10 cm at a descent rate of 30 m min 1 meant a sample rate of 10 Hz. (The original resolution target was higher and the first microprofiler had three channels running at 32 ms each in parallel.) The subsequent Neil Brown Instrument Systems Mark III CTD successively digitized conductivity, pressure, and temperature at 32 ms in each channel so that it obtained a complete sample every 96 ms, which was fast enough to resolve 10-cm thick features at a lowering speed of 30 m min 1. Practical solutions to the requirements of very high accuracy and reasonable speed are hallmarks of the very best instruments and the Mark III CTD established a standard. The range in temperature is about 32 1C, from freezing to nearly the warmest surface water. For packing efficiency, straight binary integers were used and the value of 215 is 32 768. Thus a 16-bit measurement of temperature gives 0.5 millidegree resolution and 0–32.8 1C range. (For some work, a 2 1C lower end is needed and this was later incorporated.) Conductivity varies over the same range because it tracks temperature. Sixteen bits generally permit salinities up to 38% to be measured to 0.001% precision. The depth range needed for most of the ocean is 6500 m (or a pressure range of 6500 decibar) and, with a digitizer capable of 16-bit resolution, 10-cm depth resolution is permitted, again right on target for the resolution of the sensors. But to make a measurement to a part in 65 536 (216) and have it remain accurate and stable is not easy. Furthermore, the conductivity measurement must be made at about 10 kHz to minimize electrode polarization. Neil Brown’s solution to these problems was to make all of the digitizations with transformer windings, weighted in a binary sequence and added electronically. These were driven at 10 kHz so that polarization effects were minimal. Precision in the measurements was ensured by the turns ratio in the transformer. While the Neil Brown CTD was the first of the new profiling instruments, others soon followed.

Figure 9 A CTD with rosette sampler of large Nisken bottles and reversing thermometers is being lowered off the side of a ship. Woods Hole Oceanographic Institution.

Guildline, Inc., produced an excellent lab salinometer and that technology similarly permitted a profiler to be developed. Sea-Bird Electronics, Inc., produced sensors for temperature and conductivity based upon a Wein bridge oscillator that were precise, compact, and easy to incorporate into instruments and soon Sea-Bird produced its own CTD based upon these sensors. Sea-Bird’s conductivity cell is a three-electrode cell with a slow natural flushing time but it is generally flushed with an external pump to establish a constant flushing time irrespective of lowering rate, thus improving the spatial resolution. This device is now widely used on oceanographic


716


Figure 10 Neil Brown (1927–2005), a long-time contributor to precise continuous measurements of physical properties of the ocean, shown here confronting the CTD he invented, had a great sense of humor. His use of transformer-based analog-to-digital converters provided the precision that permitted electronic measurements of temperature, conductivity, and pressure to be made in profilers.

research vessels. Ocean Sensors, Inc., has produced a small CTD with internal recording for inclusion on autonomous instruments. Idronaut, S.r.l., has a CTD which can accommodate additional sensors for oxygen, carbon dioxide, ambient light, pH, and optical backscatter, for example. In fact, the ability to add sensors to a digital instrument has been an advantage not lost on instrument builders, so that a typical profiler package on a ship may contain a CTD with a suite of these additional sensors, some sensors duplicated for redundancy, and generally includes a rosette sampler of Niskin bottles to capture water on trigger from the surface.

Extended Deployments of the CTD (Conductivity, Temperature, Depth) Battery-supplied power for the CTD and the development in the early 1970s of digital magnetic tape recorders freed the CTD from cable connection to the ship (Figure 9). The salinity and temperature profiles shown in Figure 5 were recorded on a SeaData cassette tape from a free vehicle. Now with massive solid state memory such as compact flash, replacing hard disks of the 1980s and 1990s, digital data storage is not a problem and CTDs are found on autonomous underwater vehicles, moored profilers, and on gliders and floats. The CTD is now a standard oceanographic instrument and has replaced the Nansen cast as a hydrographic tool (Figure 10). The data are sent up conducting cable as a frequency-shift-keyed signal. Multiconductor or fiber-optic armored cable has replaced the hydrographic wire for most hydrographic

surveys. Recording as an acoustic signal on tape has been replaced by direct storage of digital data on a PC or dedicated server but the decoding and computation of salinity is still generally done on deck. The algorithm has been clumsy in that it sticks close to the actual relations derived from the laboratory data sets. These were derived by going from salinity to conductivity, not the other way around. Dynamically one wants to know density and this is now computed as a second step when it could be done directly. But computational complexity is a negligible cost with even the smallest PCs. Many CTDs export salinity directly, having done the conversion internally. However, conductivity recording still permits enhanced processing if raw values are retained. A direct measurement of density might be the next step in sensor development. However, the demand for sensors of light, chemistry, and properties other than the classical ones of temperature, salinity, and pressure has seemed to be more significant and the modern CTD is often just the central element of a complex suite of sensors.

See also Gliders. Neutral Surfaces and the Equation of State. Satellite Remote Sensing of Sea Surface Temperatures. Satellite Remote Sensing: Salinity Measurements.

Further Reading Bacon S, Culkin F, Higgs N, and Ridout P (2004) IAPSO standard seawater: Definition of the uncertainty in the calibration procedure, and stability of recent batches.



Journal of Atmospheric and Oceanic Technology 24: 1785--1799. Bradshaw A and Schleicher KE (1970) Direct measurement of thermal expansion of seawater under pressure. DeepSea Research 17: 691--706. Bradshaw A and Schleicher KE (1980) Electrical conductivity of seawater. IEEE Journal of Oceanic Engineering 5: 50--56. Brown NL and Allentoft B (1966) Salinity, Conductivity and Temperature. Relationships of Seawater over the Range of 0–50%. US ONR Contract Nr-4290(00), MJO 2003 Final Report, Washington, DC. Cox RA, Culkin F, and Riley JP (1967) The electrical conductivity/chlorinity relationship in natural seawater. Deep-Sea Research 14: 203--220. Dauphinee TM (1980) Introduction. Special Issue on the Practical Salinity Scale 1978. IEEE Journal of Oceanic Engineering 2: 1--2. Forch C, Knudsen M, and Sorensen SPL (1902) Berichte uber die Konstantenbestimmungen zur Aufstellung der hydrographischen Tabellen. Kgl. Danske Videnskab. Selskabs Skrifter, 7 Rackke, Naturvidensk. Og Mathem. Afd. 12, pp. 1–151. Lewis EL and Perkin RG (1978) Salinity: Its definition and calculation. Journal of Geophysical Research 83: 466--478. Miyake M, Emery WJ, and Lovett J (1981) An evaluation of expendable salinity–temperature profilers in the eastern North Pacific. Journal of Physical Oceanography 11: 1159--1165. Molcard R and Williams AJ, III (1975) Deep stepped structure in the Tyrrhenian Sea. Memoires Societe Royale des Sciences de Liege VII: 191--210. Ridout P and Higgs N (online) An Overview of the IAPSO Standard Seawater Service. Havant: OSIL. http:// www.ptb.de/de/org/3/31/313/230ptbsem/230ptbsem_ osil_ridout.pdf (accessed Mar. 2008).

717

UNESCO (1981) International Oceanographic Tables, Vol. 3. UNESCO Technical Papers in Marine Science 39, pp. 1–111. Paris: UNESCO. Williams AJ, III (1974) Free-sinking temperature and salinity profiler for ocean microstructure studies. IEEE International Conference on Engineering in the Ocean Environment 2: 279--283.

Relevant Websites http://www.paineelectronics.com – Downhole and Differential Pressure Transducer, Sensor, and Transmitter, Paine Electronics. http://www.sippican.com – Expendable Probes, Sippican, Inc. http://www.generaloceanics.com – General Oceanics, Inc. http://www.guildline.com – Guildline Instruments. http://www.idronaut.it – Idronaut, S.r.l. http://www.agu.org – News article ‘Athelstan Spilhaus dies at 86’ (30 March 1998), Science and Society (American Geophysical Union). http://www.mnc.net – Nordic Visitors Norway. http://www.oceansensors.com – Ocean Sensors, Inc. http://www.emersonprocess.com – Rosemount Temperature Products, Emerson Process Management. http://www.seabird.com – Sea-Bird Electronics, Inc. http://www.whoi.edu – Woods Hole Oceanographic Institution.


CURRENT SYSTEMS IN THE ATLANTIC OCEAN L. Stramma, University of Kiel, Kiel, Germany Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 589–598, & 2001, Elsevier Ltd.

Introduction By the late nineteenth century our present view of the Atlantic Ocean surface circulation had already been largely worked out. The voyages of discovery brought startling observations of many of the important surface currents. During the twentieth century the focus turned to a detailed description of the surface currents and the investigation of the subsurface currents. Recently, much attention has been focused on climate research as it became clear that climate goes through long period variability and can affect our lives and prosperity. The physical climate system is controlled by the interaction of atmosphere, ocean, land and sea ice, and land surfaces. To understand the influence of the ocean on climate, the physical processes and especially the ocean currents storing and transporting heat need to be thoroughly investigated. Since the end of the twentieth century, the general circulation of the Atlantic Ocean has been considered in a climatological context. A new picture emerged with the North Atlantic Ocean being seen not only as relevant to the climate of Europe, but for its influence on the entire globe due to its unique thermohaline circulation. Its warm upper-ocean currents transport mass and heat, originating in part from the Pacific and Indian Oceans, towards the north in the South-Atlantic and the North Atlantic. Cold deep waters of the North Atlantic flow southward, cross the South Atlantic, and are exported into the Indian and even the Pacific Oceans. The research focus on ocean currents in the Atlantic Ocean at the beginning of the new millennium will be further investigation of the mean surface and deep currents of the Atlantic and its variability on short-term to decadal timescales, so that the ocean’s role in climate change can be better understood.

Basin Structure The Atlantic Ocean extends both into the Arctic and Antarctic regions, giving it the largest meridional extent of all oceans. The north–south extent from Bering Strait to the Antarctic continent is more than

718

21 000 km, while the largest zonal distance from the Gulf of Mexico to the coast of north-west Africa is only about 8300 km. Here the Atlantic Ocean is described between the northern polar circle and the southern tip of South America at 551S (see Figure 1). In the north, the Davis Strait between northern Canada and west Greenland separates the Labrador Sea from Baffin Bay to the north of Davis Strait. The Denmark Strait between east Greenland and Iceland, and the ridges between Iceland and Scotland separate the Irminger Basin and the Iceland Basin from the Greenland Sea and Norwegian Sea. In the south a line from the southern tip of South America to the southern tip of South Africa separates the South Atlantic Ocean from the Southern Ocean. Although there is no topographic justification for this separation, defining a southern ocean around Antarctica allows the separate investigation of the Antarctic Circumpolar Current and the processes near Antarctica as a whole. The Atlantic Ocean has the largest number of adjacent seas, and the larger ones are discussed in other articles (see Baltic Sea Circulation; Current Systems in the Mediterranean Sea and North Sea Circulation.) The Mid-Atlantic Ridge, which in many parts rises to o2000 m depth and reaches the 3000 m depth contour nearly everywhere, is located zonally in most places near the middle of the Atlantic and divides the Atlantic Ocean into a series of eastern and western basins (Figure 1). The basin names presented in Figure 1 are only the major ones; in more detailed investigations of special regions many more topographically identified structures exist with their related names. The major topographic features of the Atlantic strongly affect the deep currents of the deep and bottom water masses, either by blocking or guiding the flow. The Walvis Ridge off south-west Africa limits the northward flow of Antarctic Bottom Water (AABW) in the Cape Basin and consequently the major northward flow of AABW takes place in the western basins. Although the Rio Grande Rise between the Argentina Abyssal Plain and the Brazil Basin disturbs a smooth northward spreading of the AABW in the western basins, the Vema and Hunter Channels within the Rio Grande Rise are deep and wide enough to allow a continuous northward flow. The Romanche Fracture Zone at the equator allows part of the AABW to enter the eastern basins of the Atlantic, where the AABW spreads poleward in both hemispheres.


CURRENT SYSTEMS IN THE ATLANTIC OCEAN

75° W

25°W

50° W

25°E

0° Norwegian

D en St ma ra r k it

100° W

Basin

La

60°N

60°N

br

ad

Ba

sin

719

or

Biscay

40°N

Rid ge

Abyssal North

40°N

At

la

nt ic

American

Plain

Mi d

Basin Cape

20°N

20°N

Verde Abyssal Plain

0°

0°

RFZ Angola

Brazil

Abyssal Basin

20°S

20°S

Grande Rise 40°S

Argentina

ge id R

Rio

Mid Atlantic Ridge W al vi s

Plain

Cape Basin

40°S

Abyssal Plain

Scotia Ridge 60°S 100° W

Drake Passage 75° W

50° W

25° W

0°

60°S 25°E

Figure 1 Topography of the Atlantic Ocean with large-scale depth contours in 1000 m steps. RFZ, Romanche Fracture Zone.

Once a water mass is formed, there is a conservation of its angular momentum, or rather potential vorticity. For large-scale motions, in the interior of the ocean, potential vorticity reduces to f/h ¼ constant (where f is the Coriolis parameter and h the water depth). From this expression we can predict which way a current will swing on passing over bottom

irregularities – equatorward over ridges and poleward over troughs in both hemispheres. A prominent example is the interaction between the relatively narrow Drake Passage south of the South American continent and the Scotia Ridge, which connects Antarctica with South America and contains numerous islands, located about 2000 km east of Drake


720


Passage. The Antarctic Circumpolar Current accelerates to pass through the Drake Passage, meets the Scotia Ridge with increased speed, and shifts sharply northward. In subpolar and polar regions density variations with depth are small and the pressure gradient force is more evenly distributed over the water column than in the tropical and subtropical regions. As a result, the currents in the subpolar and polar regions extend to great depth. In the case of the subpolar gyre of the North Atlantic, the currents have a strong barotropic component (with little vertical velocity shear) and hence tend to follow f/h contours.

Historical Developments Charts of ocean currents from the late nineteenth century show that by then the patterns of surface circulation in regions away from the equator and polar latitudes were already well understood. This fundamental knowledge accumulated gradually through centuries of sea travel and had reached a state of near correctness by the time dedicated research cruises, full depth measurements, and the practical application of the dynamical method were begun. By the fifth century AD, mariners had probably acquired intimate knowledge of coastal currents in the Mediterranean, but little information about them is reported in Classical writing. Following the dark and Middle Ages when little progress was made, the voyages of discovery brought startling observations of many of the Atlantic’s most important ocean currents, such as the North and South Equatorial Currents, the Gulf Stream, the Agulhas Current, and others. The Gulf Stream appears to have been mapped as early as 1525 (by Ribeiro) on the basis of Spanish pilot charts. The fifteenth to seventeenth centuries were marked by attainments of knowledge that increasingly taxed the abilities of science writers to reconcile new information with accepted doctrine. Significant advances in determining the global ocean circulation beyond local mapping of currents came only after the routine determination of longitude at sea was instituted. The introduction of the marine chronometer in the late eighteenth century made this possible. Largely because of the marine chronometer, a wealth of unprecedentedly accurate information about zonal, as well as meridional, surface currents began to accumulate in various hydrographic offices. In the early nineteenth century data from the Atlantic were collected and reduced in a systematic fashion (by James Rennell), to produce the first detailed description of the major circulation

patterns at the surface for the entire mid- and lowlatitude Atlantic, along with evidence for crossequatorial flow. This work provided a foundation for the assemblage of a global data set (by Humboldt and Berghaus) that yielded worldwide charts of the nonpolar currents by the late 1830s. Heuristic and often incorrect theories of what causes the circulation in the atmosphere and oceans were popularized in the 1850s and 1860s and led to a precipitous decline in the quality of charts intended for the public (Maury; Gareis and Becker). However, errors in popular theories provided motivation for the adoption of analytical methods, which in turn led directly to the discovery of the full effect of Earth’s rotation on relatively large-scale motion and the realization of how that effect produces flow perpendicular to horizontal pressure gradients (Ferrel). The precedents for modern dedicated research cruises came in the 1870s (e.g. the Challenger cruise), as well as mounting evidence for the existence of a deep and global thermohaline circulation (Carpenter, Prestwich). With the ever-increasing numbers of observations made at and near the surface, the upperlayer circulation in nonpolar latitudes was approximately described by the late 1880s. A current map by Kru¨mmel (1887) nicely described the surface currents of the entire Atlantic. This figure is not reproducible; however, Figure 2 shows as an example an earlier and slightly less accurate map from Kru¨mmel (1882), but only for the South Atlantic Ocean.

Currents of the Atlantic Ocean Warmwatersphere The warmwatersphere, consisting of the warmer upper waters of the ocean, is the most climatologically important part of the ocean due to its direct interaction with the atmosphere. The transition from the warm- to the cold-watersphere takes place in a relatively thin layer at temperatures between 81 and 101C. The warm watersphere reaches to 500–1000 m depth in the Atlantic’s subtropics and rapidly rises towards the ocean surface poleward of about 401 latitude. The near-surface circulation is driven primarily by the wind and forced into closed circulation cells by the continental boundaries. The circulation of the Atlantic is governed by the subtropical gyres of the North and South Atlantic (Figure 3). The subtropical gyre of the North Atlantic includes the Florida Current and Gulf Stream as western boundary currents, the North Atlantic Current, the Azores and Canary Currents in the eastern Atlantic, the North Equatorial Current, and the Caribbean, Cayman and



80

60

40

20

0

20

721

40

20

20

40

40

60

60 80

60

40

20

.. W.O 0.v. Gr.

20

40

Figure 2 Chart of surface currents in the South Atlantic Ocean by Kru¨mmel (1882).

Loop Current in the Caribbean and Gulf of Mexico. The subtropical gyre of the South Atlantic includes the poleward-directed Brazil Current as western boundary current, which turns eastward at the Brazil/Falkland (Malvinas) confluence region as eastward flow near 401S named South Atlantic Current. The South Atlantic Current in part continues to the Indian Ocean and in part adds to water from the Agulhas retroflection to the northward-flowing Benguela Current and the westward-flowing South Equatorial Current. Near the coast of north Brazil the South Equatorial Current contributes in part to the Brazil Current, but in part also to the subsurface intensified North Brazil Undercurrent, responsible for the warm water flow from the Southern to the Northern Hemisphere. The two anticyclonic subtropical gyres, clockwise in the Northern and counterclockwise in the Southern Hemisphere, reach through the entire warmwatersphere and show only weak seasonal changes. Subtropical gyres, although existing longitudinally to basin-scale, also tend to have sub-basin-scale recirculation gyres in their western reaches (Figure 3). The northward extent of the South Atlantic subtropical gyre decreases with increasing depth. It is located near Brazil at 161S in the near-surface layer and at 261S in the layer of Antarctic Intermediate Water. The preferential north–south orientation of the continents bounding the Atlantic Ocean lead to meridional eastern and western boundary currents which together with the wind-induced zonal currents, westward flow under the trade winds, and

eastward flow under the midlatitude westerly winds, form the closed gyres. The western ocean boundary regions are associated with an intensification of the currents. The consequent energetic western boundary currents, the Florida Current and the Gulf Stream in the North Atlantic Ocean and the Brazil Current in the South Atlantic Ocean, have large transports and typical width scales of B100 km. The western boundary currents are intensified because the strength of the Coriolis effect varies with latitude. The western boundary currents of the Atlantic Ocean are generally so deep that they are constrained against the continental shelf edge and do not reach the shore (see Brazil and Falklands (Malvinas) Currents and Florida Current, Gulf Stream and Labrador Current). In the Tropics the two subtropical gyres are connected via a complicated tropical circulation system. The tropical circulation shows a north-westward cross-equatorial flow at the western boundary and several zonal current and countercurrent bands (Figure 4) of smaller meridional and vertical extent and a lot of vertical diversification. The north-westward flow along the western boundary starts as a subsurface flow, the North Brazil Undercurrent, which becomes surface intensified north of the northeastern tip of Brazil by near-surface inflow from the South Equatorial Current and is named North Brazil Current. The North Brazil Current crosses the equator north-westwards and retroflects eastward at about 81N. In this North Brazil Current retroflection zone, eddies detach from the current and progress



50° W

75° W

W C G

60°N

25°E

0°

25° W

C

100° W

EG

722

60°N

IC

LC

NAC 40° N

40˚N S

AzC Ca

C

G

FC

20°N

NEC

20˚N

NECC/NEUC NB

C

0˚

0° SEUC

AC

NBUC

EUC

SECC

20˚S

SEC

Be

20°S C

BC 40˚S

ACC

FAC

60°S 100° W

40°S

SAC

75° W

50° W

25° W

0°

40°S ° 25 E

Figure 3 Schematic representation of upper-ocean currents in the North and South Atlantic Oceans in northern fall. For abbreviations of current bands see Table 1.

north-westward towards the Caribbean. In northern spring, when the North Equatorial Countercurrent is weak, there seems to be a continuous flow of about 10 Sv towards the Caribbean called Guyana Current. The westward flows are regarded as different bands of the South Equatorial Current, the northern one even crossing the equator. The eastward subsurface

flows are named the Equatorial Undercurrent at the equator, and the North and South Equatorial Undercurrents at about 51 latitude. The eastward surface intensified flows at about 91 latitude are the North and South Equatorial Countercurrents. In northern fall the North Equatorial Countercurrent and the North Equatorial Undercurrent override one



723

EQ 0 m 100

200

300

400 600 800 1000 1200

5˚

4˚

3˚

2˚

1˚

S

0˚

N

1˚

2˚

3˚

4˚

5˚

Figure 4 Zonal velocity component in m s1 (eastward flow is shaded) from direct velocity measurements (ADCP) across the equator at 351W in March 1994 north of the north-eastern tip of Brazil. Current branches are indicated and transport numbers are given in Sv. The figure shows the North Brazil Current (NBC), the South Equatorial Undercurrent (SEUC), the South Equatorial Current (SEC) with branches north and south of the equator separated by the Equatorial Undercurrent (EUC), the North Equatorial Undercurrent (NEUC), the Equatorial Intermediate Current (EIC), the Northern Intermediate Countercurrent (NICC), and the Southern Intermediate Countercurrent (SICC).

another and it is difficult to distinguish between the two current bands. In the Antarctic Intermediate Water layer at about 700 m depth there are intermediate currents at the equator (Equatorial Intermediate Current), as well as north and south of the equator (Northern and Southern Intermediate Countercurrents) which flow in the opposite direction to the currents above (Figure 4). The Intertropical Convergence Zone in the Atlantic, where the trade winds of both hemispheres converge, is located north of the equator throughout the year, and reaches the South Atlantic only in southern summer and then only at the north coast of Brazil. Seasonal changes of the wind field lead obvious variations in the tropical near-surface currents; however, with different strengths. The strongest seasonal signal is observed in the North Equatorial Countercurrent. The eastward-flowing North Equatorial Countercurrent is strongest in August, when the Intertropical Convergence Zone is located at its northernmost position. At that time the North Equatorial Countercurrent crosses the entire Atlantic basins zonally, but in late boreal winter it becomes weak or even reverses to westward in the western domain. South of the Cape Verde Islands at 91N, 251W, there is a cyclonic feature named Guinea Dome throughout the year, but it is weaker in northern winter. The

Southern Hemispheric counterpart is the Angola Dome at 101S 91E. The Angola Dome is seen only in southern summer and it is imbedded in a permanent larger-scale cyclonic feature centered near 131S, 51E called Angola Gyre. The western tropical Atlantic is a region of special interest in the global ocean circulation. The meridional heat transport across the equator is accomplished by warm surface water, central water, and subpolar intermediate water from the Southern Hemisphere moving northward in the upper 900 m mainly in the North Brazil Current, and cold North Atlantic Deep Water (NADW) moving southward between 1200 m and 4000 m. These reversed and compensating water spreading paths are often referred to as part of the global thermohaline conveyor belt. A clear distinction has to be made between the cross-equatorial flow at the western boundary and the interhemispheric water mass exchange. The latter is the amount of transfer from the Southern to the Northern Hemisphere of about 17 Sv in the upper ocean, and to a small degree in the Antarctic Bottom Water, compensated by the transfer from the Northern to the Southern Hemisphere of about 17 Sv by the North Atlantic Deep Water. The cross-equatorial flow within the North Brazil Current with about 35 Sv (Table 1) is much larger, since part of


724


Table 1 Major upper-ocean currents of the Atlantic Ocean and transport in Sv (1 Sv ¼ 106 m3 s1) Current name

Abbreviation in Figure 3

Transport in Sv

Subpolar gyre East Greenland Current West Greenland Current Irminger Current Labrador Current

EGC WGC IC LC

40–45 40–45 16 40–45

North Atlantic subtropical gyre Gulf Stream North Atlantic Current Azores Current Canary Current North Equatorial Current Florida Current

GS NAC AzC CaC NEC FC

90–130 35 12 5 20 32

NECC

40

NEUC

19 (mean)

EUC NBC NBUC SEUC

20–30 35 25 5–23

AC SECC

5 7

BC SAC BeC SEC

5–22 15–30 25 20 (southern band)

FAC ACC

up to 70 110–150

Equatorial currents North Equatorial Countercurrent North Equatorial Undercurrent Equatorial Undercurrent North Brazil Current North Brazil Undercurrent South Equatorial Undercurrent Angola Current South Equatorial Countercurrent South Atlantic subtropical gyre Brazil Current South Atlantic Current Benguela Current South Equatorial Current

Southern South Atlantic Falkland (Malvinas) Current Antarctic Circumpolar Current

this cross-equatorial flow originates from the zonal equatorial circulation, retroflects north of the equator, and returns into the equatorial circulation system. Poleward of the subtropical gyre the current field of the North and South Atlantic are completely different. In the North Atlantic a cyclonic subpolar gyre is present, driven in part by the wind stress curl associated with the atmospheric Icelandic low pressure system and in part by the fresh water from the subarctic. This subpolar gyre includes the northern part of the North Atlantic Current, the Irminger Current, the East and West Greenland Currents and the Labrador Current off north-eastern North America.

In the South Atlantic the continents terminate and an eastward flow of water all around the globe within the Antarctic Circumpolar Current is driven mainly by the midlatitudes westerlies. The South Atlantic counterpart of the Labrador Current is the Falkland (Malvinas) Current, which flows equatorward along the south-eastern South American shelf edge to about 381S. However, this current differs in origin as it is essentially a meander of a branch of the Antarctic Circumpolar Current. In the Brazil/Falkland (Malvinas) confluence region the Falkland (Malvinas) Current is retroflected southward to join the Antarctic Circumpolar Current. The South Atlantic Current as southern current band of the South Atlantic subtropical gyre and the Antarctic Circumpolar Current can be distinguished as separate current bands, nevertheless mass and heat exchange between the subtropics and subpolar region takes place in this region. The currents of the North Atlantic subpolar gyre have a strong barotropic flow component, which lead to large water mass transports (Table 1). As the major method of estimating transport is by geostrophy, which provides only the baroclinic component, early transport estimates for this region with strong barotropic flow fields largely underestimated the real transports. Another prominent example is the Falkland (Malvinas) Current in the South Atlantic, where estimates including the barotropic component lead to transports of up to 70 Sv, while earlier geostrophic computations resulted in transports of about 10 Sv. Differences between transports presented in Table 1 and transport values presented elsewhere might also arise from the location where the transport is estimated, as the mass transport changes along the flow path, or from different definitions of the boundaries of the current bands. For example, the Gulf Stream is measured to the deepest depth reached by the northward flow, while the southward-flowing Brazil Current is typically estimated only for the transport in the warmwatersphere, while the southward flow underneath is estimated separately as Deep Western Boundary Current.

Currents of the Deep Atlantic Ocean The deep-ocean circulation depends heavily on the changes in density imposed by air–sea interaction. The flow in the deep ocean is driven by the equatorto-pole differences in ocean density. This thermohaline circulation, driven by temperature and salinity gradients, provides global-scale transport of heat and salt. The forcing of this flow is concentrated in a few



areas of intense production of dense water in the far North Atlantic, the Labrador Sea, and along the margin of Antarctica within so-called convection areas. The water formed at the Antarctic continent is the densest water mass in the Atlantic and, once it has

100° W

75 ° W

50° W

725

crossed the Antarctic Circumpolar Current (ACC), spreads as Antarctic Bottom Water through the South Atlantic western basins northward into the North Atlantic (Figure 5), where it can usually be found near the seafloor even north of 401N. Actually, the real Antarctic water masses are so dense that they

25° W

0°

25°E

60°N

60° N

C

40° N

40° N MW

20°N

20°N

0°

0°

20°S

20°S

40°S

40°S

CDW 60°S

60°S 100° W

75° W

50° W

25° W

0°

25°E

Figure 5 Schematic representation of the large-scale North Atlantic Deep Water flow (solid lines), Circumpolar Deep Water (CDW), and Antarctic Bottom Water (dashed lines). C denotes the convection region in the Labrador Sea, MW the entrance of Mediterranean Water to the North Atlantic. For readability of the figure no recirculation cells are drawn. Bottom topography in 2000 m steps.


726


can be followed only to about 4.51S, and Lower Circumpolar Deep Water spreads to the Northern Hemisphere. However, for historical reasons the name Antarctic Bottom Water is used generally for this water mass and is used here for consistency. In the north, Antarctic Bottom Water is modified by mixing and contributes to the North Atlantic deep water formation. The deep water of the North Atlantic (NADW) is composed of different sources in the northern Atlantic. The source for the deepest NADW layer is dense water from the Greenland Sea which overflows the Denmark Strait and is called Denmark Strait Overflow Water or lower NADW. South of the Denmark Strait the lower NADW entrains surrounding water, which in part contains modified Antarctic Bottom Water. The middle layer of NADW is a combination of overflow across the Iceland–Scotland Ridge with a light component of modified Antarctic Bottom Water. The upper layer of the NADW is caused by open-ocean convection in the Labrador Sea and is called Labrador Sea Water or upper NADW. A closer investigation of the Labrador Sea Water shows that it has two different sources. Mediterranean Water entering over the Strait of Gibraltar spreads westward in the North Atlantic and contributes saline water mainly to the upper NADW. The NADW is trapped for some years within the deep-reaching North Atlantic subpolar gyre before it enters the Deep Western Boundary Current. Then the NADW spreads southward as Deep Western Boundary Current in the western ocean basins with recirculation cells to the east. When the NADW crosses the equator towards the South Atlantic part of the NADW flows eastward along the equator and then southward within the eastern basins. However, the major portion of the NADW continues to flow southward at the Brazilian continental margin as Deep Western Boundary Current. When the NADW reaches the latitude of the ACC the NADW is modified by mixing as it is carried eastward with the ACC around the Antarctic continent. Branches of the modified NADW, now often referred to as Circumpolar Deep Water, move northward again into the Indian and Pacific Oceans.

Atlantic were influenced by the interest in the resources of the sea for food, and the search for economic sources. The research focus on ocean currents in the Atlantic Ocean at the beginning of the new millennium will be improvement in understanding the mean surface and deep currents of the Atlantic, and its variability on short-term to decadal timescales to clarify the ocean’s role in climate changes. These investigations are driven by the need to protect and manage the environment and the living conditions of all countries and are managed in large international research programs. To improve the climate prediction models it is necessary to understand the ocean’s role in climate changes better. International programs like Climate Variability and Predictability (CLIVAR) started to describe and understand the physical processes responsible for climate and predictability on seasonal, interannual, decadal, and centennial timescales, through the collection and analysis of observations and the development and application of models of the coupled climate system. Another new important focus will be to describe and understand the interactive physical, chemical, and biological processes that regulate the total Earth system. This is also the overall objective of the International Geosphere-Biosphere Program (IGBP). One core project of IGBP is GLOBEC, which is now changing from a planning to a research status with the goal of advancing understanding of the structure and functioning of the global ocean ecosystem, its major subsystems, and its response to physical forcing so that a capability can be developed to forecast the responses of the marine ecosystem to global change. Despite the future focus on the Atlantic’s role in climate changes as well as interactive processes, and although the major components of the near-surface circulation from ship drift observations have been known for more than 100 years, there is still also the need to investigate details of the Atlantic Ocean subsurface and abyssal circulation and its physical processes, which so far are unrevealed.

See also Future Aspects Ocean research is always influenced by political and economic interests. The improvement in understanding of the surface currents at the time of the voyages of discovery was caused by the need for good and safe sailing routes. The more detailed look at the currents of the surface as well as the deep

Abyssal Currents. Atlantic Ocean Equatorial Currents. Baltic Sea Circulation. Benguela Current. Brazil and Falklands (Malvinas) Currents. Canary and Portugal Currents. Current Systems in the Mediterranean Sea. Current Systems in the Southern Ocean. Florida Current, Gulf Stream and Labrador Current. Intra-Americas Seas. North Sea Circulation. Ocean Circulation.



Further Reading Krauss W (ed.) (1996) The Warmwatersphere of the North Atlantic Ocean. Berlin: Gebru¨der Borntraeger. Peterson RG, Stramma L, and Kortum G (1996) Early concepts and charts of ocean circulation. Progress in Oceanography 37: 1--115. Robinson AR and Brink KH (eds.) (1998) The Sea, Ideas and Observations on Progress in the Study of the Seas The Global Coastal Ocean, Regional Studies and Syntheses, vol. 11. New York: Wiley.

727

Segar DA (1998) Introduction to Ocean Sciences. London: Wadsworth Publishing Company. Tomczak M and Godfrey JS (1994) Regional Oceanography An Introduction. Oxford: Elsevier. Wefer G, Berger WH, Siedler G, and Webb DJ (eds.) (1996) The South Atlantic Present and Past Circulation. Berlin: Springer Verlag. Zenk W, Peterson RG and Lutjeharms JRE (eds.) (1999)New view of the Atlantic: A tribute to Gerold Siedler. Deep-Sea Research II 46: 527.


CURRENT SYSTEMS IN THE INDIAN OCEAN M. Fieux, Universite´ Pierre et Marie Curie, Paris, France G. Reverdin, LEGOS, Toulouse Cedex, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 598–604, & 2001, Elsevier Ltd.

The Overlying Atmosphere

Introduction The Indian Ocean is the smallest of all the oceans and is in several respects quite different from the others. In particular, it is bounded by the Asian continent to the north. This meridional land–sea contrast has a strong influence on the winds, resulting in a complete seasonal reversal of the winds known as the monsoon system. The characteristics of the basin and of the wind regime are determinant for the currents, and will be described first in this article. The description of the currents has been separated into two main sections: the first for the southern part of the Indian Ocean which is not affected by the monsoons and is more akin to the other subtropical oceans; and the second for the northern part which undergoes forcing through the reversal of the monsoon winds. Some information on the deep circulation and a short conclusion are then provided.

Characteristics of the Indian Ocean Basin The Indian Ocean basin is the smallest of the five great subdivisions of the world ocean with 49.106 km2 out of the 361.106 km2 of the global ocean (Figure 1). It is closed to the north around the latitude of the Tropic of Cancer by the Asian continent, which has important consequences on the ocean circulation. South of the equator, its western boundary is modified by the presence of the island of Madagascar. In the east, the basin is connected with the equatorial Pacific Ocean through the deep passages of the Indonesian Seas. The north of the Indian Ocean is made up of the large basins on either side of the Indian peninsula, the Arabian Sea in the west and the Bay of Bengal in the east which drains most of the river runoff from the Himalayas and the Indian subcontinent. The Arabian Sea is connected directly to the shallow Persian Gulf, and through the sill of Bab-el-Mandeb (110 m) to the deep Red Sea basin where high salinity waters are formed. In the south, the basin is largely open to the Antarctic Ocean

728

between South Africa and Australia. The Indian Ocean limit to the south is the Subtropical Convergence, a hydrological limit where the meridional surface temperature gradient is maximum. At depth, the complicated system of ridges separates the Indian Ocean in many deep basins (Figure 1).

Due to the presence of the Asiatic continent to the north, the atmospheric circulation is quite different from the Pacific Ocean and the Atlantic Ocean, particularly north of 101S. Seasonal heating and cooling of the atmosphere over Asia induces a seasonally varying monsoon circulation (Figure 2). For centuries it has been known that the winds north of around 101S reverse with the seasons. A long time ago the Arabic traders along the east African coast made use of the fair currents and winds during their voyages. The word ‘monsoon’ comes from the Arabic word ‘mawsin’ meaning season. As the winds are the main driver of the currents, in particular near the surface, the main characteristics of the wind seasonal variability will be described below. The wind seasonal variability over the ocean can be separated in four periods: the winter monsoon period, the summer monsoon period, and the two transition periods between the two monsoons. Between December and March–April, north of the equator, the winter (NE) monsoon blows from the north east with a moderate strength. At the equator the winds are weak and usually from the north. Between the equator and the Intertropical Convergence Zone (ITCZ) which stretches zonally near 101S between north Madagascar and south Sumatra, it blows from the north west. During that season (southern summer), the atmospheric pressure decreases over Australia and South Africa and the subtropical high pressure over the ocean, around 351S, is weaker – as are the south-east trade winds during that season. In April–May, during the transition period between the end of the NE monsoon and the beginning of the SW monsoon, the winds north of the equator calm down. At the equator moderate eastward winds blow, which contrasts with the westward winds over the equatorial Pacific and Atlantic Oceans. From June to September–October during the SW monsoon, the winds reverse completely and north of the equator the summer monsoon blows steadily from the south west. The SW summer monsoon is much


CURRENT SYSTEMS IN THE INDIAN OCEAN

729

30˚N

0˚

30˚

30˚E

90˚

60˚

120˚

Figure 1 Map of the Indian Ocean with the different basins.

40°E

40°E

60°E 80°E 100°E 120°E

60°E 80°E 100°E 120°E L

H

20°N

20°N

0°

0°

20°S

20°S

20°N

20°N

0°

0°

20°S

20°S

H

H

40°S

(A)

40°S

40°E 60°E 80°E 100°E 120°E Prevailing winds in January Mean position of ITCZ

40°S

(B)

40°S

40°E 60°E 80°E 100°E 120°E Prevailing winds in July Mean position of ITCZ

Figure 2 Prevailing winds during (A) the northern monsoon (January); (B) the southern monsoon (July); double dashed lines indicate the ITCZ.


730


stronger than the NE winter monsoon. A wind jet, also called the Finlater jet, develops along the high orography of the east African coast. As a consequence, the winds are the strongest on the western side of the Indian Ocean along the Somali coast towards the Arabian Sea, particularly north-east of Cape Guardafui (the horn of Africa) where the mean July wind speed is 12 m s 1 with peaks exceeding 20 m s 1. They are the strongest and the steadiest wind flow in the world. At the equator, the winds are moderate from the south and decrease eastward. In the southern Indian Ocean, the subtropical high pressure center intensifies and covers the whole width of the southern Indian Ocean during the southern winter (July) and the SE trades penetrate farther north than during the southern summer (January); they reach the equator in the western part of the ocean and are the strongest among the three oceans. During that season the air masses transported by the SE trade winds cross the equator in the west and continue, loaded with moisture, towards the Asian continent where they bring the awaited monsoon rainfall. October–November corresponds to the second transition period between the end of the SW monsoon and the beginning of the NE monsoon. North of the Equator, the winds vanish and the sea surface temperature can exceed 301C. At the equator moderate eastward winds blow again as during the first transition period, although they are usually slightly stronger. This particular wind regime implies that at the equator, the zonal wind is dominated by a semi-annual period associated with the westerly winds of the transition periods, while the meridional wind presents a strongly annual period associated with the monsoon reversals. The winds off the equator also present a strong annual period. Along the western Australian coast the northward winds, favorable for upwelling, are much weaker than in the eastern Pacific and Atlantic Oceans. They even drop down during the SW monsoon season. This is due to the different land–ocean distribution. The seasonal changes of the winds south of latitude 101S are smaller than to the north, and therefore the variability in the ocean circulation will also be smaller there. The next section will describe the currents in this area, before presenting the currents in the northern area.

The Currents in the Southern Part of the Indian Ocean The strong anticyclonic subtropical gyre of the southern Indian Ocean is the result of the large wind

stress curl between the Antarctic westerlies and the SE trade winds. Its northern branch is the westwardflowing South Equatorial Current (SEC) centered between 121S and 201S, fed in its northern part by the throughflow waters originating from the Indonesian Seas and corresponding to lower salinity waters (Figure 3). The SEC is the limit of the influence of the monsoon system. Its mean transport relative to the 1000 dbar level varies from 39 Sv in July–August to 33 Sv in January–February. Its latitudinal range varies between 81S and 221S in July– August and 101S–201S in January–February. The SEC impinges on both the east coast of Madagascar and on the east African coast, resulting in several intensified boundary currents along these coasts. The SEC splits into a northward flow and a southward flow east of Madagascar near 171S. The southern branch continues as the East Madagascar Current (EMC) carrying of the order of 20 Sv (0–1100 m), which ultimately joins the Mozambique Current and the Agulhas Current (AC) to the south west, besides some recirculation to the east into the subtropical gyre. The northern branch of the SEC splits again east of the African coast near cape Delgado (111S) into the southward-flowing Mozambique Current (MC) and the northward-flowing East African Coastal Current (EACC). The Mozambique Current presents intense recirculation in the northern Mozambique Channel, but ultimately feeds roughly 20 Sv into the Agulhas Current which is the strongest western boundary current in the south Indian Ocean, transporting nearly 70 Sv. The eastward-flowing south branch of the anticyclonic subtropical gyre of the southern Indian Ocean is part of the Antarctic Circumpolar Current (ACC). Nevertheless, north of the ACC, a South Indian Ocean Current (SIC) can be differentiated from the different cores of the ACC. The SIC comprises the eastward flow recirculating part of the Agulhas Current off South Africa and at depth transports North Atlantic Deep Water. The ACC transports Antarctic circumpolar waters with lower salinity than in the South Indian Ocean Current. The eastern Indian Ocean is connected with the Pacific Ocean through channels in the Indonesian Archipelago. This has large consequences on the eastern Indian Ocean circulation and is expected to be one of the causes for the southward flow west of Australia, the Leeuwin Current, which flows opposite to the wind. The mean sea level, higher on the Pacific side than on the Indian Ocean side of the Indonesian Seas, drives a throughflow towards the Indian Ocean transporting warmer and fresher waters contributing to a high dynamic height in the north of the western Australian coast. This induces



30°E

50°E

70°E

90°E

30°E

110°E

20°N

20°N

10°N

10°N

0°

0°

10°S

10°S

20°S

20°S

30°S

30°S

50°E

70°E

90°E

731

110°E

Fieux aft. Cutler & Swallow

(A)

40°E

60°E

80°E

100°E

120°E

(B)

40°E

60°E

Surface current in January 30°E

50°E

70°E

90°E

80°E

100°E

120°E

Surface current in April

110°E

30°E

20°N

20°N

10°N

10°N

0°

0°

10°S

10°S

20°S

20°S

30°S

30°S

50°E

70°E

90°E

110°E

Fieux aft. Cutler & Swallow

(C)

40°E

60°E 80°E 100°E Surface current in July

120°E

(D)

40°E

60°E 80°E 100°E Surface current in October

120°E

Figure 3 General surface circulation in the Indian Ocean: (A) during the NE monsoon; (B) during the transition period in April; (C) during the SW monsoon; (D) during the transition period in October. (Adapted from Cutler AN and Swallow JC (1984) Surface Currents of the Indian Ocean (to 251S, 1001E): compiled from historical data archived by the Meteorological Office, Bracknell, UK. Institute of Oceanographic Sciences, Wormley, UK Rep. 187, 8pp and 36 charts.)

an alongshore pressure gradient off western Australia which drives the Leeuwin Current to the south and can even overwhelm the counteracting effect of the coastal upwelling induced by the weak southerly winds. At 221S, the Leeuwin current is a 30–50 km wide and shallow (150–200 m) poleward jet close to the coast (max in May–June) with large intraseasonal interannual variability which is associated to an equatorward undercurrent.

Response of the Indian Ocean Circulation to the Wind Variability in the Northern Part North of 101S, the ocean circulation responds to the seasonally varying monsoon winds and as a consequence presents well defined seasonal characteristics. It is necessary to distinguish the circulations near the western boundaries, near the equator, in the

northern ocean interior, and the eastern boundary current systems, which have different dynamics. They will be described successively, on the basis of observations as well as numerical or analytical modeling studies. The Western Boundary Current System

As in the other oceans, the strongest currents are close to the western shores of the ocean as a result of the direction in which the earth rotates and the variation of the Coriolis parameter with latitude. North of 101S (which is the limit of the monsoon influence), there are two western boundary currents: the East African Coastal Current (EACC) which always flows northward and the Somali Current which is the most intense and the most variable. The EACC flows in continuity with the branch of the SEC which passes north of Madagascar and splits around 111S. It runs northward throughout the year


732


between latitudes 111S and 31S. Its surface speed can exceed 1 m s 1 during northern summer and its transport amounts to 20 Sv in the upper 500 m. Its northern end depends on the season. In northern winter, the EACC converges around 31S–41S with the south-going Somali Current to form the eastward South Equatorial Countercurrent. During the northern summer, the EACC merges into the northgoing Somali Current. The Somali Current is the most intense, but unlike the other western boundary currents it is highly variable due to the complete seasonal reversal of the winds. It was first studied during the International Indian Ocean Experiment in the 1960s, during INDEX (INDian ocean EXperiment which started in the 1970s) and during SINODE in the 1980s. Recently (1990–96), during WOCE (World Ocean Circulation Experiment) considerable amounts of new data were collected over the whole Indian Ocean. The Somali current develops in different phases in response to the winds. During the transition period after the NE monsoon, in April, the Somali current flows south-westward along the coast south of 51N, merging near the equator with the northward-flowing EACC. This feeds a south-eastward flow towards the ocean interior. In early May, the Somali current responds rapidly to the onset of the local southerly winds and reverses northward in continuity with the northward EACC near the equator. By mid-May, the SW wind onset propagates northward and the current turns offshore towards the east at 41N. North of that branch an upwelling wedge spreads out, bringing cold and enriched waters at the surface. Further to the north, the current flows northward from March onwards. When the onset of the strong summer monsoon winds occurs at these latitudes in June, the southern branch increases in strength and a strong anticyclonic gyre, called ‘the great whirl’ develops between 51N and 101N. Between the Somali coast and the northern branch of the great whirl a second upwelling wedge forms. Numerical models have shown that the location and motion of these structures are influenced by the distribution and strength of the wind forcing. In August–September, when the winds decrease, the southern cold wedge propagates northward along the coast and meets with the northern one (although the latter probably also moves). It is only at that time that the Somali current is continuous from the equator up to 101N and brings fresher waters into the Arabian Sea. During the transition period in October–November, the northward Somali circulation decreases.

In December–February, during the NE winter monsoon, the Somali Current reverses southward from 101N to 51S where it converges with the northward EACC to form the South Equatorial Counter Current (SECC) flowing eastward. This countercurrent exists only during the NE winter monsoon and could be compared to the other Equatorial Counter Currents in the Atlantic Ocean and in the Pacific Ocean. It develops just north of the intertropical convergence zone (ITCZ) where the winds have an eastward component. At the equator near Africa, the reversal affects only a thin surface layer below which (between 120 m and 400 m) there is a northward undercurrent, remnant of the SW monsoon season, followed again by a southward current below 400 m. There are also western boundary currents in the Gulf of Bengal off the coasts of Sri Lanka and India. From September to January, the currents are southward along the whole eastern coast of India and Sri Lanka, bringing fresh Bengal Bay water to low latitudes (61N). In February–March the currents reverse to flow to the north along these coasts with a separation from the coast in the northern Bay of Bengal (191N) in March. From April to August, the current reverses along the eastern coast of Sri Lanka where it flows to the south. Further north, from May to July, the separation from the coast of the northward current takes place around 161N instead of 191N. In July, north of 161N, reversal to the south takes place. This seasonal cycle is markedly different from the one off the Somali coast. Modeling studies show that it is a response both to local winds, to the curl of the wind stress over the Bay of Bengal, with other contributions propagating along the coast from further east, and from the vicinity of the equator. Equatorial Currents System

The winds at the equator are profoundly different in the Indian Ocean from the mostly westward winds in the Atlantic and the Pacific tropical oceans. Instead, in the Indian Ocean, there is a strong semi-annual cycle in the zonal winds and the mean zonal wind is westerly. During the two transition periods between the monsoons, a strong eastward jet (called ‘Wyrtki jet’) occurs in a narrow band, trapped within 21–31 of the equator, mostly in the central and eastern parts, driven by the equatorial westerly winds. Due to the efficiency with which zonal winds can accelerate zonal currents at the equator where the Coriolis force vanishes, the current speeds can rapidly reach >1 m s 1. The jet usually peaks in November with velocities which can reach 1.5 m s 1; it could also



reach these values in May as there is large intraseasonal and interannual variability. At the equator, in the middle of the Indian Ocean near Gan Island (731E), measurements show currents throughout the upper 100 m in phase with local winds, which reverse four times a year. The associated change of current direction produces semi-annual variations in the thermocline depth and sea level. During periods of eastward flow the thermocline rises off Africa and falls off Sumatra corresponding to opposite displacements of the sea level. The set up of the jet is apparently triggered by the westerly winds during the transition periods. This forces a local response as well as waves propagating the response further to the east (Kelvin waves) and west (Rossby waves). The stopping of the jet seems to happen progressively from east to west, as it has been observed with drifting buoys. This is interpreted dynamically as westward propagating decelerating Rossby waves which are generated when the eastward jet reaches the coast of Sumatra. During the fully developed SW monsoon in July– August, along the equator – aside from the extreme west where the strong north-eastward Somali Current occurs – the winds are southerly and light, and currents at the equator are weak and variable. An eastward equatorial undercurrent embedded in the thermocline along the Indian Ocean equator exists only during January–June and is strongest in March at the end of the NE monsoon. It is confined between 2130N and 2130S and is weak east of 801E. During the NE monsoon, it flows under a weak westward current until the eastward Wyrtki jet starts, then the whole upper layer flows eastward.

The Northern Interior Current System

During the NE monsoon in December–February, the northern Indian Ocean presents a current structure similar to those found in the other oceans. In the northern Arabian Sea, the circulation is not well defined during this season. There is a general westward flow south of 101N, the North-east Monsoon Current (NMC), extending south to about 21S, with speeds between 0.3 and 0.8 m s 1. South of Sri Lanka, the current splits into a branch continuing westward and a branch which tends to follow the western coast of India, possibly after meandering in eddies off south-west India. Further south between 21S and 81S the eastward South Equatorial Counter Current (SECC) flows eastward starting at the convergence of the southward Somali Current with the northward EACC (see above).

733

During the SW monsoon in the Arabian Sea, there is a general eastward flow in the South-west Monsoon Current (SMC), with more intense veins near 151N as well as near 9–101N (although the latter might be more developed in April–June) and near 51N. In the Arabian Sea, there is some anticyclonic recirculation to the south of the northern eastward vein, very likely forced by the curl of the wind stress, and some indication of cyclonic eddies to the north of this circulation. The flow along the western coast of India is southward during the SW monsoon associated with a poleward undercurrent along the shelf which sometimes could reach the surface. In the northern Bay of Bengal, north of 151N, the eastward currents are already set in April–May, which last until August (in particular near 151N–171N). Eastward currents also exist south of 81N in the Bay of Bengal from April to September. The currents are intensified south of Sri Lanka during both monsoons, resulting in particularly large eastward SMC and westward NMC. Numerous eddies are also present, especially in the western parts of the Arabian Sea and Bay of Bengal, associated with upwellings that are particularly strong off the Arabian coast during the SW monsoon. The Eastern Boundary Current System Affected by the Monsoons

Along the eastern boundary the currents are also seasonally variable, except along Sumatra, south of the equator, where it is always south-eastward and flows against the winds in June–September. During the transition periods the Kelvin waves associated with the Wyrtki jet continue north-westward and south-eastward as coastal trapped Kelvin waves along the Indonesian islands. South of Java they reinforce the NE monsoon-driven south-eastward Java current, but work against the response of the Java boundary current to the SE monsoon onset in May– June. So when the equatorial Kelvin waves arrive along the Java coast in October–November, the reversal is faster than in May–June. In July–September when the Java current is north-westward, there is a convergence with the south-eastward Sumatra current south of Sumatra. It is also during that season that large-scale wind-driven upwelling occurs along the coast of Java. The open eastern boundary of the Indian Ocean allows exchanges with the Pacific Ocean through the channels of the Indonesian Archipelago. This flow is called the Indonesian throughflow. It transports waters from the surface down to 1300–1400 m which are principally drawn from the northern Pacific and modified in the Indonesian archipelago


734


under both effects of exchanges with the atmosphere and dynamical mixing. The resulting water entering the Indian Ocean is a well characterized Indonesian Water. The principal route of the throughflow goes through Makassar Strait then through Lombok Strait (350 m deep between Lombok and Bali), and north and south of Timor Island (mean sills depth around 1350 m). The few partial direct measurements show strong interannual, seasonal, and intra-seasonal variability. This throughflow is entrained westward into the SEC bringing fresh and warm water to the Indian Ocean. It is also part of the source waters of the southward Leeuwin current which brings fresh and warm waters to the south along the west Australian coast. The surface flow along the western coast of India is usually south-eastward with a north-westward subsurface undercurrent, in particular during the SW monsoon with upwelling favorable winds. However, the currents reverse at the end of the year bringing a pulse of fresher Bay of Bengal water along the coast. Deep Circulation

Most of the circulation described here concerns the upper part of the ocean. The deep circulation is relatively unknown. The Indian Ocean is separated into numerous deep basins connected through narrow passages (Figure 1). As a consequence, the Antarctic Bottom Water cannot reach the northern ocean basins. Recent long-term direct measurements were carried in the Crozet-Kerguelen Gap of the South-west Indian Ridge, one of the major channels through which Antarctic Bottom Water can move equatorwards. The annual northward transport of Antarctic Water at depth greater than 1600 m amounts to 11.5 Sv which, because it has undergone large dilution through mixing, corresponds to an initial volume of Antarctic Bottom Water of 2.5–3 Sv deduced from CFC distribution. By contrast, further north, in the Amirante Passage connecting at depth the Mascarene Basin and the Somali Basin, the flow of bottom water flowing northward has been estimated to be 2.5–3.8 Sv. From the characteristics of the water masses, an intensification of the deep flow is found as deep

western boundary currents against the eastern flanks of each meridional ridge separating the numerous deep basins. Some of this deep, intermediate, and subsurface water flowing northward in the Indian Ocean is upwelled and contributed to the cooling of the surface waters. Conclusion

The northern Indian Ocean is a natural laboratory to study the effect of the wind on the oceanic circulation, as regularly twice a year the winds change direction rapidly and are particularly strong in the western boundary. The highest variability as well as the highest current speed of the world ocean are found in the Somali current. At the equator, particularly between the two monsoon seasons, westerly wind bursts entrain a strong eastward equatorial jet twice a year which could have an effect on the strength of the transport coming from the Pacific Ocean. Some of the variability of these wind bursts seems to be related to the El Ninõ–La Ninã climatic variability. Comparing the width and the external forcing of the Pacific and Indian oceans, the Indian Ocean at semi-annual frequency should behave dynamically like the Pacific at annual frequency.

See also Agulhas Current. Antarctic Circumpolar Current. El Ninõ Southern Oscillation (ENSO). Elemental Distribution: Overview. Indian Ocean Equatorial Currents. Leeuwin Current. Pacific Ocean Equatorial Currents. Somali Current. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Open University, Oceanography Course Team (1993) Ocean Circulation. Oxford: Pergamon Press. Tomczak M and Godfrey S (1994) Regional Oceanography: An Introduction. Pergamon: Elsevier. Fein JS and Stephens PL (eds.) (1987) Monsoons. Washington: John Wiley Sons.


CURRENT SYSTEMS IN THE SOUTHERN OCEAN A. L. Gordon, Lamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 613–621, & 2001, Elsevier Ltd.

Summary The Southern Ocean, encircling Antarctica, plays a major role in shaping the characteristics of the global ocean. It provides the most significant inter-ocean conduit by which waters of the three major oceans, the Atlantic, Pacific and Indian, can intermingle, acting to diminish their differences in temperature, salinity, and chemical properties. The most prominent current is the Antarctic Circumpolar Current, which transfers about 134 million m3 s1 of sea water from west to east within a latitudinal range from 501 to 601S. North of the Antarctic Circumpolar Current are the poleward limbs of the large subtropical gyres of the southern hemisphere, W 20°

referred to as the South Atlantic, South Indian, and South Pacific Currents. Within large embayments of Antarctica, notably the Weddell and Ross Seas, south of the Antarctic Circumpolar Current are large clockwise flowing gyres. The Weddell Gyre carries about 30–50 million m3 s1 of water. Along the continental margin of Antarctica is the coastal current that advects water from east to west. The coastal current is directed towards the north along the east coast of Antarctic Peninsula, forming the western boundary of the Weddell Gyre. At the northern tip of the Antarctic Peninsula the coastal current is directed into the open ocean. The coastal waters injected into the open ocean separate the Antarctica Circumpolar Current from the Weddell Gyre, in what is called the Weddell-Scotia Confluence. Besides ocean currents flowing on nearly horizontal planes, the Southern Ocean experiences major overturning of ocean water. Overturning is forced by the production of dense surface water along the margins of Antarctica, leading to the formation of Antarctic Bottom Water. Within the Antarctic Circumpolar Current, 0° E 20°

0.0

_

0.5

40°

40° _ 1.0 0.0

60°

0.5 1.0 1.5 2.0

0.0

60°

_ 0.5 _

0.5

2.64 _ 1.0

80°

80°

0.0 _ 1.25

2.2

0.0

2.52

100°

0.5 1.0

100° _ 1.0

1.5

1.0 0.5

0.5 0.0

120°

120° _ 0.5

_ 0.5

140°

140° _ 0.5

160°

160° 180°

Figure 1 Wind stress in N m2. (Reproduced with permission from Nowlin and Klinck, 1986.)


735

736

CURRENT SYSTEMS IN THE SOUTHERN OCEAN

at the Polar Front, surface waters sink under the more buoyant surface water to the north, forming Antarctic Intermediate Water, a low salinity intrusion spreading at a depth of nearly 1000 m under the main thermocline of subtropical ocean. The horizontal and vertical circulation influences the distribution of sea ice, which in turn modifies the heat, freshwater, and gas exchange between the Southern Ocean and polar atmosphere.

of 651S the winds are easterlies, marking the northern edges of the polar high pressure over Antarctica. Density changes of surface water induced by sea– air fluxes of heat and fresh water, often involving sea ice within the Southern Ocean, also produce circulation. However, buoyancy (sometimes referred to as thermohaline), circulation is sluggish and mainly occurs within the meridional vertical plane, as dense water sinking forces slow, compensatory upwelling of less dense resident water. Sinking of dense waters along the continental margins of Antarctica results in Antarctic Bottom Water. Ocean currents are for the most part in equilibrium with the distribution of density within the ocean (Figure 3) satisfying approximately the socalled ‘thermal wind equation’ (see Elemental Distribution: Overview. Ocean Circulation.) Along the Greenwich meridian the surface of equal density, or isopycnals, rise up towards the sea surface as latitude increases to the south. Strongly sloped isopycnals are coupled to strong ocean current, more precisely to strong geostrophic ocean currents, relative to the

Introduction Ocean currents are the product for the most part of the stress exerted on the sea surface by the wind. The winds are strong over the Southern Ocean, particular within the Indian Ocean and Australian sectors (Figure 1) and therefore drive a vigorous circulation (Figure 2). Strong westerlies (wind directed from west to east) extend from the subtropical high atmospheric pressure near 301S, a latitude often used to define the northern limits of the Southern Ocean, to a belt of low atmospheric pressure at 651S. South

Ch ile C urren t

_ 90°

ast

Co ti

c

arc A nt

Ke r

n Current uth India So

0 15

°

gu Gyilen re

° _ 30

y

S

ica lG

rop

t Eas tralia Ausrrent Cu

as Agulh Ag ent ul Curr h a s Ret ur n

al Cu rre nt

yre el l G ed d

180°

Ross Gyre

W

0°

Benguela Current

30 °

Weddell Scotia Confluence

nt rre Cu

0°

Sub t

ctic Circ

um

tar An

yre pical G btro Su urrent ntic C Atla ut h So t urren ar C pol

_ 15

rent Cur

°

ific ac P th ou

B C aaz u rre il nt

re

_ 60

°

_ 120

°

60

0°

12

90° Figure 2 Schematic of general circulation of the Southern Ocean. Land is black; the shaded area marks ocean depths of o3000 m.



South Atlantic Current 0

Antarctic Circumpolar Current

Weddell Gyre

14

_ 0.4

12

0.2

0

10

500

500

6

0.2

4

1000

1000 0

3

1500

1500 _ 0.2

2000

Depth (m)

2000 _ 0.4

2500

2500 3000

3000 2 _ 0.6

3500 4000

3500 4000

1 _ 0.8

0.8

4500

4500

5000

5000 Potential temperature

5500

5500 30

35

40

45

(A)

50

55

60

65

70

Latitude, °S

South Atlantic Current


0

Weddell Gyre 0

34.4 34.6

35.5 35 34.8 34.6 34.5 34.4

500

500 1000

1000 34.4 34.5

1500

1500

34.6 34.7

2000

2000

34.66

Depth (m)

34.8

2500

2500

3000

3000

3500

3500

4000

4000

4500

4500 5000

5000 Salinity

5500 30 (B)

35

40

45

50

5500 55

60

65

Latitude, °S

Figure 3 Potential (A) temperature, (B) salinity, and (C) density (sigma-0) along the Greenwich meridian.


70

737

738


South Atlantic Current 0


Weddell Gyre

26.6 26.8

500

0

27.6 27.8 27.82

26.4

500

27 27.2

1000

1000

27.84 27.4

1500

1500

27.6

2000

Depth (m)

2000 2500

2500

27.8 27.82

3000

3000

27.84

3500

3500 4000

4000

27.86

4500

4500

5000

5000 Density (sigma 0)

5500

5500 30 (C)

35

40

45

50 55 Latitude, °C

60

65

70

Figure 3 Continued

seafloor. As a rule of thumb, in the southern hemisphere higher density water occurs to the right of the direction of the ocean current. Hence the increasing density as latitude increases is linked to west to east flow of water. Along the margin of Antarctica the descent of isopycnals marks a flow towards the east. Regions of rapid changes in temperature, salinity or density mark the positions of ocean fronts, often coinciding with a strong ocean current. Maximum westerlies in the wind occur near 551S, which roughly coincides with the axis of the Antarctic Circumpolar Current. To the south of this latitude, wind-induced northward Ekman transport of surface water results in a wide region of upwelling. North of 551S the Ekman transport diminishes. This causes a region of surface water convergence. Near 551S surface water sinks under the more buoyant surface water to the north, producing Antarctic Intermediate Water. Antarctic Intermediate Water forms a low salinity layer found at the base of the thermocline of the subtropical southern hemisphere regions. Upwelling poleward of the maximum westerlies brings deeper water to the sea surface to compensate for the sinking of Antarctic Bottom Water and Antarctic Intermediate Water. The upwelling also drives two large clock-

wise-flowing, cyclonic Gyres within the large embayments of Antarctica, marking the Weddell and Ross Seas (Figure 2) and a smaller one east of Kerguelen Plateau.

Antarctic Circumpolar Current The most prominent current of the Southern Ocean is the west to east flowing Antarctic Circumpolar Current lying within a latitudinal range from 501 to 601S. It is the greatest ocean current on the Earth, covering a distance of 21 000 km, with an average transport through the Drake Passage (between South America and Antarctic Peninsula) of 134 million m3 s1 (134 Sv). The transport varies with time mirroring variations in the circumpolar wind field, from about 100 to 150 million m3 s1. Transport is enhanced south of Australia by return of water to the Pacific Ocean lost to the Indian Ocean within the Indonesian Seas, by about 10 million m3 s1. The Antarctic Circumpolar Current transport passing between Tasmania and Antarctica is estimated as 143 million m3 s1, with a range from 131 to 158 million m3 s1. The Antarctic Circumpolar Current is a deep reaching or barotropic current, meaning that it extends to the seafloor. Because of this it is said that the



739

90˚W

60

0˚W

˚W

1.4

2

0˚W 15

1.8

˚W 30

1.6

1 .4

1.8

8

12

1.2

1 2

1

0˚

180˚

1.6

0.8 1.4

1.2

˚E 30

1.2

0.8

0˚E

1.

4

15

8

1.

1

1.8

2 1.6

12

˚E

60

2

0˚E

90˚E Figure 4 Anomaly of Southern Ocean sea level height relative to the 2 107 Pa pressure surface. The values are differences in dynamic meters (approximately equivalent to geometric meters) from that of a standard ocean (01C, 35 PSU salinity). Land is black; the shaded area marks ocean depths of o3000 m. Geostrophic ocean currents in the southern hemisphere are directed so that lower sea level is to the right of the flow direction, which is aligned along lines of equal sea level height. The rise of sea level towards the north defines the eastward flowing Antarctic circumpolar current.

Antarctic Circumpolar Current ‘feels’ the shape of the sea floor and hence its path is steered by the seafloor topography (Figure 4). The flow following the southern deflection in the mid-ocean ridge reaches its southern-most position in the southwest Pacific Ocean, near 601S. Upon passing through Drake Passage it turns sharply to the north, transversing the Atlantic Ocean near 501S. As the ocean surface temperature pattern responds to the circulation pattern the surprising result is that the bottom topography is ‘projected’ in the sea surface temperature pattern. Rather than a broad diffuse flow, the Antarctic Circumpolar Current is composed of a number of high speed filaments, separated by zones of low flow, or even reversed flow (towards the west). The jets are typically 40–50 km wide. Surface currents average about 30–40 cm s1 within the axes, and speeds of over 100 cm s1 are common. The high speed filaments are marked by ocean fronts, where the temperature and salinity stratification changes rapidly

with latitude (Figure 5). Between these fronts are zones of similar stratification. The primary axis occurs at the polar front (Figure 6). Meanders of the flow axes and associated fronts displace these features by at least 100 km to either side of their mean position. Meanders occasionally produce detached eddies, in which pools of water ringed by a high speed current from one zone invade an adjacent zone. A characteristic of the Antarctic Circumpolar Current is its high degree of eddy activity (Figure 7). The most active eddy fields are observed where the Antarctic Circumpolar Current crosses submarine ridges or plateaus, as south of Australia, in the southwest Atlantic and south of Africa. The correlation of the eddy currents and of the ocean temperature leads to significant poleward flux of ocean heat by the eddy processes. The poleward heat flux measured in the Drake Passage, if extrapolated all around Antarctica, is 0.3 PW, which can account for most of the meridional heat flux across 601S. However caution is suggested as the


740


68

64

66

58˚W

60

62

c cti tar n ba e Su Zon

Cape Horn

56˚S

t

27

28

n Fro

nt

ba

Su

tic arc

29 30

lar Po ntal Fro one Z

32

58 t

34

lar Po

36

n Fro

38

40

c cti tar An Z one

42 43

44

tal en n i nt ter y Co Wa dar un Bo

45 46 47 48

5

10 15 cm s

60

tal en tin e n Co Zon

_1

62

South Shetland Islands

Figure 5 Vertical averaged geostrophic speeds of the upper 2500 m directed normal to a line across the Drake Passage. The positions of the front and stratification zones are shown. The highest flows defining the axes of the Antarctic Circumpolar Current coincide with the position of the ocean fronts. (Reproduced with permission from Clifford, 1983.)

Drake Passage eddy field may not be typical of the full circumpolar belt. Meanders and eddies of the Antarctic Circumpolar Current also act to carry wind-delivered momentum downward to the seafloor, where pressure forces (often referred to as form drag) acting on the slopes of bottom topographic features act to compensate the force of the wind. The downward transfer of momentum is integrally linked to the meridional fluxes of heat and fresh water by the eddy field, by baroclinic instability.

Weddell Gyre The Weddell Gyre is the largest of the cyclonic Gyres occupying the region between the Antarctic Circumpolar Current and Antarctica, stretching from the Antarctic Peninsula to 301E. The clockwise flow pattern is linked to doming of isopycnals and upwelling of deep water within its central axis (Figure 3). As the deep water is warmer than the surface layer the Gyre injects heat into the surface layer,


CURRENT SYSTEMS IN THE SOUTHERN OCEAN 10°W

45°S

741

10°E

30°W

30°E 55°S

50°W

50°E 65°S

70°E

70°W

90°W

90°E

110°E

110°W 70°S 1987 1988 1989 1990 1991 1992 1993

130°E 55°S 150°E

150°W 170°W

45°S

170°E

Figure 6 Position of the Antarctic polar front for the years 1987–1993 as revealed by satellite images of sea surface temperature. (Reproduced with permission from Moore et al., 1999.)

which limits the winter sea ice cover to a thickness of only 0.5 m. Along the Antarctic margins intense cooling of surface water lead to the formation of the coldest, densest ocean water masses of global importance: Antarctic Bottom Water. A branch of the Antarctic Circumpolar Current turns southward near 301E forming the eastern limb of the Weddell Gyre. Some of this water turns westward along the coast of Antarctica (the rest continues to flow eastward, but at a more southern latitude than the axis of the Antarctic Circumpolar Current). Westward flowing coastal current is characteristic all around Antarctica, with a surface speed of about 10 cm s1 as detected by satellite tracking of the drift of icebergs calved from Antarctica. Within the Weddell Gyre the coastal current westward flow is blocked by the Antarctic Peninsula. Upon encountering the southern base of the peninsula, the coastal current turns northward, forming the western boundary current of the Weddell Gyre. At the northern tip of the Antarctic Peninsula, the western boundary current composed of the cold, low salinity stratification characteristic of Antarctic continental margin, is injected into the open ocean. This feature, called the Weddell-Scotia Confluence

(Figure 2) separates the Antarctic Circumpolar Current from the interior of the Weddell Gyre. It can be traced as a low salinity band to the Greenwich Meridian. Along the sea floor Antarctic Bottom Water escapes from the Gyre, flowing northward within deep crevices in the seafloor morphology, into the Scotia Sea, South Sandwich Trench and south of Africa. Export of Bottom Water is compensated by import of circumpolar water along the eastern boundary. Surface currents of the Weddell Gyre are weak, usually 10 cm s1, but the flow extends to the seafloor, as a strongly barotropic current. There is some evidence that the current increases along the seafloor of the continental slope, with speeds of up to 20 cm s1, associated with plumes of dense shelf water descending into the deep ocean as Antarctic Bottom Water. Observations of ocean currents during the period from 1989 to 1992 across the mouth of the Weddell Sea, stretching from Kapp Norvegia (711200 S; 111400 W) to the northern tip of the Antarctic Peninsula, find Gyre transport of about 30 million m3 s1 (30 Sv), most of which is contained in narrow jets following along the continental slope. An additional 10 million m3 s1 (10 Sv) of transport


742

CURRENT SYSTEMS IN THE SOUTHERN OCEAN Topex / Poseidon sea level standard deviation (m)

0.00

0.06

0.12

0.18

0.24

Figure 7 Variability of sea surface height from mean sea level, as revealed by Topex Poseidon satellite altimetric measurements of sea level for the period 1992–1999. Values are in meters. Variability of sea level is caused by meanders and eddies of the geostrophic flow field, or changes in its strength. (Provided by Donna Witter, Associated Research Scientist at Lamont-Doherty Earth Observatory.)

is likely around the central axis of the Gyre, making a total recirculation transport around the Gyre of 40 million m3 s1 (40 Sv). Export from the Gyre is not known exactly, but can be estimated as 5 million m3 s1 (5 Sv) within the bottom layer.

See also Agulhas Current. Antarctic Circumpolar Current. Atlantic Ocean Equatorial Currents. Icebergs. Indian Ocean Equatorial Currents.

Further Reading Belkin IM and Gordon AL (1996) Southern ocean fronts from the Greenwich Meridian to Tasmania. Journal of Geophysical Research 101(C2): 3675--3696. Clifford MA (1983) A Descriptive Study of the Zonation of the Antarctic Circumpolar Current and its Relation

to Wind Stress and Ice Cover. MS thesis, Texas A & M University. Gille S (1994) Mean sea surface height of the Antarctic Circumpolar Current from Geosat data: methods and application. Journal of Geophysical Research 99: 18 255--18 273. Gille S and Kelly K (1996) Scales of spatial and temporal variability in the Southern Ocean. Journal of Geophysical Research 101: 8759--8773. Gordon AL, Molinelli E, and Baker T (1978) Large-scale relative dynamic topography of the Southern Ocean. Journal of Geophysical Research 83: 3023--3032. Hofmann EE (1985) The large-scale horizontal structure of the Antarctic Circumpolar Current from FGGE drifters. Journal of Geophysical Research 90: 7087--7097. Nowlin W and Klinck J (1986) The physics of the Antarctic Circumpolar Current. Reviews of Geophysics 24(3): 469--491. Moore JK, Abbott M, and Richman J (1999) Location and dynamics of the Antarctic Polar Front from satellite sea surface temperature data. Journal of Geophysical Research 104(C2): 3059--3073.



Orsi AH, Nowlin WD, and Whitworth T (1993) On the circulation and stratification of the Weddell Gyre. Deep-Sea Research 40: 169--203. Orsi AH, Whitworth T, and Nowlin WD (1995) On the meridional extent and fronts of the Antarctic Circumpolar Current. Deep-Sea Research 42(5): 641--673.

743

Whitworth T (1980) Zonation and geostrophic flow of the Antarctic Circumpolar Current at Drake Passage. DeepSea Research 27: 497--507. Whitworth T and Nowlin WD (1987) Water masses and currents of the Southern Ocean at the Greenwich Meridian. Journal of Geophysical Research 92: 6462--6476.


CURRENT SYSTEMS IN THE MEDITERRANEAN SEA P. Malanotte-Rizzoli, Massachusetts Institute of Technology, Cambridge, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 1, pp 605–612, & 2001, Elsevier Ltd.

Introduction In the last two decades the Mediterranean Sea has been the object of renewed interest in the oceanographic community thanks to the formulation and execution of international collaborative programs, such as the UNESCO/IOC sponsored Programme de Recherche Internationale en Me´diterraneé Occidentale (PRIMO) in the Western basin and Physical Oceanography of the Eastern Mediterranean (POEM) Programme in the Eastern Mediterranean, followed up by the recent effort undertaken by the European community under the Marine Science and Technology (MAST) banner. Two main reasons form the basis of this scientific effort. The Mediterranean is a midlatitude semienclosed sea, exchanging water and other properties with the North Atlantic Ocean, which plays a crucial role in the global thermohaline circulation through the formation of North Atlantic Deep Water (NADW) in the polar Greenland and Labrador Seas. The upper intermediate layer of the North Atlantic is replenished with a very salty water mass that spreads out from the Mediterranean through the connecting narrow Straits of Gibraltar. This salty water is formed in the easternmost Mediterranean, the Levantine basin, as Levantine Intermediate Water (LIW). The Gibraltar Straits are thus a point source of heat and salt for the North Atlantic at all depths from 1000 m to more than 2500 m. Tongues of this warm, salty water extend through the Atlantic interior, both northward along the coast of Europe and westward towards America on all isopycnals from s1 ¼ 31.938 kg m3 to s3 ¼ 41.44 kg m3, thus crucially preconditioning the NADW convective cells in the polar seas. The second reason is that many dynamical processes which are fundamental to the world ocean circulation also occur within the Mediterranean, such as deep convection cells completely analogous to the NADW cell, with convective sites in both the Western and Eastern basins. Both of these basins are

744

moreover endowed with a deep thermohaline circulation, the equivalent of the global conveyor belt. Thus the Mediterranean provides a laboratory basin for general circulation studies. The two basins, Western and Eastern, can be studied quite independently, as they are connected through the shallow Sicily Straits, with the deepest threshold at B250 m, which prevents direct communication between the subsurface layers. This paper first provides a short review of the climatology characterizing the entire basin. The western and eastern basins are then discussed separately. Particular attention is devoted to the Eastern Mediterranean, where a major transient has occurred in the last decade that documents the existence of multiple states for the Eastern Mediterranean internal thermohaline circulation.

Morphology and Climatology The Mediterranean Sea (Figure 1A) is an enclosed basin connected to the Atlantic Ocean by the narrow and shallow Strait of Gibraltar (width B 13 km; sill depth B300 m). It is composed of two similar size basins, western and eastern, connected by the Strait of Sicily (width B35 km; sill depth B250 m). The Western Mediterranean has a triangular shape, with the Ligurian Sea at its apex and a large topographic plateau in the Balearic Sea (Figure 1B). The islands of Corsica and Sardinia separate the Balearic Sea from the Tyrrhenian Sea, where the bottom relief has the more complex shape of a deep, corrugated valley. The Balearic and Tyrrhenian Seas join in the south in a wide passage between Sardinia and Sicily that leads to the Sicily Straits and the eastern basin. The Eastern Mediterranean has a more complicated structure than the western, with a much more irregular, complex topography constituted by a succession of deep valleys, ridges, and localized pits (Figure 1B). Four sub-basins can be defined in the Eastern Mediterranean (Figure 1A): the Ionian, the Levantine, the Adriatic, and the Aegean Seas. The Ionian Sea is the deepest in its central part, ending in the shallow Gulf of Sirte at its southernmost end. The Cretan passage leads from the Ionian into the Levantine basin, that reaches its maximum depth in a localized depression south-east of the island of Rhodes.


CURRENT SYSTEMS IN THE MEDITERRANEAN SEA

745

45˚N Ligurian Sea

Gulf of Lions

ric

ria

tic

Corsica Tyr a

40˚

Se

Black Sea

Ad

rhe

Sardinia

nia

nS

Otranto

Sea of Marmara ea

Aegean Sea

ea

l

Ba

Bosporus

Dardanelles

St

rai t

Gibraltar

of

Sicily

Sic

ily

Ionian Sea

Alboran Sea

35˚

Rhodes Crete

Cyprus Levantine Sea

Gulf of Sirte

30˚N (A)

5 ˚W

0˚

15˚

5˚ E

25˚

30˚

35˚E

46˚N

38˚

30˚N 10˚W (B)

2 ˚E

14˚

26˚

38˚E

Figure 1 Morphology of the Mediterranean Sea: (A) geography of the basin; (B) bathymetry.

The hydrology and the circulation of the Mediterranean Sea have been known in overall generality for some time. For instance it is well known that the Mediterranean basins, both western and eastern, are evaporation basins (lagoons), with freshwater flux from the Atlantic through the Gibraltar Straits and into the Eastern Mediterranean through the Sicily Straits. The Atlantic Water (AW) mass entering through Gibraltar increases in density because evaporation exceeds precipitation and becomes Modified Atlantic Water (MAW) in its route to the Levantine basin. New water masses are formed here via convection events driven by intense local cooling and evaporation from winter storms. Bottom water

is produced in localized convection sites, for the western basin in the Gulf of Lions (Western Mediterranean Deep Water, WMDW) and for the eastern basin in the southern Adriatic (Eastern Mediterranean Deep Water, EMDW). Recent observations also indicate Levantine Deep Water (LDW) formation in the north-eastern Levantine basin during exceptionally cold winters, where Levantine Intermediate Water (LIW) is regularly formed seasonally. Evidence has emerged that LIW formation occurs over much of the Levantine basin but preferentially in the north probably due to meteorological factors. The LIW is the important water mass which circulates westward through both the eastern and western basins and


746

0 m


36.5 37.5

36.5

500

37.5 38.75

38.5

Sicilian Sill

Sardinian Sill

1000

Gilbraltar Sill

Alboran Sill

> 39.0

38.5

38.9 38.75

Malta Sill 38.7

38.4

< 38.65

1500 < 38.4 < 38.65

2000 Figure 2 The open thermohaline cell of the Mediterranean upper layer.

contributes predominantly to the efflux from Gibraltar to the Atlantic, mixed with some EMDW and WMDW. The western and eastern basins are connected in the upper layer, B200 m thick, through the ‘open thermohaline cell’ of the basin, schematized in Figure 2. While the entire Mediterranean is a ‘lagoon’ for the North Atlantic, the eastern basin itself is a ‘lagoon’ for the western one. The Atlantic Water (AW) mass enters the Mediterranean at Gibraltar with a typical salinity of S ¼ 36.15 PSU and temperature T ¼ 151C. The AW becomes Modified Atlantic Water (MAW) through diffusive processes in its pathway eastward and can be identified as a subsurface salinity minimum below B30 m depth. At the Sicily Straits, the saltier MAW has a salinity S r 37.5 PSU reaching a maximum of S o 38.9 PSU in the Levantine basin. Here in its northern part, winter episodes of cold, dry winds blowing from the mainland under surface cooling and evaporative fluxes lead to the formation of LIW, which has 39.0 r S r 39.2 PSU and 151C o y o 161C at the formation sites, the most important of which is the well-known Rhodes gyre. The LIW return route is westward in the layer between 200 and 600 m depth. LIW becomes progressively colder and fresher. At the Sicily Straits typical LIW core values are y ¼ 14.31C and S r 38.8 PSU. At the Gibraltar Straits LIW is diluted to S r 38.5 PSU, and spreads out in the North Atlantic, becoming Upper North Atlantic Intermediate Water. Secondary pathways of LIW will be discussed separately for the two basins.

The Western Mediterranean The upper thermocline circulation (upper B200 m) of the Western basin is schematized in Figure 3A. The MAW in the Alboran Sea describes a quasipermanent anticyclonic gyre in the west and a more variable circuit in the eastern Alboran. Further east, the MAW is entrained in the strong meandering

Algerian current, whose instabilities lead to the formation of anticyclonic eddies (diameter B50–100 km) all along the coast of Algeria. These eddies grow in size; some may detach from the coast and drift into the interior of the Balearic Sea. A quasi-stationary cyclonic path of MAW has been observed around the Balearic Sea leading to the formation of the Western Corsican Current west of Corsica. A steady cyclonic path of MAW is also present in the Tyrrhenian Sea, that intrudes into Northern Ligurian Sea, where it joins the Western Corsican Current producing a return south-westward flow along the Italian, French, and Spanish coasts, towards the Alboran Sea, that is called the Northern Current. The latter shows strong seasonal variability, becoming more intense and narrower in wintertime when it develops intense meanders, and splits into multiple branches in the southern Balearic sea. Figure 3B shows the pathway of LIW emerging from the Sicily Straits into the Western Mediterranean in the intermediate layer, 200–600 in depth. LIW follows a cyclonic route all around the Tyrrhenian Sea, and splits into two branches at the northern tip of Corsica. One branch enters directly into the Ligurian Sea, the second circulates around Sardinia and Corsica, merges with the previous branch and successively flows cyclonically around the Balearic Sea. This major LIW branch enters the Gulf of Lions, where it plays a crucial role in preconditioning the winter convective cell of WMDW located here. WMDW has been observed to form in the Gulf of Lions basically every year, under winter episodes of cold, dry Mistral wind blowing from France. Here the mixed, ventilating chimney (B100 km in diameter) can reach 2000 m depth. It must be pointed out that the mean LIW pathway is still controversial. Numerical simulations as well as data analysis indicate a major direct route of LIW from the Sicily Straits to Gibraltar. On the other hand, strong observational evidence suggests the pattern presented in Figure 3B, with the LIW cyclonic circuit around the Tyrrhenian, the islands of



747

40

(A)

_5

35 0

5

10

15

40

(B)

_5

35 0

5

10

15

40

(C)

_5

35 0

5

10

15

Figure 3 Circulation of the Western Mediterranean: (A) upper thermocline circulation; (B) intermediate layer circulation with LIW pathways; (C) deep thermohaline circulation with WMDW pathways.


748


Sardinia and Corsica and finally the Balearic Sea. Here, in the southern, eastern part, the LIW pathway bifurcates, with one branch proceeding towards and exiting from Gibraltar and a second returning eastward along the Algerian coast. Figure 3C schematizes the thermohaline cell of the Western Mediterranean, driven by the winter convection in the Gulf of Lions. Even though the exact routes of WMDW are also under debate, the pattern of Figure 3C is based on available observations and indicates spreading at depth of WMDW both towards the Sicily and the Gibraltar Straits, following a circuitous cyclonic route that leads it throughout the Balearic and Tyrrhenian Seas. The deep WMDW flow is obviously affected by topography. In the Tyrrhenian Sea, the WMDW joins the Tyrrhenian Dense Water present in the deep layers. At Gibraltar, upwelling of WMDW occurs, mixing with the overlying LIW, and contributing (what it is believed to be a small proportion) to the outflow from Gibraltar into the northern Atlantic.

established the existence of three dominant scales interacting in the general circulation pattern; the basin-scale, i.e. the intermediate and deep thermohaline circulation; the sub-basin scale, characterizing the upper thermocline; and the mesoscale, defined by a ubiquitous and energetic eddy field. The POEM observational evidence has moreover shown that the Eastern Mediterranean has undergone a startling transition in the intermediate and deep basin circulations between the 1980s and the 1990s, the first documented example of the existence of multiple thermohaline states. The upper thermocline circulation, on the other side, which is embedded in the Mediterranean open thermohaline cell, has remained very consistent throughout the two decades. The building blocks of the upper thermocline circulation are sub-basin-scale gyres and permanent, or quasi-permanent, cyclonic and anticyclonic structures interconnected by intense jets and meandering currents. The schematic representation of the upper thermocline circulation is given in Figure 4. All the structures depicted in Figure 4 are robust and persisted in the 1980s and the 1990s, albeit with modulations in strength and areal extension. Some differences were present but only in the Levantine Sea. At the Sicily Straits, the entering MAW is advected by the strong Atlantic Ionian Stream (AIS)

The Eastern Mediterranean The field work for the POEM program which was carried out in the period 1985–95, has definitively 12˚E

16˚E

20˚E

24˚E

28˚E

32˚E

36˚E

ASW 40˚

38˚ MAW AIS ISW 36˚

IA

MIJ

C AM

LSW PA LS

MAW

CC

MMJ 32˚

Rh Iérapetra

34˚

es

od

W

CC

W.rus p Cy

J

MM

Shikmona

MersaMatruh

Figure 4 Eastern Mediterranean: Schematic representation of the upper thermocline circulation. AIS, Atlantic Ionian Stream; AMC, Asia Minor Current; ASW, Adriatic Surface Water; CC, Cretan Cyclone; IA; lonian anticyclones; ISW, lonian Surface Water; LSW, Levantine Surface Water; MAW, Modified Atlantic Water; MIJ, Mid-Ionian Jet; MMJ, Mid-Mediterranean Jet; PA, Pelops anticyclone.



jet, which forms a broad meander in the Ionian Sea, bifurcating into two main branches. One branch turns directly southward towards the African coast enclosing an intense anticyclonic area with multiple centers, the Ionian anticyclones (IA) which penetrate deeply into the intermediate layer. The second AIS

749

branch protrudes into the north-eastern extremity, then turns southward forming the strong MidIonian Jet (MIJ) that crosses the entire Ionian sea meridionally, thereafter veering eastward through the Cretan channel where it becomes the Mid-Mediterranean Jet (MMJ). Strong permanent features are

42

38

38.78

38.

7

40

.7 2

38

00

39.

8

8

39.0

38.9

.9

0

38.9

08

39.

92

. 38

38

38

2

.88

.88

38

8.8

38

.84

38

8

Latitude N

.86

.7 38

6 .7

38

38

36

.96

34

39.1

2

39.00

92

38. 38.88

32

30 10

12

14

16

18

20

26 28 24 Longitude E _ Salinity at density = 29.05 kg m 3

(A)

22

32

30

34

36

38

36

38

42

40 .74

38 38.94

34

39.

08

39.94

39.06

39.1 0

10

.98

38

39.0 2

39.

.86 38

38.8

4

38.

39.06

.82

Latitude N

.86

38

82

38

36

38.8

0

38

39.02

0 .0

39

39.00

32

30 10

(B)

12

14

16

18

20

22


30

32

34

Figure 5 Intermediate layer circulation: (A) circulation in 1987 with LIW pathways; (B) circulation in 1991 with LIW and CIW pathways.


750


the Pelops anticyclone (PA) that has a strong barotropic component and penetrates to 800–1000 m depth. The Cretan cyclone, located south of Crete, is on the other side confined to the upper thermocline. A further permanent structure in the Cretan passage is the strong Ierapetra anticyclone, also South of Crete. The MMJ from the Cretan passage intrudes into the Eastern Levantine where it separates a northern overall cyclonic region from a southern anticyclonic region. The northern region comprises two well defined, permanent cyclones, the Rhodes gyre, site of

LIW and LDW formation, and the western Cyprus cyclone. The southern anticyclonic area also comprises multiple centers, the strongest and most robust of which is the Mersa-Matruh anticyclone, located just south of the Rhodes gyre. A quasi-permanent structure, the Shikmona anticyclone, is present in the easternmost Levantine. In the 1990s, the only major difference was constituted by the appearance of a third anticyclonic center in the southern Levantine, of which the MMJ constituted the Northern rim. This anticyclone pushed westward the Mersa-Matruh

Aegean

Adriatic

(A)

Ionian

Black

Levantine

42

78

38.86

0

38.82

84

38.80

38

.8

34

38

38.

38.

38.76

Latitude N

38.72 38.72

36

38.84

38.74

38

38.74

40

38.76

2

38.74

38.7

32

30 10

(B)

12

14

16

18

20

22


30

32

34

36

38

Figure 6 Deep thermohaline circulation: (A) schematic representation of the deep thermohaline cell in 1987; (B) deep pathways of CDW in 1991.



anticyclone, thus forming with the Ierapetra gyre a three-lobed intense anticyclonic area that induced a stronger meandering in the MMJ, confining it to its northern rim. The dramatic transition occurred in the intermediate and deep layer circulations. Definitive observational evidence has been presented that this change was already present in 1991 and persisted through 1995–96, while the 1985–87 situation was completely different. In the intermediate layer, 250–600 dbar depth, characterized by LIW spreading on the isopycnal horizons sy ¼ 29.00–29.05– 29.10 kg m3, the 1987 LIW circulation is depicted in Figure 5A for sy ¼ 29.05 kg m3. LIW, formed in the northern Levantine in the Rhodes gyre, follows its ‘classical’ pathway, spreading towards the Sicily Straits through the Cretan channel and the Ionian interior. A second major route produced by the veering by the Pelops anticyclone is northward along the Greek coastline towards the Otranto Straits. Here LIW enters the southern Adriatic Sea to precondition the deep convective cell where Adriatic Deep Water (ADW) occurs. ADW spreads at depth out of the Otranto Strait to become EMDW. The situation in 1991 (shown in Figure 5B, again the salinity distribution on sy ¼ 29.05 to kg m3) is completely different. Now Cretan Intermediate Water (CIW) formed inside the Cretan/Aegean sea, substitutes for LIW, exiting from the western Cretan Arc Straits in a well defined tongue and filling the Ionian interior. The LIW is still formed in the northern Levantine, but its westbound pathway is blocked by the three-lobed anticyclonic region now present in the southern Levantine, which induces a local LIW cyclonic recirculation inside the Levantine itself. This startling change is due to the fact that in the 1990s the ‘driving engine’ of the intermediate, transitional and deep layer circulations became the interior of the Cretan/Aegean Sea, with CIW and Cretan Deep Water (CDW) forming there, spreading out and filling the abyssal layers of the entire Eastern Mediterranean. In 1987, the EMDW was formed in the southern Adriatic as ADW, spread out from the Otranto Strait into the entire Eastern Mediterranean, upwelled to the transitional/intermediate layers and returned as LIW in the upper warm pathway to the southern Adriatic, thus closing the internal

751

thermohaline cell. The 1987 ‘conveyor belt’ of the basin is schematized in Figure 6A. In 1991 the transitional and deep water masses were also formed in the Cretan/Aegean Sea, and they spread out from the western and eastern Cretan Arc Straits on all the horizons sy ¼ Z29.15 kg m3, as shown in Figure 6B These denser isopycnals rose to much shallower depths in 1991 than in 1987, thus greatly increasing by advection the salt content of the intermediate layer. In the Ionian Sea, CDW pushes the old and slightly denser EMDW of southern Adriatic origin to the west and downward to the near bottom layer. However, the closing pathways of the Eastern Mediterranean deep thermohaline cell in the 1990s are not yet clearly identified.

See also Mediterranean Sea Circulation.

Further Reading Klein B, Roether W, and Manca BB (1999) The large deep water transient in the Eastern Mediterranean. Deep-Sea Research I 46: 371--414. Malanotte-Rizzoli P and Robinson AR (eds.) (1994) Ocean Processes in Climate Dynamics: Global and Mediterranean ExamplesN: ATO-ASI Series C, vol. 419. Kluwer Academic Publisher Malanotte-Rizzoli P, Manca BB, and Ribera d’Alcala M (1997) A synthesis of the Ionian Sea hydrography, circulation and water mass pathways during POEMPhase I. Progress in Oceanography 39: 153--204. Malanotte-Rizzoli P, Manca BB, and Ribera d’Alcala M (1999) The Eastern Mediterranean in the 80s and in the 90s: the big transition in the intermediate and deep circulations. Dynamics of Atmospheres and Oceans 29: 365--395. Millot C (1999) Circulation in the Western Mediterranean. Journal of Marine Systems Special volume 20: 423--442. POEM Group (1992) The general circulation of the Eastern Mediterranean. Earth Science Review 32: 285--309. Reid JJ (1994) On the total geostrophic circulation of the North Atlantic oceanf: low patterns, tracers, and transports. Progress in Oceanography 33: 1--92. Robinson AR and Malanotte-Rizzoli P (1993) Physical Oceanography of the Eastern Mediterranean. Deep-Sea Research (Special Issue) 40: 1073--1332.


Encyclopedia of Ocean Sciences, Second Edition, Volume 1: A-C

Encyclopedia of Ocean Sciences, Second Edition, Volume 1-6

Encyclopedia of Ocean Sciences, Second Edition, Volume 3: G-M

Encyclopedia of Ocean Sciences, Second Edition, Volume 5: S

Encyclopedia of Ocean Sciences, Second Edition, Volume 6: T-Z

Encyclopedia of Ocean Sciences, Second Edition, Volume 2: D-F