Encyclopedia of Ocean Sciences,

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

530 downloads 4663 Views 38MB 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 4: N - R 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


NEKTON W. G. Pearcy, Oregon State University, Corvallis, OR, USA R. D. Brodeur, Northwest Fisheries Science Center, Newport, OR, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Marine nekton are the swimmers in the sea, as opposed to plankton that drift with currents. Although the main criterion is swimming ability, most definitions of nekton, as with plankton, are not precise. They are based on swimming speeds, lengths, Reynolds number, shape to reduce drag, or simply those animals too swift or agile to be captured in plankton nets. Nekton include fishes, cephalopods, seabirds, marine mammals, reptiles, and crustaceans. They inhabit all the ecological zones of the ocean, from the epipelagic near-surface waters to the deep-sea abyss, including both pelagic and epibenthic animals. Methods of assessing the distribution and abundances of nekton include nets, baited hooks or traps, acoustics, visual systems, lasers, egg/larval surveys, tagging/recapture, and food habits of their predators. Because this edition includes articles on the biology of fishes, fisheries, birds, and marine mammals, the emphasis here will be on the smaller nekton, or micronekton (usually 2–20 cm in length). Although most micronekton are not harvested, krill (euphausiids) and lanternfishes are sometimes fished commercially. Some problems for nektonic animals in the sea include overcoming the forces of drag while moving through a viscous medium, counteracting sinking through buoyancy adaptations, feeding, and avoiding being eaten. The reader is referred to articles on biology of fishes that address some of these subjects. The major taxonomic groups of micronekton are fishes, cephalopods, and large crustaceans. They are important components of marine ecosystems and vital links between zooplankton and higher trophic levels. Many are carnivorous; some are herbivores. As a group they are important to us as food, either directly through fisheries or indirectly as food for commercially important species. Fishes, squids, and euphausiids are important prey for seabirds, larger fishes, and marine mammals. Oceanic and coastal micronekton are important, but poorly understood,

intermediate links in the food webs from zooplankton to apex predators for many marine ecosystems including the tropical Pacific, the Gulf of Alaska, and the Barents Sea (Figure 1). Euphausiids are especially important as prey for many species of fishes, seabirds, and marine mammals. In the Antarctic, krill (mainly Euphausia superba; Figure 2) are a major food for whales, penguins, and some pinnipeds. Although most euphausiids are herbivorous, most of the micronektonic fishes and squids are carnivorous, preying on euphausiids, copepods, amphipods, and other zooplankton. Distributions

Many studies have shown that the distributions of mesopelagic micronekton, like zooplankton, generally coincide with specific water masses in the ocean, often forming faunal assemblages. However, some species are found in all oceans while others have distributions restricted to waters of the continental shelf or slope. Species richness of the micronektonic fauna usually increases from high to low latitudes. The biology and morphology of micronekton differ among the different pelagic realms: epipelagic (0–200 m), mesopelagic (200–1000 m), bathy/abyssopelagic (41000 m), and epibenthic. Epipelagic micronekton include the juvenile phases of many species, as well as the mature stages of small fishes such as Engraulidae (anchovies), Clupeidae (herrings, sardines), and Osmeridae (smelts) (see Figure 1). These epipelagic species are most common over productive regions of the continental shelves and coastal upwelling regions where they are often subjected to commercial fisheries.

Midwater Micronekton Many families of fishes, squids, and crustaceans are considered micronekton (Figure 3). Myctophidae (lanternfishes) and Gonostomatidae (bristlemouths) are common families of mesopelagic fishes. It has been estimated that the worldwide biomass of mesopelagic fishes alone is over 900 million metric tons. Many of these animals have distributions in oceanic (waters beyond the continental shelves) in all the world’s oceans. Generally, mesopelagic micronekton are most abundant over the continental slopes and their numbers decrease both inshore and offshore. Diel migrations also provide an abundance of food resource for shallow-water or coastal fishes.


1

2

NEKTON

Tropical Pacific Primary producers

Zooplankton

Micronekton

Macronekton

Apex predators

Sea turtles Meroplankton

Epipelagic

Seabirds Marlin and sailfish

Euphausids Amphipods

Mesopelagic migrant

Copepods

Mesopelagic nonmigrant

Skipjack

Dolphins

Phytoplankton

DOM

Chaetognath Salps

Picoplankton Detritus

Yellow fin

Young tuna and scombrids

Bathypelagic migrant

Piscivorous fish

Bathypelagic nonmigrant

Large squids

Blue shark Other sharks Bigeye

Microzooplankton Larvaceans

Albacore

Swordfish Killer whales

Gulf of Alaska Euphausiids

Pollock Eulachon Pacificcod

Copepods Capelin Chaetognath Phytoplankton

Other

Arrowtooth flounder Halibuth

Pteropods

Purpoises Toothed whales Baleen whales Seabirds

Rockfish Juvenile gadids Larvaceans

Steller sea lion Other

Shrimps

Harbor seals

Flathead sole

Other

Barents Sea

Euphausiids

Herring

Cod

Seabirds

Capelin Copepods Phytoplankton

Polar cod

Other gadids

Baleen whales

Amphipods Juv. gadids Pinnipeds Other

Shrimps

Fatfish

Figure 1 Simplified food web for tropical Pacific, Gulf of Alaska, and Barents Sea showing the intermediate position of micronekton in the food webs. Adapted from Ciannelli L, Hjermann DO, Lehodey P, Ottersen G, Duffy-Anderson JT, and Stenseth NC (2005) Climate forcing, food web structure, and community dynamics in pelagic marine ecosystems. In: Belgrano A, Scharler UM, Dunne J, and Ulanowicz RE (eds.) Aquatic Food Webs, an Ecosystem Approach, pp. 143–169. Oxford, UK: Oxford University Press.


NEKTON

Figure 2 Euphausiid swarm. Photo courtesy of Bruce Robison, Monterey Bay Aquarium Research Institute (MBARI).

Off Oregon, rockfishes (Sebastes spp.) often prey on euphausiids and sergestid shrimp and myctophid fishes that are carried by currents inshore and then trapped against the shallow seafloor or offshore banks or seamounts during daytime. Off Hawaii, the animals forming layers are impinged along the sides or on the tops of seamounts where they are easy prey (Figure 4). In Hawaiian coastal waters, a mesopelagic boundary layer community consists of a distinct resident community of micronekton distributed along a narrow band where the upper slope meets the oceanic realm. Based on acoustical measurements from moorings, it exhibits both diel vertical and horizontal migrations. Remarkable inshore migrations at night extend within 1 km from shore. Again, this provides a trophic link between neritic and oceanic systems. Diel Vertical Migrations

Many species of epipelagic and mesopelagic micronekton undertake diel, nocturnal, or daily vertical migrations, swimming toward the surface at night and descending into deeper water during periods of daylight. This common and unique behavior of mesopelagic micronekton is shown in Figure 4, where a distinct sound-scattering layer migrates toward the surface during twilight, while other layers do not migrate. Interestingly, not all individuals migrate even within the same species. Some are basically nonmigratory. Some studies have observed diel migrations of sound scatterers between the depths of 500 and 1000 m. However, there is little evidence for migrations of animals living below about 1000 m in the open ocean at bathy/abyssopelagic depths. These diel migrations were first observed in the open ocean by physicists who were perplexed by

3

sound-scattering layers at mid-depths during the day that looked like false bottoms, and often moved toward the surface at night. They called them ‘deepscattering layers’. Different animals reflect sound depending on the frequency of sound used and the sound velocity and density contrast of the animals. We know that the animals that reflect the sound (10–50 kHz) in these layers are usually fishes or siphonophores, often with gas-filled swimbladders or floats that effectively resonate sound. In the open ocean, migrations are generally from mesopelagic depths (200–1000 m) into epipelagic waters, or even to the surface, at night. Because of the huge biomass involved, these migrations of micronekton account for the largest daily movement of biomass on Earth (except for the masses of Homo sapiens that commute daily!). Sunlight is the primary sensory cue that initiates or guides diel vertical migrations. This conclusion is supported by the conformity of the vertical movements of scattering layers and specific isolumes or light intensities, as well as responses to artificial light or solar eclipses. However, endogenous rhythms or changes of light intensity have also been suggested as triggers to initiate migrations. Adaptive significance of diel vertical migrations Why migrate vertically? The main theories concern the predator following prey and the latter avoiding predation. Food is available in greater quantity in shallower waters that are rich with zooplankton but these sunlit waters are dangerous during the daytime because of large visual predators. Hence this behavior is thought to be an energetic trade-off between the risk of predation and feeding. Vulnerability to large visual predators is reduced at night in near-surface water and in deep, dimly lit water during the daytime. Other theories involve an energy bonus for animals feeding in warmer nearsurface waters and digesting and assimilating food in deeper cooler waters. Mid-water fishes often migrate vertically in the Antarctic even when the water column has near-uniform temperature and their major prey, E. superba, remains near the surface and does not migrate. This suggests that avoidance of visual predators is a major selection factor for diel vertical migrations in these waters. These ubiquitous vertical migrations in the open ocean contribute to the vertical transport of energy and elements. Carbon in zooplankton, as well as anthropogenic materials such as pesticides or radioisotopes, are ingested by micronekton in near-surface waters and are defecated in deeper water. This rapid transport of materials has been called a ‘biological pump’.


4

NEKTON

Figure 3 Photos of in situ mesopelagic fishes, squids, and shrimp taken from a remotely operated vehicle (ROV) Tiburon. All images & MBARI. From top, left to right: lanternfish Stenobrachius leucopsarus, bristlemouth Cyclothone signata, viperfish Chauliodus macouni, squids Galiteuthis, Gonatidae, and shrimp Sergestes similis. All images & MBARI.

Bioluminescence

Many species of oceanic micronekton, especially in the meso/bathypelagic regions, are bioluminescent. These include most families of euphausiids, many pelagic mysids, shrimps, fishes, and squids. Bioluminescence is produced by simple or complex light

organs, photophores, or luminous secretions, either by symbiotic bacteria or intrinsic organs under nervous control. The color of most bioluminescence in mesopelagic animals is blue-green, c. 470–490 nm. Notable exceptions are some stomiatoid fishes that emit a red light, a frequency that is invisible to most


NEKTON

Depth (m)

Migratory layer ≈ 120 m

9 Oct. 2004

0

5

250

500

Daytime layer ≈ 550 m

750

18:10

18:37

19:04 Time (local)

19:31

Nonmigratory layer ≈ 550 m

Figure 4 A 38-kHz echogram off Hawaii showing the complexity of sound-scattering layers, with a layer migrating toward the surface at twilight and nonmigratory layers near the surface and along the flanks of the seamount.

deep-water nekton – it functions as an underwater snooperscope. Many of the mesopelagic fishes, squids, and crustaceans have photophores arrayed along the ventral portions of their bodies (see fishes in Figure 3). This is thought to provide a ventral ‘countershading’ or counter-illumination that reduces the silhouette of their bodies from predators that lurk below. This theory is supported by experiments that have shown that the intensity of light emitted by the photophores closely matches downwelling light from above. Other bioluminescent organs are thought to be lures to attract prey or to promote schooling or mate recognition. Bioluminescence occurs in the ocean, even down to the deepest depths. If a light meter is lowered to measure light intensity, downwelling light from the sun decreases exponentially, but then light intensity may increase with depth at mid-depths from flashes of bioluminescent organisms. Because bioluminescence is the predominant source of light below 500–1000 m, fishes and crustaceans in the deep ocean usually have either jet-black (fishes) or crimson (crustaceans) pigments, presumably to absorb bioluminescent flashes. Here silvery or reflective bodies, as found in most epipelagic fishes, would reflect these flashes and stand out as mirrors that would attract predators. Life History and Biology

In the large volume that comprises the mid-water domain of the world’s oceans, finding and recognizing mates to reproduce with can be difficult, especially when overall abundances are low. Many fishes depend on random encounters of males and females, perhaps recognizing their own kind by their

photophore patterns. Some male anglerfishes have evolved a bizarre behavior where they actually attach as a dwarf form to a female once they locate them. They then draw nutrition from the female and continue in this parasitic relationship for the rest of their lives. Reproductive patterns are not well known in many mid-water nekton. Myctophids are known to spawn year-round in the tropics but spawn during a single season at higher latitudes. Many myctophids undertake extended seasonal north–south migrations between feeding and spawning grounds, as seen in epipelagic fishes. Most larvae and early juveniles occur in the epipelagic zone, are often nonmigratory, and are found progressively deeper as they grow older, metamorphosing into the adult stage in the mesopelagic zone. They usually mature after 1–3 years but their typical life spans are not generally known. With the exception of Antarctic krill, which can live to be about 7 years old, most nektonic crustaceans and squid probably live only a few years.

Food Habits Feeding modes in midwater nekton can range from filter-feeding to highly carnivorous. Most deep-water animals are carnivorous (note the viperfish in Figure 5). Numerous morphological or behavioral adaptations have evolved to overcome the paucity of food in the deep ocean. Bioluminescence is used to both locate prey, as discussed previously, and attract prey. Anglerfishes often use lighted ‘lures’ suspended over the mouth by fin rays modified to work as ‘fishing poles’ to attract gullible prey. Most species have upward-facing eyes and mouths to locate overhead prey that may be backlit against the


6

NEKTON

Figure 5 Head-on view of the viperfish, Chauliodus macouni. Image & 2006 MBARI.

Figure 6 Whiptail gulper, Saccopharynx. Note large mouth and extended stomach.

downwelling light. In many cases, the jaws are unhinged or extremely large relative to the size of the organism to swallow large prey. In an extreme example, the whiptail gulper fish (Figure 6) possesses a mouth and stomach that are highly extensible to consume prey bigger than themselves similar to snakes, and then gradually digest the prey item until the next meal comes along, perhaps months later. Given the large number of piscivorous predators in these regions, it is not surprising that many micronekton have developed adaptations for predator avoidance. Their bioluminescent countershading and light absorptive pigmentation reduce their visibility to most predators. Bioluminescent flashes may also confuse predators. Some animals align themselves vertically in the water column to minimize their

cross-sectional area visible to predators from below. The squid Galiteuthis (see Figure 3), which has a transparent body, often holds its opaque arms and tentacles directly over the large downward-oriented light organs on its head, presumably to minimize the silhouette of its arms and tentacles by counterillumination. Other elongate fishes curl up into a ball to mimic other unpalatable species or appear larger than they actually are to a predator. Squids have developed a different strategy in that they can release a black or luminous ink that may temporarily confuse their predators or act as ‘decoys’, allowing their getaway from predators.

See also Cephalopods.Crustacean Fisheries

Further Reading Benoit-Bird KJ and Au WWL (2006) Extreme diel horizontal migrations by a tropical nearshore resident micronekton community. Marine Ecology Progress Series 319: 1--14. Brodeur RD and Yamamura O (2005) Micronekton of the North Pacific. PICES Scientific Report 30: 115pp. Ciannelli L, Hjermann DO, Lehodey P, Ottersen G, DuffyAnderson JT, and Stenseth NC (2005) Climate forcing, food web structure, and community dynamics in pelagic


NEKTON

marine ecosystems. In: Belgrano A, Scharler UM, Dunne J and Ulanowicz RE (eds.) Aquatic Food Webs, an Ecosystem Approach, pp. 143--169. Oxford, UK: Oxford University Press. Marshall NB (1971) Explorations in the Life of Fishes. Cambridge, MA: Harvard University Press.

7

Nixon M and Messenger JB (eds.) (1977) Symposia of the Zoological Society of London, No. 38: The Biology of Cephalopods. New York: Academic Press. Robison BH (2003) What drives the diel vertical migrations of Antarctic midwater fishes? Journal of the Marine Biological Association of UK 83: 639--642.


NEPHELOID LAYERS I. N. McCave, University of Cambridge, Cambridge, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction A remarkable feature of the lower water column in most deep parts of the World Ocean is a large increase in light scattering and attenuation conferred by the presence of increased amounts of particulate material. This part of the water column is termed the bottom nepheloid layer (BNL). Another class of nepheloid layers found especially at continental margins are intermediate nepheloid layers (INLs) (Figures 1 and 2). These occur frequently at high levels off the upper continental slope and at the depth of the shelf edge. From here, they spread out across the continental margin. These INLs are similar to the inversions observed in the BNL on some profiles (Figure 1). (The surface nepheloid layer (SNL), not treated here, is simply the upper ocean layer in which particles are produced by biological activity, and which may have material from river plumes close to shore.) The increase in light scattering is perceived relative to minimum values found at mid-water depths of 2000–4000 m (shallower on continental margins). The increased scattering is due to fine particles. This has been determined by particle-size measurements and filtration of seawater with determination of concentration by weight and volume. Most data on the distribution and character of nepheloid layers have been acquired by optical techniques, principally by the Lamont photographic nephelometer and the SeaTech transmissometer, and more recently the WetLabs transmissometer and light scattering sensor (LSS). The optical work has revealed that the BNL is up to 2000-m thick (can be more in trenches) and generally has a basal uniform region, the bottom mixed nepheloid layer (BMNL), corresponding quite closely to the bottom mixed layer defined by uniform potential temperature (Figure 1). Above the BMNL there is a more or less exponential fall-off in intensity of light scattering up to the clear-water minimum marking the top of the BNL. Both bottom and INLs are principally produced by resuspension of bottom sediments. Their distribution indicates the dispersal of resuspended sediment in the ocean basins and is thus a signature of both material and water transport away from boundaries (some of

8

which may be internal such as ridges and seamounts). Most concentrated nepheloid layers occur on the continental shelf, upper slope, or deep continental margin. They indicate the locus of active resuspension and redeposition by strong bottom currents and internal waves.

Optics of Nephelometers: What They ‘See’ Detection of deep-ocean nepheloid layers has been mainly through measurement of light scattering. The Lamont nephelometer has made the largest number of profiles in all oceans but is no longer in use. It used an incandescent bulb as the source and photographic film as the detector of the light scattered from angles between y ¼ 81 and 241 from the forward axis of the light beam. The film was continuously wound on as the instrument was lowered, resulting in an averaging of the received signal over about 25-m depth. The short-lived Geochemical Ocean Section Study (GEOSECS) nephelometer used a red (l ¼ 633 nm) laser source and a photoelectric cell to detect light scattered from y ¼ 3–151 off the axis of the beam. The SeaTech (now WetLabs, Inc.) LSS measures infrared light (880 nm) backscattered (1801) from particles in the sample volume using a solar-blind silicon detector. Because of multiple scattering, nephelometers do not yield precise optical parameters. Optical transmission with a narrow beam can yield the attenuation coefficient (c). The most commonly used SeaTech and WetLabs transmissometers have a red light source (l ¼ 660 or 670 nm) and usually a 0.20– 0.25-m path length. Most of the contribution to the total scattering b comes from near-forward angles (low values of y). Jerlov (1976) shows that, for surface waters, 47% of b occurs between y ¼ 01 and 31, 79% between 01 and 151, and 90% between 01 and 301. The GEOSECS instrument records about 32% of b and the Lamont nephelometer about 16%. The total scattering is given by Mie theory (assumed spherical particles) as a function of particle size d and relative refractive index (relative RI) n, and wavelength of light l. The relative indices of refraction of suspended material are dominated by components with n ¼ 1.05 and 1.15, values probably characteristic of organic and mineral matter, respectively (e.g., RI of seawater 1.34, quartz 1.55, ratio n ¼ 1.15). Particles from clear ocean waters and weak nepheloid layers (concentration Co40 mg m3) tend


NEPHELOID LAYERS

9

c 1.0

0.8

0.6

0.4

1.80

1.85

41 24

4700

300

88 16

109

4800

200

179

Pressure (dbar)

215

Meters above bottom

157

234

252

4900

100

227

BMNL 5000 2

4

8 μm

16

32

0

Figure 1 Data from the SeaTech transmissometer in the Atlantic showing the BMNL and a nepheloid layer comprising multiple steps in temperature and turbidity. Turbidity is given as c the attenuation coefficient, and y is potential temperature in 1C. Also shown on the right are particle-size spectra determined by Coulter counter. Reproduced from McCave IN (1983) Particulate size spectra, behavior and origin of nepheloid layers over the Nova Scotian Continental Rise. Journal of Geophysical Research 88: 7647–7666.

to have particle-size distributions by volume which are flat, equivalent to k ¼ 3 in a particle number distribution of Junge type, N ¼ Kd k where N is the cumulative number of particles larger than diameter d and K and k are constants. However, this distribution does not appear to be maintained at sizes finer than about 2 mm where k decreases toward 1.5. In concentrated nepheloid layers, this distribution does not occur at all and a peaked distribution with a peak between 3 and 10 mm is encountered.

Morel has calculated scattering according to Mie theory for suspensions with Junge distributions and several indices of refraction. Recalculation into cumulative curves of percentage scattering in Figure 3 illustrates the fact that most recorded scattering is produced by fine particles. The cases shown are for values of k of 2.1, 3.2, and 4.0 and a two-component (peaked) distribution with k ¼ 2.1 up to a ¼ pdn/ l ¼ 32 and k ¼ 4.0 for larger sizes (a ¼ 32 is equivalent to d ¼ 5.6 mm for l ¼ 633 nm). In each case, three


10

NEPHELOID LAYERS

(a)

0 Knorr 51 sta. 698 Rockall Trough 54° 23.1′ N, 15° 18.7′ W INL.

Depth (m)

1000

2000 BNL 900m

{

INV 2 INV 1

Nephels, arbitrary scale 3000 (b)

3.0

2.8

2.6

3.2

2200 Nephel inversion 2

Higher-salinity core

2300 NEADW

Depth (m)

2400

2500

Nephel inversion 1 Knorr 51 Station 698 Rockall Trough 54° 28.1′ N 15° 18.7′ W 24 Aug. 1975

2600

AABW Source _ High silica_

2700 40

50

60

70

80

15

90

20

25

30

35

Silica (μg at/l)

Ne

Figure 2 (a) Full-depth profile taken in the Rockall Trough with the GEOSECS nephelometer. An INL and two inversions are apparent. (b) Detail of the lower 500 m of the profile in (a) showing the relationship between the nephel (turbidity) inversions and hydrography. Reproduced from McCave IN (1986) Local and global aspects of the bottom nepheloid layers in the world ocean. Netherlands Journal of Sea Research 20: 167–181.

forward-scattering angles are given. It is clear that scattering close to the beam is more sensitive to large particles than that at 201. (At yo0.51 we are essentially dealing with a transmissometer.) In the case of k ¼ 2.1 only about 22% of the scattering is from

smaller sizes (ao32) at y ¼ 20 1. Thus the curves for y ¼ 10 1 and 20 1 are generally representative, and the distribution for both k ¼ 3.2 and the composite case show that 95% of the scattering is by particles o5 mm for l ¼ 633 nm.


NEPHELOID LAYERS

16

32

64

128 256

4

8

16

11

32

100 k = 3.2

Cumulative percent scattering

k = 2.1

32

32 32

20°

20° 10°

k = 4.0

2°

20° 10°

2°

50

2° 2°

32

k = 2.1 for < 32

20° 10°

k = 4.0 for < 32

10°

0 1

2

4

8

16

32

64

128 256

0.5

1

2

4

8

1

2

4

8

16

32

= Πdnw / Figure 3 Cumulative percentage of scattering calculated from the data of Morel (1973) by McCave (1986). The right-hand case is for a peaked distribution with the peak at a ¼ 32, equivalent to 5.6 mm for l ¼ 633 nm and n ¼ 1.15. Note that material larger than the peak contributes virtually nothing to the scattering in this case. Reproduced from McCave IN (1986) Local and global aspects of the bottom nepheloid layers in the world ocean. Netherlands Journal of Sea Research 20: 167–181.

So it is dominantly the fine fraction of particles in nepheloid layers that is seen and recorded by nephelometers. These particles have very low settling velocities, less than B5 106 m s1. Although larger particles are present, they are rare and their properties and behavior cannot be invoked to explain features of distribution shown by nephelometers. The SeaTech 0.25 m path length transmissometer has been used for most of the modern work on the structure and behavior of nepheloid layers. The transmission T is related to beam attenuation coefficient c over path length l as T ¼ecl. The major control of c is due to particles. Attenuation is due to absorption a and scattering b, thus c ¼ a þ b. The value of c for pure seawater is about 0.36 m1 for this instrument operating at l ¼ 660 nm, and any excess is due to particulate effects. Because scattering very close to the beam is more sensitive to large particles, the transmissometer is more sensitive to larger particles and the nephelometer to smaller ones (Figure 4).

Nepheloid Layer Features The principal features of nepheloid layers that must be accounted for are the facts that the concentration is generally highest close to the bed, decreasing upward, but also that this is not universally the case, because: inversions, upward increases of concentration, are

also found; steeper gradients in concentration as well as inversions are often found at the boundaries of distinct density (temperature and/or salinity) changes, but their frequency decreases upward. The thickness of the BNL is generally in the region of 500–1500 m, and exceptionally up to 2000 m. This is clearly greater than the thickness of the bottom mixed layer (Figures 2 and 4), a fact which rules out the possibility of simple mixing by boundary turbulence being a sufficient mechanism for BNL generation. In several cases, the nepheloid layer is seen to transcend water masses. That is to say, the nepheloid layer shows a general decline in turbidity upward through interleaved water masses of differing sources and temperature/salinity characteristics though there may be a steeper turbidity gradient at the boundary between water masses. The highest suspended sediment concentrations occur in the BMNL in regions of strong bottom currents where they are typically 100–500 mg m3 (this is the same as mg l1). In general, deep western boundary currents and regions of recirculation carry high particulate loads. However, high turbidity is also found beneath regions of high surface eddy kinetic energy (variance of current speed) when located over strong thermohaline bottom currents. High surface eddy kinetic energy is connected with high bottom eddy kinetic energy; thus, intermittent variability, when added to a strong steady


12

NEPHELOID LAYERS

mixing (which occurs mainly in the BMNL) up to a kilometer above the bed. It is not possible to mix sediment across sharp density steps without breaking them down. However, these features are explicable if the layers are recently separated from the bottom. Some layers marked by steps in potential temperature contain excess radon-222 (originating from bottom sediments) with a 3.8 day half-life, suggesting detachment of bottom layers within 2–3 weeks before sampling. It is anticipated that with time these layers become thinner by mixing at their boundaries and by lateral spreading to yield, eventually, a uniform stratification. In this, the upper part of the BNL, sheared-out mixed layers that have, on average, come further from the sloping sides of the basins and from regions with less frequent resuspension, have lower concentrations. The basal layers are on average more recently resuspended and also gain material by fallout from above; thus, there is an overall decreasing particulate concentration, and increasing age upward.

−1

PMC (μg l ) 0

0

50

100

150

200

250

Surface nepheloid layer Attenuation excess Coarser/organicdominated particles

200

400

600

Shallow intermediate nepheloid layer

Depth (m)

800

1000

t −1

Clear water (~10 μg l ) Good agreement between instruments

1200

1400

1600

Deep intermediate nepheloid layer 1800

Bottom nepheloid layer Scattering excess Finer/inorganic dominated particles

2000 26.00

26.25

26.50

26.75

27.00

27.25

27.50

27.75

Decay of Concentration: Aging of Particulate Populations 28.00

3

t (kg m )

Figure 4 Profiles of particulate matter concentration calculated from beam attenuation (solid line), and light scattering (dotted line) against depth, together with the density structure (st) of the water column (dashed line). Reproduced from Hall IR, Schmidt S, McCave IN, and Reyss JL (2000) Particulate matter distribution and Th-234/U-238 disequilibrium along the Northern Iberian Margin: Implications for particulate organic carbon export. DeepSea Research I 47: 557–582.

component, may be responsible for very high current speeds which produce intense sediment resuspension.

Separated Mixed-layer Model Nepheloid layer structure is consistent with a quasivertical transport mechanism involving turbulent mixing in bottom layers of B10–50-m thickness (see Turbulence in the Benthic Boundary Layer), followed by their detachment and lateral advection along isopycnal (equal density) surfaces. The detachment occurs in areas of steep topography as well as in areas of lower gradient at benthic fronts where sloping isopycnals intersect the bottom. In many nepheloid layers, there are sharp upward increases in sediment content associated with steps in other properties such as temperature and salinity (Figure 1). Both the step structure and inversions in particulate matter concentration are incompatible with vertical turbulent

The particles composing the nepheloid layer may be modified due to aggregation with similar-sized particles and scavenging by larger rapidly settling ones. Aggregation may also be caused through biological activity, although little is known about such processes at great depths. Particles tend to settle and to be deposited onto the bed, from the bottom mixed layer. The larger particles should be deposited in a few weeks to months, 10–20 mm particles taking 50– 20 days to settle from a 60-m-thick layer. This will not affect the layer perceived by nephelometers so quickly because the timescale of fine particle removal initially involves Brownian aggregation with a ‘halflife’ of several months to years. The direct rate of deposition of very fine particles (0.5–1 mm) from a layer which remained in contact with the bed would be very slow. Concentration would halve in about 8 years. Thus the rate of decrease in concentration of 0.5–1-mm particles is due more to their being moved to another part of the size spectrum by aggregation (and then deposited) than to their being deposited directly. The fine material in dilute nepheloid layers has a mean residence time measured in years, demonstrated by the residence time of particle-reactive short half-life radionuclides such as 210Pb (t1/2 ¼ 22.3 years). In more concentrated nepheloid layers, a large proportion of this material will be removed in under a year, and in the BMNL residence times are tens to a hundred or so days estimated via 234Th (t1/2 ¼ 24.1 days). The


NEPHELOID LAYERS

dilute nepheloid layers in tranquil parts of the oceans could thus contain material that was resuspended very far away. The contribution of this material to the net sedimentation rate of these tranquil regions may not be negligible. The rate of deposition in the central South Pacific of only 0.5–2 mm ky1 could include up to 1 mm ky1 of fine material from the nepheloid layer. With aging, the individual detached layers comprising the nepheloid layer lose material by aggregation and settling and lose their identity by being thinned through shearing. An originally discontinuous vertical profile of concentration with inversions is converted to one of relatively smooth upward decline

60°

40° E

in concentration. Present understanding of particle aggregation and sinking rates suggests that this takes a few years to achieve.

Chemical Scavenging by Particles in Nepheloid Layers Many chemical species are particle-reactive and rapidly become adsorbed onto surfaces. This is why a number of elements are present in only trace quantities in seawater as outlined by Robert Anderson in 2004. The phenomenon of ‘boundary scavenging’, preferential removal of particle active species at

100°

80°

120° Mid-ocean and aseismic ridges and plateaus

0.2

0°

0.

0.2

E ˚

2

20° N

Excess turbidity [log(Eb /Ec)] values: 1.0

0.2

2

0.6

0.

0.4

0.4

4

0.

2

0.6 0.

0.

2

0.2

40

13

0.6 0.4

0.6 1.0

1.0

0.8

1.0

0.6

0.6

60° S 0.8

4

0.

0.6

0.4 0.6

0.

6

0.4 0.4 0.8

Figure 5 Distribution of excess turbidity for the Indian Ocean expressed as log(E/Ec) where E is the maximum light scattering near the bed and Ec is the value at the clear water minimum. A value of 1 thus represents a factor of 10 increase from the clear-water value. Reproduced from McCave IN (1986) Local and global aspects of the bottom nepheloid layers in the world ocean. Netherlands Journal of Sea Research 20: 167–181; based on Lamont nephelometer data presented by Kolla V, Sullivan L, Streeter SS, and Langseth MG (1976) Spreading of Antarctic bottom water and its effects on the floor of the Indian Ocean inferred from bottom water potential temperature, turbidity and sea-floor photography. Marine Geology 21: 171–189; and Kolla V, Henderson L, Sullivan L, and Biscaye PE (1978) Recent sedimentation in the southeast Indian Ocean with special reference to the effects of Antarctic bottom Water circulation. Marine Geology 27: 1–17.


NEPHELOID LAYERS

scavenging in the BNL on the nearby continental margin. The 234Tht (‘t’ means total, i.e., dissolved and particle-associated) profile showed a 234Tht/238U disequilibrium (the difference between 234Tht and 238 U activity) up to at least 500 mab (meters above bottom), and that secular 234Tht/238U equilibrium was reached between 500 and 1000 mab. That is to say radiochemical evidence suggests lateral advection on the 100-day timescale (4–5 234Th half-lives) below 500 mab at a site quite close to the continental margin (about 250 km away). Above that, the timescale is longer.

The Turbidity Minimum The source of nepheloid layers at heights over a few hundred meters above the bed is believed to be the sides of the ocean basins and protrusions from the bottom of the oceans such as seamounts, ridges, and rises. Major source regions are (1) deep zones of stronger currents, which in most areas are at depths greater than 3000 m; and (2) shallow areas where material resuspended at the shelf edge and upper slope by surface and internal wave motions, spreads

15

seaward. Tidal motion on the rough topography of mid-ocean ridges is known to yield enhanced mixing as explained by Kurt Polzin and colleagues in 1997, and is also probably responsible for resuspension. Between these zones are depths of 1000–3000 m where stratification and currents are weak, there is no primary production of organic matter, and the ocean is often undersaturated with respect to calcite, aragonite, and opal. There is thus little particle supply from the side, and a decrease in downward transport due to dissolution and bacterial consumption of CaCO3, SiO2, and organic carbon, leading to a nepheloid minimum where (in the Atlantic) concentrations are 5–20 mg m3. The concentrations given by nephelometers at the clear-water minimum differ by a factor of about 3 between areas under high surface productivity (high) and mid-gyre regions (low). The flux of large biogenic particles from the surface in these areas appears to provide a net supply of smaller particles to mid-depth (rather than scavenging them). This is also seen to be the case in temporal variability at a point, such that higher mid-water turbidity occurs in summer under higher productivity, and is lower in winter.

71° W 22° N

59° W 17° N 2.2°

4

= 2.0°

1.8° 5

6

Caicos outer ridge log E/ED 0.7_ 0.8

7

0.6_ 0.7 0.5_ 0.6

8

Puerto Rico trench axis

Depth (km)

1.6°

1.4° South slope of Puerto Rico trench

0.4

Figure 7 Interflow of suspended sediment generated in the Western Boundary Undercurrent, having detached from the bed on Caicos drift, flowing over the Puerto Rico trench. Reproduced from Tucholke BE and Eittreim SL (1974) The western boundary undercurrent as a turbidity maximum over the Puerto Rico Trench. Journal of Geophysical Research 79: 4115–4118.


NEPHELOID LAYERS

Concentration and Spreading in the Atlantic and Indian Oceans The broad picture of high-speed inflows on the western sides of ocean basins with a distributed return flow proposed by Henry Stommel has been confirmed by many hydrographic studies and current measurements. These broad patterns of the inflow of cold bottom water are very similar to the excess BNL concentration (excess over the clear-water minimum) for the Indian Ocean, and net BNL particulate load (mass per unit area) for the Atlantic based on Lamont nephelometer data (Figures 5 and 6). It is also apparent that the concentration or load of sediment generally declines going northward along the paths of the major inflows of Antarctic bottom water. However, there are several points of very high concentration or load which are not obviously related to likely increases in mean bottom flow velocity, for example, the high in the center of the Argentine Basin (Figure 6). Richardson and colleagues showed in 1993 that these turbidity highs are sometimes caused by intermittently high velocities under regions where high surface eddy kinetic energy was propagated downward, resulting in high abyssal eddy kinetic energy, but at other times and places there was no correlation with current speeds. Thus some uncertainty remains over the role of high abyssal eddy kinetic energy versus locally accelerated thermohaline flow or deep recirculation loops in producing high concentration. Nevertheless, the major feature of concentrated BNLs is that they do delineate the paths of high-velocity bottom currents extremely well and they decline in concentration away from high-velocity boundary regions and away from the energetic Southern Ocean.

Boundary Mixing, Intermediate Nepheloid Layers, and Inversions Steps and inversions are best shown by modern highsampling-rate sensors (Figures 1, 2, and 4), though striking examples may also be found in older data. INLs are most common off the continental slope and submarine canyons. In both cases, focusing of internal waves (see Internal Waves and Internal Tides) or tides is believed to be responsible, due to the amplification of bottom flow velocities at points where the slope of the seabed S matches the wave characteristic slope C. This causes both resuspension and boundary mixing (see Turbulence in the Benthic Boundary Layer and Estimates of Mixing). The suspensions so generated have both fine and coarse particles at the time of generation but as they spread

seaward they rain out the larger ones over the slope, leaving only finer material with slow sinking speeds. Not all intermediate layers detach on slopes. Perhaps the biggest INL in the ocean occurs where the turbid plume of the North Atlantic Western Boundary Undercurrent leaves the bottom at the end of Caicos Drift off the Bahamas at about 5200 m depth and proceeds at that level over the 8000-m-deep Puerto Rico trench (Figure 7). Pronounced inversions are seen at depths between 1 and 4 km along the continental margin close to the Congo (Zaire) River. A comparable situation is found over Nitinat Fan off the NW USA. This area lies off two submarine canyons which are the most

% Transmission 0

62

64

66

62

64

66

1

2

3 Depth (km)

16

4

5

6

7 9° 17.7′ S, 80° 36.1′ W (Peru trench)

8° 20.4′ S, 81°01.8′ W (Peru trench)

Figure 8 Profiles of the nepheloid layer in the Peru trench made with a 1 m path length transmissometer. Reproduced from Pak H, Menzies D, and Zaneveld JRV (1979) Optical and Hydrographical Observations off the Coast of Peru during May– June 1977, Data Report 77, Ref. 79-14, pp. 1–93. Corvallis, OR: Oregon State University.


NEPHELOID LAYERS

likely sources of the intermediate layers. Thus, canyons may also act as point sources for supply of turbid layers which mix out into the ocean interior.

Trenches and Channels A final example of the influence of the sides of a basin on its nepheloid layers comes from the extremely thick nepheloid layers found in trenches and some deep-sea passages. The nepheloid layer thickness is that of the trench depth plus the thickness of the nepheloid layer over the adjacent seafloor for the Kuril–Japan trench. The result is a layer 2600-m thick. Nepheloid layers 2700-m thick have been recorded in the Peru trench between 81 and 101 S (Figure 8). Flow through Vema Channel between the Argentine and Brazil basins yielded a well-mixed water column in temperature and turbidity with excess radon-222 up to 400 m above the bed. Models of boundary layer development show that this cannot occur by vertical turbulent diffusion. Turbid layers must have spread quickly from the sides to the center of the channel, giving added support to the detached mixed-layer model. The fact that there is more suspended material at depths greater than about 4000 m partly reflects the fact that waters at those depths are in contact with a much greater area of seabed in proportion to their volume than the shallower parts of the oceans. For 4–5 km depth the value is 0.83 km2 km3, whereas for 2–3 km it is 0.11 km2 km3. One might say that deeper waters feel more bed. The global distribution of nepheloid layers is, for instrumental reasons, the global distribution of fine particles (o2 mm), because that is what the instruments that map them ‘see’, but they also contain larger particles which play an important role in sediment dynamics, especially in the lowest 10 m of the BMNL.

See also Dispersion and Diffusion in the Deep Ocean. Floc Layers. Internal Tidal Mixing. Internal Waves. Marine Snow. Particle Aggregation Dynamics. Transmissometry and Nephelometry. Turbulence in the Benthic Boundary Layer.

Further Reading Amin M and Huthnance JM (1999) The pattern of crossslope depositional fluxes. Deep-Sea Research I 46: 1565--1591. Anderson RF (2003) Chemical tracers of particle transport. In: Elderfield H (ed.) Treatise on Geochemistry, Vol. 6:

17

The Oceans and Marine Geochemistry, pp. 247--273. Oxford, UK: Pergamon. Anderson RF, Bacon MP, and Brewer PG (1983) Removal of 230Th and 231Pa at ocean margins. Earth and Planetary Science Letters 66: 73--90. Armi L and D’Asaro E (1980) Flow structures of the benthic ocean. Journal of Geophysical Research 85: 469--484. Bacon MP and Rutgers van der Loeff MM (1989) Removal of thorium-234 by scavenging in the bottom nepheloid layer of the ocean. Earth and Planetary Science Letters 92: 157--164. Baker ET and Lavelle J W (1984) The effect of particle size on the light attenuation coefficient of natural suspensions. Journal of Geophysical Research 89: 8197--8203. Biscaye PE and Eittreim SL (1977) Suspended particulate loads and transports in the nepheloid layer of the abyssal Atlantic Ocean. Marine Geology 23: 155--172. Dickson RR and McCave IN (1986) Nepheloid layers on the continental slope west of Porcupine Bank. Deep-Sea Research 33: 791--818. Gardner WD (1989) Periodic resuspension in Baltimore Canyon by focusing of internal waves. Journal of Geophysical Research 94: 185--194. Hall IR, Schmidt S, McCave IN, and Reyss JL (2000) Particulate matter distribution and Th-234/U-238 disequilibrium along the Northern Iberian Margin: Implications for particulate organic carbon export. Deep-Sea Research I 47: 557--582. Jerlov NG (1976) Marine Optics. Amsterdam: Elsevier. Johnson DA, McDowell SE, Sullivan LG, and Biscaye PE (1976) Abyssal hydrography, nephelometry, currents and benthic boundary layer structure in Vema Channel. Journal of Geophysical Research 81: 5771--5786. Kolla V, Henderson L, Sullivan L, and Biscaye PE (1978) Recent sedimentation in the southeast Indian Ocean with special reference to the effects of Antarctic bottom water circulation. Marine Geology 27: 1--17. Kolla V, Sullivan L, Streeter SS, and Langseth MG (1976) Spreading of Antarctic bottom water and its effects on the floor of the Indian Ocean inferred from bottom water potential temperature, turbidity and sea-floor photography. Marine Geology 21: 171--189. McCave IN (1983) Particulate size spectra, behavior and origin of nepheloid layers over the Nova Scotian continental rise. Journal of Geophysical Research 88: 7647--7666. McCave IN (1986) Local and global aspects of the bottom nepheloid layers in the world ocean. Netherlands Journal of Sea Research 20: 167--181. McPhee-Shaw E (2006) Boundary–interior exchange: Reviewing the idea that internal-wave mixing enhances lateral dispersal near continental margins. Deep-Sea Research II 53: 42--59. Menard HW and Smith SM (1966) Hypsometry of ocean basin provinces. Journal of Geophysical Research 71: 4305--4325. Morel A (1973) Indicatrices de Diffusion Calculeés par la Theórie de Mie pour les Syste`mes Polydisperses, en Vue de l’Application aux Particules Marines, Report 10, pp.


18

NEPHELOID LAYERS

1–75. Paris: Laboratoire d’Oceánographie Physique, University of Paris VI. Pak H, Menzies D, and Zaneveld JRV (1979) Optical and Hydrographical Observations off the Coast of Peru during May–June 1977, Data Report 77, Ref. 79-14, pp. 1–93. Corvallis, OR: Oregon State University. Polzin KL, Toole JM, Ledwell JR, and Schmitt RW (1997) Spatial variability of turbulent mixing in the abyssal ocean. Science 276: 93--96. Richardson MJ, Weatherly GL, and Gardner WD (1993) Benthic storms in the Argentine Basin. Deep-Sea Research 40: 957--987. Thorndike EM (1975) A deep sea, photographic nephelometer. Ocean Engineering 3: 1--15.

Tucholke BE and Eittreim SL (1974) The western boundary undercurrent as a turbidity maximum over the Puerto Rico Trench. Journal of Geophysical Research 79: 4115--4118. Turnewitsch R and Springer BM (2001) Do bottom mixed layers influence 234Th dynamics in the abyssal nearbottom water column? Deep-Sea Research I 48: 1279--1307. Zaneveld JRV, Roach DM, and Pak H (1974) The determination of the index of refraction of oceanic particulates. Journal of Geophysical Research 79: 4091--4095.


NETWORK ANALYSIS OF FOOD WEBS J. H. Steele, Woods Hole Oceanographic Institution, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1870–1875, & 2001, Elsevier Ltd.

Introduction Photosynthesis transforms energy from sunlight into calories within marine plants, predominantly phytoplankton and seaweeds. The plants use this energy to take up carbon and essential nutrients, such as nitrogen and phosphorus, from sea water to produce organic materials.This organic matter forms the base for the food web composed of herbivores thateat those plants and the carnivores that prey on the herbivores. There can beseveral trophic levels of carnivores, including all the fish species that weharvest. In the open sea, away from the coast and theseabed, microscopic single-celled phytoplankton dominate the plant life, so the organisms tend to get bigger as each trophic level feeds on the onebelow: from small herbivorous crustaceans to larger invertebrates, to small andlarge fish, and finally to human beings and marine mammals. When any animal consumes food, most of the energy in that food is used for metabolism; some of the remainder is excretedas waste products and only a small fraction goes to growth. In young, coldblooded animals in the sea, growth can be relatively efficient: 20–30% of energy intake. In older animals growth is replaced by reproduction, which, after all, is the whole point of the life cycle. As an approximate overall figure we usually assume that the energy converted into growth and reproduction is about 10% of the total energy intake. Thus, in asimple trophic pyramid, the energy in successive trophic levels would be 100:10:1:0.1. From this one can see why we are encouraged to eat plants on land, and why fish from the sea are energetically expensive in terms of plant calories. In practice it is very difficult to measure directly the energy content of marine organisms and, especially the energy transfers between trophic levels. However, carbon is the essential building block for organic matter, being taken up from inorganic form at photosynthesis and returned to sea water during respiration. The carbon content of organisms and the rates of uptake of inorganic carbon can be measured using radioactive carbon, carbon-14, as a tracer;

transfers through the food can then bemeasured. Carbon is therefore frequently used as a proxy for the more elusive concept of energy flow. All organisms also require many essential elements, and many of these are in short supply in sea water. In particular, inorganic nitrogen and phosphorus, as nitrate and phosphate, are regarded as limiting factors in photosynthesis and thus in the rate at which energy and carbon are supplied to the food web. Since organic carbon, nitrogen and phosphorus have a roughly constant ratio in marine organisms (the Redfield ratio), nitrogen can also be used as a proxy for energy flow; it will not, however be considered here. Carbon and nitrogen fluxes are also important in relation to other issues concerning marine food webs. The biologically mediated flux of carbon to deeper water is important for the calculation of global carbon budgets and climate change. Eutrophication in coastal waters produces imbalance in the food webs of these regions.

Units Ideally, all presentations would be in units of energy but, as discussed, the actual measurements are often in carbon or, for fisheries, in biomass (wet weight). Conversion from one unit to another will vary with the organism but, as an approximation: 10 kcal ¼ 1g carbon ¼ 10 g biomass:

½1

History The first attempt to produce an energy budget for an aquatic food web was made by Lindeman in 1942, for a freshwater lake. In 1965 a budget for the North Sea (Figure 1), showed possible pathways from primary production to commercial fisheries. The main conclusion was that the ecological efficiencies needed to be quite high for the budget to balance. This conclusion depends on the number of pathways introduced; for example, the category ‘other carnivores’ is put in to account for the presence of gelatinous invertebrate predators, such as ctenophores, that are ubiquitous but the biomass of which is difficult to estimate. Without this box, the herbivore efficiency could be less than 20%. In 1969 Ryther wrote a seminal paper that outlined the pathways at the global scale, by dividing the marine ecosystems into three types, upwelling,


19

20

NETWORK ANALYSIS OF FOOD WEBS

Net primary production

Table 1 Primary production, number of trophic levels to commercial fish species, efficiency of transfer of energy (or carbon) between trophic levels and resultant ratios of fish to primary production for three major marine categories

(Detritus) Meiobenthos

Macrobenthos

Other carnivores

Demersal fish

Pelagic herbivores Other carnivores

Pelagic fish

Primary production (gC m2 year1) Trophic levels Ecological efficiency (%) Fish production: primary production (%)

Large fish Yield to humans

(A)

100 20

Coastal

Upwelling

50

100

300

5 10 0.01

3 15 0.3

1.5 20 12

80

8 2

Ocean

12

1 2

16

3

4

0.4

0.17

0.3

12

0.6

1.8

0.04

gelatinous predators, there is still relatively little quantitative information on biomass and production; this has led to an extensive development of mathematical treatments to infer energy or carbon flows.

0.04 (B)

0.13

0.3

25%

23%

20%

13%

10%

10%

Quantitative Methods

15%

(C) Figure 1 Energy or carbon fluxes in the food web for the North Sea.(A) A simplified food web. (B) Transfers of particulate organic carbon assuming a primary production of 100 gC m2 peryear. Each box indicates the production by that component available to its predators, including humans. (C) The transfer or ecological efficiencies calculated from the output as percentage of input.

continental shelf, and open ocean(Table 1). The major implication of Ryther’s calculations was that there were no great untapped fish resources in the open ocean. He estimated that the potential sustained yield offish to humans was unlikely to be greater than 100 million tons. In 1998, the yield was about 90 million tons. The reason for the low open-oceanyield of fish – and the controversial part of Ryther’s calculation– is his estimate of five trophic levels from primary production to fish. This choice was based on the very small size of open-ocean phytoplankton. In the intervening years, research on the base of open-ocean food webs has shown that much of the primary production is recycled through the smallest sized categories in the food web – the microbial loop. For the intermediate trophic levels, such as the

Early calculations, such as those illustrated in Figure 1, were put together by a very informal series of iterations until all the flows balanced and the ecological efficiencies were not unreasonable. The major advance has been the development of numerical methods to make more objective estimates of best fit. Essentially, the ecosystem is considered to be at steady state, so that, for any box, such as those in Figure 1, there has to be a balance between rates of energy flow or carbon flux entering and leaving: consumption ¼ predation þ metabolism þ exportðinputÞ

½2

The information for each of these terms for each box canbe qualitatively different. In particular, for animals: consumption ¼ growth þ reproduction þ metabolism ½3 These complexities are eliminated if it is assumed that, for each box, there is a constant ecological efficiency, ei : ei ¼

ðgrowth þ reproductionÞ consumption

½4

Generally, for the system to be soluble, the terms in eqn [2] for all boxes must be expressed as a set of linear equations. For this purpose the variables are usually taken to be the energy, carbon or biomass flux through each box, Ti , that is available to higher



trophic levels. Linearity requiresthat the rate processes in eqn [2] are constants independent of Ti . Ti X ¼ bij Tj þ cj i ¼ 1; n; ½5 ei where bij is the fraction of Tj consumed by Ti and ci is the constant rate of external input to Ti ; ci would be, for example, the rate of primary production. Note that, as in Figure 1, all the variables are rates of throughput in the food web and the parameters are non-dimensional.These matrix inversion techniques can also be used with nutrient flows that involve recycling. An alternative top-down approach is often used for flow calculations where emphasis is on the higher trophic levels, and fish yields are the defining input. For these situations,consumption, Ci , is expressed as: ðconsumptionÞ biomass biomass Ci Bi ¼ Bi

consumption ¼

½6

Then the biomass in each box, Bi , becomes the state variable. The rate process Ci =Bi is assumed constant, for each box. Then: ai Bi ¼

X

pij Bj þ di ;

½7

where ai represents production per unit biomass, the ‘P/B’ ratio, assumed constant for any box; pij is the unit consumption rate of Bj on Bi , assumed to be independent of the magnitude of Bj or Bi ; di is the export, assumed constant for each box. If all the parameters are known for either eqns [2] and [7] then the set of equations canbe solved for Ti or Bi. In practice it is never as simple as that. Usually a number of parameters are unknown or ill defined; for example, as upper or lower bounds.Then iterative procedures can be used to obtain best fits. The selection of the number of boxes and the content of each box is the critical process. The assumption of a linear, steady-state ecosystem is the critical constraint for this type of analysis involving a large number of boxes. This approach is complemented by the highly nonlinear analysis used for modelling offisheries and plankton dynamics.

Examples of Carbon and Biomass Networks The emergence of the microbial loop as asignificant feature of pelagic food webs, together with the availability of computer based inverse methods, resulted in a focus on flow analysis of the lower levels

21

of the pelagic ecosystem; levels that were represented simply as phytoplankton-to-zooplankton in Figure 1. The results of analyses (Figure 2) illustrate how this part of the system isexpanded into seven boxes with 19 links. Two examples in Figure 2, from the continental shelf around Britain, show the kinds of patterns that arise from these calculations. The authors point out differences between the English Channel and the Celtic Sea. The former puts 85% of primary production through the microbial loop, whereas the latter has 40% going directly to the mesoplankton (predominantly copepods). However,for both examples, the exports from this part of the food web are very similar: 3% to higher trophic levels via the mesoplankton, and 30% to the benthos as detritus. The major change from the earlier calculations is the dominant role of the microbial loop, involving recycling of most of the primary production.Only the ‘new’ production from nitrate (NO3) in Figure 2A fuels the export of detritusand mesoplankton to higher trophic levels. The alternative approach (eqn [7]) has been used for a wide range of aquatic ecosystems, including lakes and coral reefs as well as ocean systems. One example (Figure 3) shows calculations for the continental shelf around the Gulf of Mexico. There is no microzooplankton box and the general flow patterns are not dissimilar to those of Figure 1; 30% of primary production goesdirectly to detritus, as does 40% of zooplankton food, presumably as fecal material. Detritus then feeds the demersal fish through the benthos. Note, however, that once again detritus is the dominant biomass (85% of the total). It would appear that all the solutions of these linear systems – whether as energy or biomass, for lower or higher trophic levels – require a major role for detritus. Detritus is a difficult variable to define and measure. This box can contain phytoplankton cells, zooplankton feces, marine snow and other residues. It is assumed that detritus is broken down by bacteria but the ai or ei values are not well known. It is rarely sampled directly, eitheras biomass or for rate processes, and so is an empty box that can be used to balance flows.

Discussion In terrestrial ecosystems most of our food comes directly from plants, with the remainder from herbivores – animalsthat eat plants. In the sea, the fish we eat are nearly all carnivores. Many feed on the small herbivorous crustaceans, such as krill, but some – the most highly prized, like tuna – themselves feed on carnivores, such as smaller fish. So what we hope to take from the sea, the potential fishery yield,


22


Sea surface A

NH4

Mi

NO3

B

Me

DOM

P Export

D Euphotic zone Aphotic zone

(A)

Sea surface GPP

A 169

720

71

257 91

Me 95

28 104

Mi 126

160 31.5 41.5

41.5 107 9.6

25.4 141

36 11.4

P 46

B 35.8 DOM 41.9 194 15.1

Export D

Euphotic zone Aphotic zone

216

(B)

Sea surface GPP

A 71.9

729

72

163 285 39 99

Me 227 23 136

Mi 67.9

22.4

103

62

75 73

60

P 31.4 51.8 15.8 4.1

B 23.4 DOM 60.5 210 17.1

X dBi ¼ ai Bi pij Bj di dt

Export D 227

(C)

depends very much on the patterns and magnitudes of the flow of energyor carbon from primary production. Our understanding of these patterns cantherefore provide valuable estimates on the limits of what we can take from thesea, as well as increasing our insight into the ecological processes. The selection of boxes and the arrows between them depends on our knowledge of which prey– predator interactions are quantitatively significant. It is obvious from the examples here that thereis a considerable compaction into large, often heterogeneous groups, such as mesoplankton or pelagic fish. Thus the level of organization is very differentfrom that for biodiversity studies, or even for descriptions of the full complexity of the food web. However, in producing a manageable number of boxes, it is important not to fold significant prey– predator interactions into the same box. Thus, in Figure 1, prey–predator interactions at the microbial scale were ignored. On the other hand, it has been pointed out that, for the North Sea, the inclusion of both the detrital box and the invertebrate carnivores does not provide enough energy flow for the fishery yield. The definition of boxes will remain acentral problem, but this is also the great strength of this approach: it requires attention to all aspects of the food web. The major limitation is that the method assumes that the system is in steady state. This is usually achieved by taking yearly averages and ignoring shorter seasonal variability and longer decadaltrends. For the former, several researchers have constructed dynamic planktonmodels of the nonlinear interactions between nutrients and detritus based onthe seven boxes or variables in Figure 2. The longer term changes are especially important for fisheries. It is possible to transform the linearized eqn [7] for equilibrium states into a time varying system by writing:

Euphotic zone Aphotic zone

Figure 2 (A) Generic model of a plankton food web in the upper layer of a stratified water column. (B) The inverse solution for carbon fluxes in summer at a station in the English Channel. Values inside the boxes are respiration flows (mgC m2 per day).(C) Flows at a station in the Celtic Sea. A, autotrophs; B, bacteria; D, detritus; DOM, dissolved organic matter; solid arrows denote intercompartmental transfer of carbon; dashed arrows indicate flows of inorganic nitrogen into the system; Me, mesozooplankton; Mi, microzooplankton; P, Protozoans. (Vezina and Platt, 1988.)

½8

This linearized approximation to an essentially nonlinear system can be used to indicate the directions that changes may take when the system departs from a previous steady state, but is unlikely to be adequate for the very large switches, or regime shifts, that occur in the relative abundance of different fish stocks. An alternative is to assign very different values to the boxes for the fish stock biomass and then recalculate the network. This technique can be used to estimate the status of ecosystems before the impact of human predation on fish or marine mammals. As an example,the Gulf of Mexico flows were



1

2

3

444.8

17.3

0.1

311.7

1776.51

Detritus B = 236.0

200

140.01

0.2

B = 0.13

54.1

0.7 1.3 209.7

0.03 0

10031

0.01

11.51

Pelagic predators B = 0.05

0.0 0.02 0.3

Phytoplankton B = 4.7

5420

1970.61

B = 3.6

391.4

06 0.01 13.5 0.4 213.7 313.6 Zooplankton 28.5

256.8

0.01 0

0

Billfish B = 0.01

0.7

10.81

Dolphin B = 0.02

0.1

1.5

0.03

1270.31

Pelagic fish B = 12.5

0.2 2.7 0.02

10.71

0.1

2.7 Mackerels 0

0.7

0.02 13.31 0

B = 0.04

0 0.5 Tunas

4.4

00.4

0

Macro- 9.8 crustaceans B = 1.0

0.4

8.0

0.03

Sharks B = 0.08

10.71

0.1

0.2

18.71

B = 0.7

2.7

0.8 0.2

40.4

2.0 20

Benthos plants B = 3.7

18.3

133.01

6.5

Benthos B = 5.0

88.7

11

36.7

18.33

36.2

10.9

153.51

Demersal fish B = 3.5

17.6

0.6 0.4 4.4

0.2 0.04 1.6 0.04 Demersal 6.4 0.3 predators

Other exports

Respiration

Harvest

Flow to detritus

Figure 3 Gulf of Mexico ecosystem model: trophic compartments with values for biomass (g m2) and connecting flows (gm2 per year1). (Christensen and Pauly, 1993.)

Trophic level

4

NETWORK ANALYSIS OF FOOD WEBS 23

24


recalculated, with the biomass of the top predators increased by a factor of 10. The major change was that the utilization of detritus within the ecosystem increased from 11% to 70%. Thus the ‘detritus’ box appears to act as a reservoir for over-exploited systems.

Conclusions Calculations of the overall energy, orcarbon, fluxes through a marine ecosystem provide a valuable check on estimates for the potential productivity associated with each component of the food web. These calculations can set limits on expected yields to humans; they can act as links between detailed, but necessarily static, descriptions of biodiversity, and models of the dynamics of individual populations. It is necessary to bear in mind that our knowledge of the food webs used in these calculations are provisional and the outcome of the calculations is dependent on the specification of this structure; thus the full benefits from this approach depend on further studies of each ecosystem.

Further Reading Christensen V and Pauly D (eds) (1993) Trophic models of aquatic ecosystems. ICLARM Conf. Proc. 26: 390pp. Christensen V and Pauly D (1998) Changes in models of aquatic ecosystems approaching carrying capacity. In: Ecosystems management for sustainable marine fisheries. Ecol. Appl 8 (suppl). Fasham MJR (1985) Flow analysis of materials in the marine euphotic zone. In: Ulanowicz RE and Platt T (eds) Ecosystem theory for biological oceanography. Can. Bull. Fish Aquat. Sci. 213: 260pp. Lindeman RL (1942) The trophic-dynamic aspect of ecology. Ecology 23: 399--418. Ryther JH (1969) Photosynthesis and fish production in the sea. Science 166: 72--76. Steele JH (1974) The Structure of Marine Ecosystems. Boston: Harvard University Press. Vezina AF and Platt T (1988) Food web dynamics in the ocean. I. Best-estimates of flow networks using inverse methods. Mar. Ecol. Prog. Ser 42: 269--287. Wulff F, Field JG, and Mann KH (eds.) (1989) Network Analysis in Marine Ecology. Berlin: Springer-Verlag.

See also Carbon Cycle. Copepods. Diversity of Marine Species. Eutrophication. Krill. Large Marine Ecosystems. Marine Snow. Microbial Loops. Photochemical Processes. Upwelling Ecosystems.


NEUTRAL SURFACES AND THE EQUATION OF STATE T. J. McDougall and D. R. Jackett, CSIRO Marine and Atmospheric Research, Hobart, TAS, Australia & 2009 Elsevier Ltd. All rights reserved.

Introduction The equation of state of seawater relates the in situ density r of a parcel of seawater to its salinity S, temperature t, and pressure p. Because in situ temperature t is not a conservative variable, it is often more convenient to express the equation of state in terms of the potential temperature y: potential temperature is the temperature that a fluid parcel would have if its pressure were changed adiabatically and without exchange of matter, from the in situ pressure p to a fixed reference pressure pr . The equation of state then takes the functional form r ¼ r(S, y, p). Even without the exchange of heat and salt so that a fluid parcel retains its salinity and potential temperature, the fact that seawater is compressible means that the parcel’s in situ density r will change as it moves around the ocean experiencing different pressures. In order to counteract the effect of compressibility, Wu¨st in 1933 introduced ‘potential density’ which is the density that a fluid parcel would have if its pressure were changed adiabatically and without exchange of salt to an arbitrarily chosen, but fixed, reference pressure. Oceanographers often use sea surface pressure as the reference pressure, pr ¼ 0 and potential density is often given the symbol ry ¼ r(S, y, 0). However, because the compressibility of seawater is not constant, when a fluid parcel is moved a small distance along a potential density surface, a vertical force must be supplied to keep the parcel on that surface; in other words, a potential density surface is not a surface of neutral or zero buoyancy. Because of the theoretical difficulties in forming a neutral surface throughout extended regions of the ocean, we initially restrict attention to the requirement for a tangent plane to possess the property of neutral buoyancy. A neutral tangent plane in threedimensional space is defined so that an in situ parcel of seawater can be moved small distances along the neutral tangent plane without experiencing a vertical buoyant restoring force. The partial derivatives of

density are used to define the three coefficients 1 @r 1 @r ; a¼ ; b¼ r @Sy;p r @yS;p

and

1 @r k¼ ½1 r @pS;y

where b is the saline contraction coefficient, a is the thermal expansion coefficient, and k is the adiabatic (and isentropic) compressibility (rk ¼ c2 where c is the speed of sound in seawater). When a fluid parcel is moved a small distance adiabatically and without change in salinity in the neutral tangent plane, neither the potential temperature nor the salinity of the parcel changes during this movement, so the parcel’s in situ density changes only in response to its change in pressure, multiplied by the compressibility k. The definition of the neutral tangent plane in terms of the lack of buoyant restoring forces implies that the parcel’s in situ density must be the same as that of the seawater that surrounds the parcel. In this way, the two-dimensional gradients of in situ density and pressure in the neutral tangent plane obey rn ln r ¼ krnp, and expanding the gradient of in situ density in terms of the gradients of salinity, potential temperature, and pressure using eqn [1], we find that the spatial property gradients in the neutral tangent plane obey both rn ln r ¼ krn p

and

brn S ¼ arn y

½2

Here rn is the two-dimensional spatial gradient operator for properties measured in the sloping neutral tangent plane although the horizontal elements of distance are measured in the geopotential surface. This definition of the neutral tangent plane in terms of the lack of buoyant restoring forces implies that the normal n to the neutral tangent plane is in the direction ary brS ¼ krp rln r

½3

which is written in terms of three-dimensional spatial property gradients. Oceanographers care about neutral tangent planes because the turbulent diffusivity that represents vertical (or diapycnal) diffusion is approximately 8 orders of magnitude smaller than the diffusivity that represents the turbulent mixing of properties along neutral tangent planes. In order to separately understand the very different lateral and vertical mixing processes it is necessary to be able to


25

26

NEUTRAL SURFACES AND THE EQUATION OF STATE

determine the slope of the relevant mixing directions (i.e., the slope of the neutral tangent plane) to an accuracy of about 10 4. As explained by McDougall and Jackett, it is the smallness of the rate of dissipation of mechanical energy in the ocean that confirms the common assumption that the strong lateral diffusion achieved by mesoscale eddies occurs along neutral tangent planes in the ocean interior.

Requirements for a Neutral Surface to Exist Intuitively one might expect that since neutral tangent planes exist everywhere in the ocean, these planes would join to form a series of ‘neutral surfaces’. However, it is well known that surfaces normal to the vector n are only well defined if n r n is everywhere zero. Since the normal to the neural tangent plane is in the direction ary brS the requirement that a neutral surface exists in some region of physical space is that neutral helicity defined by H ðary brSÞ r ðary brSÞ

½4

is zero everywhere in that region. McDougall and Jackett have shown that neutral helicity can be written as H ¼ bTb rp rS ry

½5

where the thermobaric parameter Tb which represents the key nonlinearity of density, is Tb ¼ ap ða=bÞbp ¼ bða=bÞp

½6

From figure 9(b) of McDougall’s 1987 article (featured in the Journal of Geophysical Research) we see

that a value of TbE2.7 108 K1 db1 ¼ 2.7 1012 K 1 Pa 1 is typical of the bulk of the ocean’s volume. Equation [5] says that if neutral helicity is zero, the line of intersection of the S and y planes, rS ry, must lie in the isobaric surface; approximately, this means that rS ry must be a horizontal vector. A sketch of the S, y, r, and p planes and the lines rS ry and rp rr is shown in Figure 1. The neutral tangent plane is always well defined and it includes both the lines ry rS and rp rr, but a neutral surface is well defined only if these two lines coincide. On the face of it, this requirement seems very restrictive; one would think that the distributions of salinity and temperature would demand more freedom than to be bound by this mutual constraint. Nevertheless, McDougall and Jackett showed that the ocean’s hydrography is close to obeying this seemingly unreasonable constraint, implying that all six colored planes of Figure 1 intersect along the single line ry rS//rp rr. Because neutral helicity is found to be quite small in the ocean one is able to construct properly defined surfaces whose tangent planes give very good approximations to the neutral tangent planes. One way of forming such approximately neutral surfaces is described in the section titled ‘Neutral density surfaces compared with potential density surfaces’.

The Helical Nature of Neutral Trajectories When one forms an approximately neutral surface (such as can be found by using the Neutral Density software of Jackett and McDougall), one indeed

Neutral tangent plane

∇

×∇ S

S

∇p ×

∇

Figure 1 Sketch of the planes of constant salinity S, potential temperature y, potential density ry, all intersecting along the line ry rS and the planes of constant pressure p, and in situ density r, that intersect along the line rp rr The neutral tangent plane includes both the lines ry rS and rp rr. A neutral surface exists if and only if these two lines coincide so that the line ry rS lies in the isobaric surface. From McDougall TJ (1988) Neutral-surface potential vorticity Progress in Oceanography 20: 185–221.



27

urface

Neutral helix

Sea s

b Approximately neutral surface

tant p cons

a e θ,S

co

, S

co

ns

ta

nt

ns

tan

t

c d

nt

p consta

z y Bottom

x Figure 2 The neutral helix depicts circulation such as would occur in diapycnal mixing. The helical nature means that mean vertical motion surfaces.

a trajectory of the mean the absence of interior of this neutral trajectory occurs through density

finds a mathematically well-defined surface but such a surface will not exactly obey the neutral property that fluid parcels on the surface can be moved small distances in the surface without feeling buoyant restoring forces. This is illustrated in Figure 2, which shows schematically a spiraling motion along a sequence of neutral directions following the mean circulation. Even in the absence of small-scale diapycnal mixing processes, flow along this neutral helix will achieve mean dia-surface transport though any real density surface that one cares to construct. This spiraling motion amounts to a new diapycnal advection mechanism that operates independently of the diapycnal motion caused by breaking internal waves. A present challenge is to properly quantify this upwelling mechanism. Further insight into the cause of the helical nature of neutral trajectories can be gleaned from Figure 3. The first leg of the neutral trajectory (from a to b) occurs at constant pressure with the salinity and temperature decreasing in a neutral manner according to eqn [2]. The next leg has the salinity and temperature constant as the pressure increases; the third leg has increasing temperature and salinity (in accordance with eqn [2]) at constant pressure. The fourth leg (from d to e) has constant temperature and salinity while pressure decreases. The reason why the final point along the neutral trajectory, e, does not return to point a can be understood from the Sy diagram of Figure 3. The changes in salinity and potential temperature that occur along legs ab and cd occur at different pressures, so these legs appear in the Sy diagram of Figure 3 as potential density

a Slope = (pa)

d,e

Slope = z /S z

b,c

Slope = (pc)

S Figure 3 An idealized example of how mean vertical advection arises from the ill-defined nature of neutral surfaces. From McDougall TJ and Jackett DR (1988) On the helical nature of neutral trajectories in the ocean. Progress in Oceanography, 20, 153–183.

isolines, but they are potential densities referenced to different pressures. These lines have different slopes because the ratio of the expansion coefficients, b/a, is a function of pressure.

Neutral Density Surfaces Compared with Potential Density Surfaces This article has so far concentrated on the theoretical reasons why neutral surfaces do not exist in the ocean. The field of oceanography has used several different types of surfaces that to varying degrees approximately possess the neutral property. The first and most obvious surface to consider is a surface of constant in situ density. It is easy to demonstrate that seawater is sufficiently compressible in comparison


28


with the vertical stratification of the ocean so that an in situ density surface is often quite close to being isobaric. In fact, sound fixing and ranging (SOFAR) floats are much less compressible than seawater and thus these floats have constant mass and almost constant volume so that their density is almost constant. Hence these floats move around the ocean close to in situ density surfaces, and these floats are often described as ‘isobaric floats’. Potential density surfaces have been the most commonly used approximation to neutral mixing directions for much of last century. Potential density is defined to be the density that a parcel of seawater would have if its pressure were changed in an adiabatic and isohaline fashion from its in situ pressure to a fixed reference pressure. Typically reference pressures of 0, 2000, and 4000 dbar are chosen. In the last half of the nineteenth century, Reid and Lynn improved on the straightforward use of potential density by linking successive potential density surfaces that were referenced to different reference pressures. The choices in their method were made based on latitude, and importantly, the particular potential density surface that spanned the equator was common to the definitions that applied to the Northern and Southern Hemispheres. They showed that at the sea surface outcropping of their patched potential density surfaces, the in situ density was different by up to 0.14 kg m3 between the Northern and Southern Hemispheres. This offset of both in situ density and potential density between the hemispheres must be respected by any variable or surface that approximates the neutral property. By so doing, such a ‘neutral density’ variable becomes nonthermodynamic, since it depends on not only salinity, temperature, and pressure, but also location. Jackett and McDougall published a computer algorithm in 1997 for determining a new density variable called neutral density gn for ocean data. Underlying this algorithm is an accurately labeled world ocean data set that has been prelabeled with the neutral density variable. A gn label for an arbitrary {S, T, p, lat, long} ocean observation is determined by first finding the depths on the surrounding four casts in the reference data set where the hydrographic properties of the {S, T, p} observation and the four points in the reference data set are on the same neutral tangent plane. The gn label of the {S, T, p, lat, long} observation is then taken as a weighted average of the interpolated gn values in the prelabeled data set. For the global neutral density variable gn defined by Jackett and McDougall, the global reference data set was based on the climatological atlas of Levitus, with slight modifications having been made to account for extremities in

density and an adequate resolution of the Antarctic Circumpolar Current. Since the normal n to the neutral tangent plane is in the direction of brS ary, the variable gn in the reference data set is taken to approximately satisfy rgn ¼ brðbrS aryÞ

½7

where b is an unknown scalar function of space. This equation is a system of first-order hyperbolic partial differential equations, generally solved using the method of characteristics. The characteristic directions of eqn [7] turn out to be the neutral tangent planes which satisfy brnS ¼ arny and along which gn is approximately constant, so the method of characteristics for this problem reduces to fitting approximately neutral surfaces to the three-dimensional reference data set, with the values of gn on these surfaces being determined by a knowledge of gn down just one cast. The noncharacteristic cast we choose in our global data set is located at 161 S, 1881 E, down which we defined gnref ðzÞ ¼ ry ð0Þ 1000 kg m3

Z0

rg1 N 2 dz

½8

z

being a simple integration of eqn [7] in the vertical with a b profile of unity. The equation of state used for the calculations of r, a, b, and N2 in eqns [7] and [8], is the UNESCO equation (see Appendix 5. Properties of Seawater). The ambiguity in defining neutral density manifests itself in that gn is not exactly constant on local neutral tangent planes. When one integrates along a neutral trajectory (which everywhere is parallel to neutral tangent planes) around loops of global scale one finds that the vertical excursion of such a helix is of the order of tens of meters, consistent with the small size of neutral helicity eqn [5]. The final important stage in generating the gn field in the global reference data set consists of averaging this error over the entire globe. This was achieved by iteratively averaging to a steady state the values of the initial gn field on local tangent planes calculated in the northward, southward, eastward, and westward directions at each location: in effect, a cellular automata of the world ocean. Notice that the initial value of b ¼ 1 down the Pacific Ocean reference cast in eqn [8] was slightly altered during this averaging process. Figure 4 shows various density surfaces along the 3281 E longitude meridian in the Atlantic Ocean taken from the Koltermann et al. hydrographic atlas of the world ocean. The red line shows a neutral trajectory from south to north calculated by joining accurately defined neutral tangent planes in this



29

0

Pressure (dbar)

1000

2000

3000

4000

5000

6000 −80

−60

−40

−20

0 Latitude

20

40

60

Figure 4 Surfaces of constant gn (blue solid line), sy and s2 (dashed lines with s2 being the lower dashed line), and a neutral trajectory (red line) along 3281 E in the Atlantic Ocean. All the surfaces coincide at 250 dbar and 591 S.

north–south direction, while the blue line shows the gn surface obtained by interpolating the gn field after labeling with neutral density. The two surfaces coincide at 591 S at 250 dbar pressure. The two dashed lines are potential density surfaces referred to pressures of 0 dbar (sy) and 2000 dbar (s2), again coinciding with the neutral trajectory at 591 S. The figure demonstrates the dramatic improvement in the accuracy of neutral density gn in approximating the neutral trajectory over the more traditional potential density surfaces. The maximum pressure difference between the gn surface and the neutral trajectory is 50 dbar while the maximum pressure difference for the sy surface is 1250 dbar and for the s2 surface 970 dbar. A more global measure of the error between the gn surfaces and the local neutral tangent planes can be made by considering the effective diapycnal diffusivity of density when salinity and potential temperature are fluxed along gn surfaces rather than along-neutral-tangent planes. This fictitious diapycnal diffusivity 2 can be shown to be DVS ¼ Krgn z rn z , where rgn z and rnz are the two-dimensional lateral slopes of the gn surface and of the neutral tangent plane, respectively, and K is a lateral diffusivity, which we take as 1000 m2 s1. Here we have designated this false diapycnal diffusivity DVS since George Veronis and Henry Stommel first noticed this effect of mixing along a direction other than the local neutral tangent planes (they were concerned with the incorrect mixing direction being along geopotentials). In Figure 5, we have plotted the

frequency distribution of the logarithm to the base 10 of the fictitious diapycnal diffusivity DVS for various surfaces: (1) for all data in the Koltermann et al. global ocean atlas excluding the Arctic Ocean and marginal seas; and (2) for the modified Levitus reference data set. The fictitious diapycnal diffusivity was evaluated for lateral mixing along gn, sy, and s2 surfaces for (1) and only along the gn surfaces for (2) (‘labeled Levitus’ in Figure 5). The best surfaces in approximating the neutral tangent plane are clearly the two neutral density surfaces, the more accurate of these being the neutral density surface in the Levitus data that underlies the definition of neutral density. The fraction of the ocean volume exhibiting fictitious diapycnal diffusivities greater than a fixed value can be estimated by the area under these curves in Figure 5 to the right of the fixed value. It is clear that there are very relatively places where the fictitious diapycnal diffusivity exceeds 105 m2 s1 for neutral density surfaces.

Equation of State The definition of practical salinity in terms of conductivity was settled in 1978 and this is central to the evaluation of seawater density using the International Equation of State of 1980 (see Appendix 5. Properties of Seawater). These definitions and the associated algorithms have served the field of oceanography very well for almost 30 years. Recently there have been some advances in the science of determining ‘salinity’ and in evaluating many


30


0.35

0.3

0.25

2

Frequency

n 0.2

0.15

n (labeled Levitus)

0.1 n n (labeled Levitus)

0.05

0 −15 −14 −13 −12 −11 −10

−9

−8

−7

−6

−5

−4

−3

−2

−1

0

log (D) Figure 5 Frequency distribution of the logarithm of the false diapycnal diffusivity DVS arising from mixing along various coordinate surfaces.

thermodynamic properties of seawater. A working group established jointly by International Association for the Physical Sciences of the Oceans (IAPSO) and Scientific Committee on Oceanic Research (SCOR) in 2006 is examining new definitions and releases for salinity and for thermodynamic properties such as density. This is not the appropriate place to describe these developments in detail but the two recent developments that are driving this work are briefly outlined. The first development is the realization that all the accurate measurements of various thermodynamic properties of seawater (e.g., specific volume, heat capacity, sound speed) can be incorporated into a single thermodynamic function called a Gibbs function. All the thermodynamic properties of interest are obtained by simple mathematical manipulations (usually differentiation) of this Gibbs function. The salient advantages of using a Gibbs function to describe thermodynamic properties are that (1) all the derived thermodynamic properties are automatically consistent with each other and (2) the accurate data of all measurable thermodynamic properties contribute to the accuracy of many other derived thermodynamic quantities. The second development that is driving the redefinition of the thermodynamics of seawater is the realization that the differences in the composition of seawater that exist between the major ocean basins can and should be incorporated into a more accurate definition of salinity. The thermodynamic properties

of seawater depend on the ‘absolute salinity’, namely the number of grams of dissolved material in a kilogram of seawater. For seawater of standard composition, ‘absolute salinity’ is larger than ‘practical salinity’ by about 0.47%. In addition, for a given practical salinity, the absolute salinity in some ocean basins can differ by 0.02 g kg1. We expect to soon be able to estimate this spatial variation in the difference between absolute salinity and practical salinity. Since the thermodynamic properties of seawater are most consistently represented by a Gibbs function, and since this Gibbs function is best regarded as a function of absolute salinity, temperature, and pressure (rather than in terms of practical salinity), the time seems ripe to adopt a thermodynamic definition of seawater based on the Gibbs function and in terms of absolute salinity.

Summary The dynamical features in the ocean that correspond to the weather systems in the atmosphere are the socalled mesoscale eddies that typically have radii of 100 km and cause lateral mixing along neutral tangent planes with a lateral diffusivity of about 1000 m2 s1. In contrast, the diffusivity of tracers in the ocean in the diapycnal direction is 8 orders of magnitude less than this. Because of this wide disparity between the rates at which the ocean mixes



along and across ‘isopycnals’, it is important to be able to accurately determine the correct slopes of the surfaces in which the large mixing occurs so as not to have this large lateral diffusivity leak into the diapycnal direction, thus causing unwanted false diapycnal mixing. This article began by summarizing the reasons why neutral surfaces do not exist in the ocean: the neutral tangent planes that locally define the direction of neutral mixing are well defined but these planes do not link up to form a mathematically welldefined surface. Rather, oceanic motion occurs along neutral trajectories that are helical in nature and have the ability to cause a mean diapycnal advection. This diapycnal advection process is independent of the amount of diapycnal mixing present in the ocean. For the purpose of defining approximately neutral surfaces, the path-dependent nature of neutral surfaces is actually a small issue compared with the differences that exist between the hemispheres in potential density. The patched potential density procedure of Red and Lynn is relatively accurate although it is a time-consuming procedure for forming approximately neutral surfaces. The neutral density variable of Jackett and McDougall also builds in the spatial variations of water mass properties that occur in the present ocean and this software provides a practical way of forming approximately neutral surfaces in today’s ocean. The equation of state (i.e., the expression for the density of seawater in terms of the various thermodynamic parameters, commonly salinity, temperature, and pressure) was defined by UNESCO in 1981 and has served oceanography well for almost 30 years. Efforts are now underway to provide a new internationally accepted equation of state in the next year or so, this time it being based on a Gibbs potential for seawater.

See also Fluid Dynamics, Introduction, and Laboratory Experiments. Ocean Thermal Energy Conversion (OTEC). Satellite Remote Sensing of Sea Surface Temperatures. Sea Surface Exchanges of Momentum, Heat, and Fresh Water Determined by Satellite Remote Sensing. Tidal Energy. Water Types and Water Masses. Wave Energy.

31

120 g kg1. Deep Sea Research I (doi:10.1016/ j.dsr.2008.07.004). Jackett DR and McDougall TJ (1997) A neutral density variable for the world’s oceans. Journal of Physical Oceanography 27: 237--263. Koltermann KP, Gouretski V, and Jancke K (2004) Hydrographic Atlas of the World Ocean Circulation Experiment (WOCE), Vol. 3: Atlantic Ocean. http:// www.bsh.de/aktdat/mk/AIMS (accessed Mar. 2008). McDougall TJ (1987) Neutral surfaces. Journal of Physical Oceanography 17: 1950--1964. McDougall TJ (1987) Thermobaricity, cabbeling, and water-mass conversion. Journal of Geophysical Research 92: 5448--5464. McDougall TJ (1988) Neutral-surface potential vorticity. Progress in Oceanography 20: 185--221. McDougall TJ and Jackett DR (1988) On the helical nature of neutral trajectories in the ocean. Progress in Oceanography 20: 153--183. McDougall TJ and Jackett DR (2005) An assessment of orthobaric density in the global ocean. Journal of Physical Oceanography 35: 2054--2075. McDougall TJ and Jackett DR (2005) The material derivative of neutral density. Journal of Marine Research 63: 159--185. McDougall TJ and Jackett DR (2007) The thinness of the ocean in S Y p space and the implications for mean diapycnal advection. Journal of Physical Oceanography 37: 1714--1732. Millero FJ, Feistel R, Wright DG, and McDougall TJ (2008) The composition of standard seawater and the definition of the reference-composition salinity scale. Deep Sea Research I 55: 5072 (doi:10.1016/j.dsr. 2007.10.001). Phillips HB (1956) Vector Analysis, 236pp. New York: Wiley. Reid JL and Lynn RJ (1971) On the influence of the Norwegian–Greenland and Weddell seas upon the bottom waters of the Indian and Pacific oceans. Deep Sea Research and Oceanographic Abstracts 18: 1063--1088. UNESCO (1981) The Practical Salinity Scale 1978 and the International Equation of State of Seawater 1980. UNESCO Technical Papers in Marine Science 36. Wu¨st G (1933) Das Bodenwasser und die Gliedernng der Atlantischen Tiefsee. Wiss. Ergebn, Dtsch. Atlant. Exped. Meteor 6(I): 1--107

Relevant Websites http://www.cmar.csiro.au – Neutral Density, Version 3.1: Jan. 1997, CSIRO Marine and Atmospheric Research.

Further Reading Feistel R (2008) A Gibbs function for seawater thermodynamics for 6 1C to 80 1C and salinity up to


NITROGEN CYCLE D. M. Karl, University of Hawaii at Manoa, Honolulu, HI, USA A. F. Michaels, University of Southern California, Los Angeles, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1876–1884, & 2001, Elsevier Ltd.

Introduction The continued production of organic matter in the sea requires the availability of the many building blocks of life, including essential major elements such as carbon (C), nitrogen (N), and phosphorus (P); essential minor elements such as iron, zinc, and cobalt; and, for many marine organisms, essential trace organic nutrients that they cannot manufacture themselves (e.g., amino acids and vitamins). These required nutrients have diverse structural and metabolic function and, by definition, marine organisms cannot survive in their absence. The marine nitrogen cycle is part of the much larger and interconnected hydrosphere–lithosphere–atmosphere–biosphere nitrogen cycle of the Earth. Furthermore, the oceanic cycles of carbon, nitrogen, and phosphorus are inextricably linked together through the production and remineralization of organic matter, especially near surface ocean phytoplankton production. This coordinated web of major bioelements can be viewed as the nutrient ‘super-cycle.’ The dominant form of nitrogen in the sea is dissolved gaseous dinitrogen (N2) which accounts for more than 95% of the total nitrogen inventory. However, the relative stability of the triple bond of N2 renders this form nearly inert. Although N2 can serve as a biologically-available nitrogen source for specialized N2 fixing microorganisms, these organisms are relatively rare in most marine ecosystems. Consequently, chemically ‘fixed’ or ‘reactive’ nitrogen compounds such as nitrate (NO3 ), nitrate (NO2 ), ammonium (NH4 þ ), and dissolved and particulate organic nitrogen (DON/ PON) serve as the principal sources of nitrogen to sustain biological processes. For more than a century, oceanographers have been concerned with the identification of growth-and production-rate limiting factors. This has stimulated investigations of the marine nitrogen cycle including both inventory determinations and pathways and controls of nitrogen transformations from one form

32

to another. Contemporaneous ocean investigations have documented an inextricable link between nitrogen and phosphorus cycles, as well as the importance of trace inorganic nutrients. It now appears almost certain that nitrogen is only one of several key elements for life in the sea, neither more nor less important than the others. Although the basic features of the marine nitrogen cycle were established nearly 50 years ago, new pathways and novel microorganisms continue to be discovered. Consequently, our conceptual view of the nitrogen cycle is a flexible framework, always poised for readjustment.

Methods and Units The analytical determinations of the various dissolved and particulate forms of nitrogen in the sea rely largely on methods that have been in routine use for several decades. Determinations of NO3 , NO2 , and NH4 þ generally employ automated shipboard, colorimetric assays, although surface waters of open ocean ecosystems demand the use of modern highsensitivity chemiluminescence and fluorometric detection systems. PON is measured by high-temperature combustion followed by chromatographic detection of the by-product (N2), usually with a commercial C–N analyzer. Total dissolved nitrogen (TDN) determination employs sample oxidation, by chemical or photolytic means, followed by measurement of NO3 . DON is calculated as the difference between TDN and the measured dissolved, reactive inorganic forms of N (NO3 , NO2 , NH4 þ ) present in the original sample. Gaseous forms of nitrogen, including N2, nitrous oxide (N2O), and nitric oxide (NO) are generally measured by gas chromatography. Nitrogen exists naturally as two stable isotopes, 14 N (99.6% by atoms) and 15N (0.4% by atoms). These isotopes can be used to study the marine nitrogen cycle by examination of natural variations in the 14N/15N ratio, or by the addition of specific tracers that are artificially enriched in 15N. Most studies of oceanic nitrogen inventories or transformations use either molar or mass units; conversion between the two is straightforward (1 mole N ¼ 14 g N, keeping in mind that the molecular weight of N2 gas is 28).

Components of the Marine Nitrogen Cycle The systematic transformation of one form of nitrogen to another is referred to as the nitrogen cycle


NITROGEN CYCLE

(Figure 1). In the sea, the nitrogen cycle revolves around the metabolic activities of selected microorganisms and it is reasonable to refer to it as the microbial nitrogen cycle because it depends on bacteria (Table 1). During most of these nitrogen transformations there is a gain or loss of electrons and, therefore, a change in the oxidation state of nitrogen from the most oxidized form, NO3 ( þ 5), to the most reduced form, NH4 þ ( 3). Transformations in the nitrogen cycle are generally either energy-requiring (reductions) or energy-yielding (oxidations). The gaseous forms of nitrogen in the

33

surface ocean can freely exchange with the atmosphere, so there is a constant flux of nitrogen between these two pools. The natural, stepwise process for the regeneration of NO3 from PON can be reproduced in a simple ‘decomposition experiment’ in an enclosed bottle of sea water (Figure 2). During a 3-month incubation period, the nitrogen contained in particulate matter is first released as NH4þ (the process of ammonification), then transformed to NO2 (first step of nitrification), and finally, and quantitatively, to NO3 (the second step of nitrification). These transformations Atmosphere Ocean

ion

N2

fix

at

N2O

DENITRIFICATION

tro

ge

n

NH3

Ni

NO −

−

NO3 reduction

NO2 reduction

Biosynthesis

Ammonification

NO3

NO2 Ammonium oxidation

Nitrite oxidation

NITRIFICATION Export

Organic N _3

−

−

+

NH4

Import _2

_1

0 2 3 1 Oxidation state of nitrogen

4

5

Figure 1 Schematic representation of the various transformations from one form of nitrogen to another that compose the marine nitrogen cycle. Shown at the bottom is the oxidation state of nitrogen for each of the components. Most transformations are microbiological and most involve nitrogen reduction or oxidation. (Adapted from Capone, ch. 14 of Rogers and Whitman (1991).)

Table 1

Marine nitrogen cycle Credits

Process

Bacteria

Phytoplanktona

Zooplankton/Fishb

NH4þ production from DON/PON (ammonification) NH4þ /NO2 /NO3 /DON assimilation PON ingestion NH4 þ -NO2 (nitrification, step 1) NO2 -NO3 /N2O (nitrification, step 2) NO3 /NO2 -N2/N2O (denitrification) N2- NH4 þ /organic N (N2 fixation)

þ þ þ þ þ þ

þ þ þ þ

þ þ

a

Phytoplankton – eukaryotic phytoplankton. Zooplankton – including protozoans and metazoans. Abbreviations: NH4 þ , ammonium; NO2 , nitrite; NO3 , nitrate; N2O, nitrous oxide; N2, dinitrogen; DON, dissolved organic N; PON, particulate organic N; organic N, DON and PON. b


Nitrogen (arbitrary units)

34

NITROGEN CYCLE

+

NH4

−

−

NO2

NO3

200 100

PON

0 0

1

2

and abundant planktonic cyanobacteria that coexist in tropical and subtropical marine habitats have devised alternate metabolic strategies: Synechococcus prefers NO3 and Prochlorococcus prefers NH4þ. In fact, Prochlorococcus cannot reduce NO3 to NH4þ, presumably because the critical enzyme systems are absent.

3

Time (months) Figure 2 Stepwise decomposition of particulate organic nitrogen (PON) in arbitrary units versus time during a dark incubation. Nitrogen is transformed, first to NH4þ by bacterial ammonification and finally to NO2 and NO3 by the two-step process of bacterial nitrification. These same processes are responsible for the global ocean formation of NO3 in the deep sea. These data are the idealized results of pioneering nitrogen cycle investigators, T. von Brand and N. Rakestraw, who unraveled these processes more than 50 years ago.

are almost exclusively a result of the metabolic activities of bacteria. This set of regeneration reactions is vital to the nitrogen cycle, and since most deep water nitrogen (excluding N2) is in the form of NO3, bacterial nitrification must be a very important process (see Nitrogen Distributions in the Sea, below). Nitrogen Assimilation

Several forms of nitrogen can be directly transported across cell membranes and assimilated into new cellular materials as required for biosynthesis and growth. Most microorganisms readily transport NH4þ, NO2, NO3, and selected DON compounds such as amino acids, urea, and nucleic acid bases. By comparison, the ability to utilize N2 as a nitrogen source for biosynthesis is restricted to a very few species of specialized microbes. Many protozoans, including both photosynthetic and heterotrophic species, and all metazoans obtain nitrogen primarily by ingestion of PON. Once inside the cell or organism, nitrogen is digested and, if necessary, reduced to NH4þ. If oxidized compounds such as NO3 or NO2 are utilized, cellular energy must be invested to reduce these substrates to ammonium for incorporation into organic matter. The process of reduction of NO3 (or NO2) for the purpose of cell growth is referred to as assimilatory nitrogen (NO3/NO2) reduction and most microorganisms, both bacteria and phytoplankton, possess this metabolic capability (Table 1). In theory, there should be a metabolic preference for NH4þ over either NO3 or NO2, based strictly on energetic considerations. However, it should be emphasized that preferential utilization of NH4þ does not always occur. For example, two closely related

Nitrification

As nitrogen is oxidized from NH4þ through NO2 to NO3, energy is released (Figure 1), a portion of which can be coupled to the reduction of carbon dioxide (CO2) to organic matter (CH2O) by nitrifying bacteria. These specialized bacteria, one group capable only of the oxidation of NH4þ to NO2 and the second capable only of the oxidation of NO2 to NO3, are termed ‘chemolithoautotrophic’ because they can fix CO2 in the dark at the expense of chemical energy. Other related chemolithoautotrophs can oxidize reduced sulfur compounds, and this pathway of organic matter production has been hypothesized as the basis for life at deep-sea hydrothermal vents. It is essential to emphasize an important ecological aspect of NH4þ/NO2 chemolithoautotrophy. First, the oxidation of NH4þ to NO2 and of NO2 to NO3 usually requires oxygen and these processes are ultimately coupled to the photosynthetic production of oxygen in the surface water. Second, the continued formation of reduced nitrogen, in the form of NH4þ or organic nitrogen, is also dependent, ultimately, on photosynthesis. In this regard the CO2 reduced via this ‘autotrophic’ pathway must be considered secondary, not primary, production from an ecological energetics perspective. Marine nitrifying bacteria, especially the NO2 oxidizers are ubiquitous in the world ocean and key to the regeneration of NO3, which dominates waters below the well-illuminated, euphotic zone. However they are never very abundant and, at least for those species in culture, grow very slowly. Certain heterotrophic bacteria can also oxidize NH4þ to both NO2 and NO3 during metabolism of preformed organic matter. However, very little is known about the potential for ‘heterotrophic nitrification’ in the sea. Denitrification

Under conditions of reduced oxygen (O2) availability, selected species of marine bacteria can use NO3 as a terminal acceptor for electrons during metabolism, a process termed NO3 respiration or dissimilatory NO3 reduction. This process allows microorganisms to utilize organic matter in low-O2


NITROGEN CYCLE

or anoxic habitats with only a slight loss of efficiency relative to O2-based metabolism. A majority of marine bacteria have the ability for NO3 respiration under the appropriate environmental conditions (Table 1). Potential by-products of NO 3 respiration are NO2, N2, and N2O; if a gas is formed (N2/N2O) then the process is termed denitrification because the net effect is to remove bioavailable nitrogen from the local environment. The total rate of denitrification is generally limited by the availability of NO3, and a continued supply of NO3 via nitrification is dependent upon the availability of NH4þ and free O2. Consequently denitrification typically occurs at boundaries between low-O2 and anoxic conditions where the supply of NH4þ from the anoxic zone sustains a high rate of NO3 production via nitrification to fuel- sustained NO3 respiration and denitrification. Recently a new group of microorganisms has been isolated that are capable of simultaneously using both O2 and NO3/NO2 as terminal electron acceptors. This process is termed ‘aerobic denitrification.’ Likewise, there are exceptional microorganisms that are able to carry out anaerobic nitrification (oxidation of NH4þ in the absence of O2). It appears difficult to establish any hard-and-fast rules regarding marine nitrogen cycle processes. N2 Fixation

The ability to use N2 as a growth substrate is restricted to a relatively small group of microorganisms. Open ocean ecosystems that are chronically depleted in fixed nitrogen would appear to be ideal habitats for the proliferation of N2-fixing microorganisms. However, the enzyme that is required for reduction of N2 to NHþ 4 is also inhibited by O2, so specialized structural, molecular, and behavioral adaptations have evolved to promote oceanic N2 fixation. Fixation of molecular nitrogen in the open ocean may also be limited by the availability of iron, which is an essential cofactor for the N2 reduction enzyme system. Changes in iron loading are caused by climate variations, in particular the areal extent of global deserts, by the intensity of atmospheric circulation, and more recently by changes in land use practices. Conversion of deserts into irrigated croplands may cause a change in the pattern and intensity of dust production and, therefore, of iron transport to the sea. Humanity is also altering the global nitrogen cycle by enhancing the fixation of N2 by the manufacturing of fertilizer. At the present time, the industrial fixation of N2 is approximately equivalent to the pre-industrial, natural N2 fixation rate.

35

Eventually some of this artificially fixed N2 will make its way to the sea, and this may lead to a perturbation in the natural nitrogen cycle. On a global scale and over relatively long timescales, the total rate of N2 fixation is more or less in balance with total denitrification, so that the nitrogen cycle is mass-balanced. However, significant net deficits or excesses can be observed locally or even on ocean basin space scales and on decade to century timescales. These nitrogen imbalances may impact the global carbon and phosphorus cycles as well, including the net balance of CO2 between the ocean and the atmosphere.

Nitrogen Distributions in the Sea Required growth nutrients, like nitrogen, typically have uneven distributions in the open sea, with deficits in areas where net organic matter is produced and exported, and excesses in areas where organic matter is decomposed. For example, surface ocean NO3 distributions in the Pacific basin reveal a coherent pattern with excess NO3 in high latitudes, especially in the Southern Ocean (south of 601 S), and along the Equator (especially east of the dateline), and generally depleted NO3 concentrations in the middle latitudes of both hemispheres (Figure 3). These distributions are a result of the balance between NO3 supply mostly by ocean mixing and NO3 demand or net photosynthesis. The very large NO3 inventory in the surface waters of the Southern Ocean implies that factors other than fixed nitrogen availability control photosynthesis in these regions. It has been hypothesized that the availability of iron is key in this and perhaps other regions of the open ocean. The much smaller but very distinctive band of elevated NO3 along the Equator is the result of upwelling of NO3-enriched waters from depth to the surface. This process has a large seasonal and, especially, interannual variability, and it is almost absent during El Ninõ conditions. Excluding these high-latitude and equatorial regions, the remainder of the surface waters of the North and South Pacific Oceans from about 401N to 401S are relatively depleted in NO3. In fact surface (0–50 m) NO3 concentrations in the North Pacific subtropical gyre near Hawaii are typically below 0.01 mmol l1 (Figure 4). Within the upper 200 m, the major pools of fixed nitrogen (e.g., NO3, DON, and PON) have different depth distributions. In the sunlit surface zone, NO3 is removed to sustain organic matter production and export. Beneath 100 m, there is a steep concentration versus depth gradient (referred to as the nutricline), which reaches a


36

NITROGEN CYCLE

70˚N 60˚N 50˚N 40˚N 30˚N 20˚N

Latitude

10˚N 0˚ 10˚S 20˚S 30˚S 40˚S 50˚S 60˚S 70˚S 80˚S 105˚E 120˚E 135˚E 150˚E

165˚E 180˚E 165˚W 150˚W 135˚W 120˚W 105˚W Longitude

90˚W

75˚W

60˚W

Figure 3 Mean annual NO3 concentration (mmol l1) at the sea surface for samples collected in the Pacific Ocean basin and Pacific sector of the Southern Ocean. (From Conkright et al. 1998.)

maximum of about 40–45 it mmol l1 at about 1000 m in the North Pacific Ocean. PON concentration is greatest in the near-surface waters where the production of organic matter via photosynthesis is highest (Figure 4). PON includes both living (biomass) and nonliving (detrital) components; usually biomass nitrogen is less than 50% of the total PON in near-surface waters, and less than 10% beneath the euphotic zone (4150 m). DON concentration is also highest in the euphotic zone (B5– 6 mmol l1) and decreases systematically with depth to a minimum of 2–3 mmol l1 at 800–1000 m. The main sources for DON in the surface ocean are the combined processes of excretion, grazing, death, and cell lysis. Consequently, DON is a complex mixture of cell-derived biochemicals; at present, less than 20% of the total DON has been chemically characterized. Dissolved N2 (not shown) is always high (B800 mmol l1) and increases systematically with depth. The major controls of N2 concentration are temperature and salinity, which together determine gas solubility. Marine life has little impact on N2 distributions in the open sea even though some microorganisms can utilize N2 as a growth substrate and others can produce N2 as a metabolic by-

product. These transformations are simply too small to significantly impact the large N2 inventories in most regions of the world ocean. Another important feature of the global distribution of NO3 is the regional variability in the deep water inventory (Figure 5). Deep ocean circulation can be viewed as a conveyor-belt-like flow, with the youngest waters in the North Atlantic and the oldest in the North Pacific. The transit time is in excess of 1000 y, during which time NO3 is continuously regenerated from exported particulate and dissolved organic matter via coupled ammonification and nitrification (Figure 2). Consequently, the deep Pacific Ocean has nearly twice as much NO3 as comparable depths in the North Atlantic (Figure 5). Nitrous Oxide Production

Nitrous oxide (N2O) is a potent greenhouse gas that has also been implicated in stratospheric ozone depletion. The atmospheric inventory of N2O is presently increasing, so there is a renewed interest in the marine ecosystem as a potential source of N2O. Nitrous oxide is a trace gas in sea water, with typical concentrations ranging from 5 to 50 nmol l1.


NITROGEN CYCLE

_1

_1

Nitrate (μmol l )

37

_1

DON (μmol l )

PON (μmol l )

0 20 40 60

Depth (m)

80 100 120 140 160 180 200 0

0.5

1.0

1.5 4.5

5.0

6.0 0

5.5

0.1

0.2

0.3

0.4

Figure 4 Average concentrations (mmol l1) of NO3, DON, and PON versus water depth for samples collected in the upper 200 m of the water column at Sta. ALOHA (22.751N, 158.01W). These field data are from the Hawaii Ocean Time-series program and are available at http://hahana.soest.edu/hot_jgofs.html).

0

Primary Nitrite (NO 2 ) Maximum

An interesting, almost cosmopolitan feature of the world ocean is the existence of a primary NO2

1000

N. Pacific

N. Atlantic 2000

Depth (m)

Concentrations of N2O in oceanic surface waters are generally in slight excess of air saturation, implying both a local source and a sustained ocean-to-atmosphere flux. Typically there is a mid-water (500– 1000 m) peak in N2O concentration that coincides with the dissolved oxygen minimum. At these intermediate water depths, N2O can exceed 300% saturation relative to atmospheric equilibrium. The two most probable sources of N2O in the ocean are bacterial nitrification and bacterial denitrification, although to date it has been difficult to quantify the relative contribution of each pathway for a given habitat. Isotopic measurements of nitrogen and oxygen could prove invaluable in this regard. Because the various nitrogen cycle reactions are interconnected, changes in the rate of any one process will likely have an impact on the others. For example, selection for N2-fixing organisms as a consequence of dust deposition or deliberate iron fertilization would increase the local NHþ 4 inventory and lead to accelerated rates of nitrification and, hence, enhanced N2O production in the surface ocean and flux to the atmosphere.

3000

4000

5000

6000 0

5

10

15

20

25

30

35

40

45

_1

Nitrate (μmol l ) Figure 5 Nitrate concentrations (mmol l1) versus water depth at two contrasting stations located in the North Atlantic (31.81N, 50.91W) and North Pacific (301N, 160.31W) Oceans. These data were collected in the 1970s during the worldwide GEOSECS expedition, stations #119 and #212, respectively.


38

NITROGEN CYCLE

maximum (PNM) near the base of the euphotic zone (B100–150 m; Figure 6). Nitrite is a key intermediate between NO3 and NH4þ, so there are several potential pathways, both oxidative and reductive, that might lead to its accumulation in sea water. First, phototrophic organisms growing on NO3 may partially reduce the substrate to NO2 as the first, and least energy-consuming, step in the assimilatory NO 3 reduction pathway. However, the next step, reduction of NO2 to NH4þ, requires a substantial amount of energy, so when energy is scarce (e.g., light limitation) NO2 accumulates inside the cells. Because NO2 is the salt of a weak acid, nitrous acid (HNO2) forms in the slightly acidic intracellular environment, diffuses out of the cell and ionizes to form NO2 in the alkaline sea water. This NO3-NO2 phytoplankton pump, under the control of light intensity, could provide a source of NO 2 necessary to create and maintain the PNM. Alternatively, local regeneration of dissolved and particulate organic matter could produce NH4þ (via ammonification) that is partially oxidized in place to produce a relative excess of NO2 (the first step of

nitrification). Kinetic controls on this process would be rates of NH4þ production and NO2 oxidation to NO3 (the second and final step in nitrification). Sunlight, even at very low levels, appears to disrupt the normal coupling between NO2 production and NO2 oxidation, in favor of NO2 accumulation. Finally, it is possible, though perhaps less likely, that NO3 respiration (terminating at NO2), followed by excretion of NO2 (into the surrounding sea water might also contribute to the accumulation of NO2) near the base of the euphotic zone. Because the global ocean at the depth of the PNM is characteristically well-oxygenated, one would need to invoke microenvironments like animal guts or large particles as the habitats for this nitrogen cycle pathway. The use of 15N-labeled substrates, selective metabolic inhibitors, and other experimental manipulations provides an opportunity for direct assessment of the role of each of these potential processes. In all likelihood, more than one of these processes contributes to the observed PNM. Whatever the cause, light appears to be an important determinant that might explain the relative position, with regard to depth, of this global feature.

0

Nitrogen Cycle and Ocean Productivity

Light

40

Depth (m)

80 Chlorophyll-a

120 −

− NO2

NO3

160

200 Concentration Figure 6 Schematic representation of the depth distributions of sunlight, chlorophyll-a, NO 2 , and NO3 for a representative station in the subtropical North Pacific Ocean showing the relationship of the primary NO 2 maximum (PNM) zone (shaded) to the other environmental variables. (Modified from J. E. Dore, Microbial nitrification in the marine euphotic zone, Ph.D. Dissertation, University of Hawaii, redrawn with permission of the author.)

Because nitrogen transformations include both the formation and decomposition of organic matter, much of the nitrogen used in photosynthesis is locally recycled back to NH4þ or NO3 to support another pass through the cycle. The net removal of nitrogen in particulate, dissolved, or gaseous form can cause the cycle to slow down or even terminate unless new nitrogen is imported from an external source. A unifying concept in the study of nutrient dynamics in the sea is the ‘new’ versus ‘regenerated’ nitrogen dichotomy (Figure 7). New nitrogen is imported from surrounding regions (e.g., NO3 injection from below) or locally created (e.g., NH4þ/organic N from N2 fixation). Regenerated nitrogen is locally recycled (e.g., NH4þ from ammonification, NO2 / NO3 from nitrification, or DON from grazing or cell lysis). Under steady-state conditions, the amount of new nitrogen entering an ecosystem will determine the total amount that can be exported without the system running down. In shallow, coastal regions runoff from land or movement upward from the sediments are potentially major sources of NH4þ, NO3 and DON for water column processes. In certain regions, atmospheric deposition (both wet and dry) may also supply bioavailable nitrogen to the system. However, in


NITROGEN CYCLE

39

N2

Gas exchange

N2

Atmosphere

Euphotic zone

N2 fixation PRIMARY PRODUCERS

CONSUMERS

−

NO3

+

NH4

Recycling DON Nitracline Export

Export Ammonification

Diffusion and advection

DON and PON

+

NH4

DON and PON Aphotic zone

−

NO3

Nitrification Figure 7 Schematic representation of the major pools and transformations/fluxes of nitrogen in a typical open ocean ecosystem. New sources of bioavailable N (NO3 and N2 in this presentation) continuously resupply nitrogen that is lost via DON and PON export. These interactions and ocean processes form the conceptual framework for the ‘new’ versus ‘regenerated’ paradigm of nitrogen dynamics in the sea that was originally proposed by R. Dugdale and J. Goering.

most open ocean environments, new sources of nitrogen required to balance the net losses from the euphotic zone are restricted to upward diffusion or mixing of NO3 from deep water and to local fixation of N2 gas. In a balanced steady state, the importation rate of new sources of bioavailable nitrogen will constrain the export of nitrogen (including fisheries production and harvesting). If all other required nutrients are available, export-rich ecosystems are those characterized by high bioavailable nitrogen loading such as coastal and open ocean upwelling regions. These are also the major regions of fish production in the sea.

See also Atmospheric Input of Pollutants. Nitrogen Isotopes in the Ocean. Phosphorus Cycle. Primary Production Processes.

Further Reading Carpenter EJ and Capone DG (eds.) (1983) Nitrogen in the Marine Environment. New York: Academic Press. Conkright ME, O’Brien TD, Levitus S, et al. (1998) NOAA Atlas NESDIS 37, WORLD OCEAN ATLAS 1998, vol. II: Nutrients and Chlorophyll of the Pacific Ocean. Washington, DC: US Department of Commerce. Harvey HW (1966) The Chemistry and Fertility of Sea Waters. London: Cambridge University Press. Kirchman DL (ed.) (2000) Microbial Ecology of the Oceans. New York: Wiley-Liss. Rogers JE and Whitman WB (eds.) (1991) Microbial Production and Consumption of Greenhouse Gases: Methane, Nitrogen Oxides, and Halomethanes. Washington, DC: American Society for Microbiology. Schlesinger WH (1997) Biogeochemistry: An Analysis of Global Change. San Diego: Academic Press. Wada E and Hattori A (1991) Nitrogen in the Sea: Forms, Abundances, and Rate Processes. Boca Raton, FL: CRC Press.


NITROGEN ISOTOPES IN THE OCEAN D. M. Sigman and K. L. Karsh, Princeton University, Princeton, NJ, USA K. L. Casciotti, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Nitrogen has two stable isotopes, 14N and 15N (atomic masses of 14 and 15, respectively). 14N is the more abundant of the two, comprising 99.63% of the nitrogen found in nature. Physical, chemical, and biological processes discriminate between the two isotopes, leading to subtle but measurable differences in the ratio of 15N to 14N among different forms of nitrogen found in the marine environment. Nitrogen is a central component of marine biomass and one of the major nutrients required by all phytoplankton. In this sense, biologically available (or ‘fixed’, i.e., non-N2) N is representative of the fundamental patterns of biogeochemical cycling in the ocean. However, N differs from other nutrients in that its oceanic sources and sinks are dominantly internal and biological, with marine N2 fixation supplying much of the fixed N in the ocean and marine denitrification removing it. The N isotopes provide a means of studying both the input/output budget of oceanic fixed N and its cycling within the ocean. In this overview, we outline the isotope systematics of N cycle processes and their impacts on the isotopic composition of the major N reservoirs in the ocean. This information provides a starting point for considering the wide range of questions in ocean sciences to which the N isotopes can be applied.

Terms and Units Mass spectrometry can measure precisely the ratio of the N isotopes relative to a N reference containing a constant isotopic ratio. The universal reference for N isotopes is atmospheric N2, with an 15N/14N ratio of 0.367 65%70.000 81%. Natural samples exhibit small deviations from the standard ratio, which are expressed in d notation (in units of per mil, %):

15

d Nð%Þ ¼

40

ð15 N=14 NÞsample ð15 N=14 NÞstandard

! 1

1000 ½1

In this notation, the d15N of atmospheric N2 is 0%. Special terms are also used to characterize the amplitude of isotopic fractionation caused by a given process. Isotope fractionation results from both equilibrium processes (‘equilibrium fractionation’) and unidirectional reactions (‘kinetic fractionation’). Nitrogen isotope variations in the ocean are typically dominated by kinetic fractionation associated with the conversions of N from one form to another. The kinetic isotope effect, e, of a given reaction is defined by the ratio of rates with which the two N isotopes are converted from reactant to product: eð%Þ ¼ ð14 k=15 k 1Þ 1000

½2

where 14k and 15k are the rate coefficients of the reaction for 14N- and 15N-containing reactant, respectively. For e{1000%, e is approximated by the difference in d15N between the reactant and its instantaneous product. That is, if a reaction has an e of 5%, then the d15N of the product N generated at any given time will be B5% lower than the d15N of the reactant N at that time.

Measurements The isotopic analysis of N relies on the generation of a stable gas as the analyte for isotope ratio mass spectrometry. Online combustion to N2 is currently the standard method for the preparation of a N sample for isotopic analysis. With ‘off-the-shelf’ technology, a typical sample size requirement is 1–2 mmol N per analysis. Gas chromatography followed by combustion to N2 is improving as a technique for specific organic compounds, amino acids in particular, although the polarity of many N compounds remains a challenge. Liquid chromatography is also being explored. There are standard methods of collection for most bulk forms of particulate N (PN) in the ocean. Shallow and deep samples of suspended PN are filtered onto glass fiber filters. Sinking PN is collected by sediment traps. Zooplankton can be picked from filtered samples or net tows, and particulates can be separated into size classes. In the case of dissolved forms of N, the species of interest must be converted selectively to a gas or other extractable form for collection. Since the 1970s, the d15N values of marine nitrate (NO3 ), nitrite (NO2 ), and ammonium (NH4 þ ) have been analyzed by conversion to ammonia gas and collection of the cationic ammonium form for


41

NITROGEN ISOTOPES IN THE OCEAN

Models Two simple models, the ‘Rayleigh’ model and the ‘steady-state’ model, are frequently used to interpret N isotope data from the ocean. In both of these models, the degree of consumption of the reactant N pool (f ) is a key parameter, and the d15N of the initial reactant N pool (d15 Ninitial ) and kinetic isotope effect (e) are the two central isotopic parameters. If a transformation proceeds with a constant isotope effect and if the reactant and product N pools are neither replenished nor lost from the system during the progress of the transformation, then the process can be described in terms of Rayleigh fractionation kinetics, which define the isotopic variation of the reactant N pool (d15 Nreactant ; eqn [3]), the instantaneously generated product N (d15 Ninstantaneous ; eqn [4]), and the integrated product N pool (d15 Nintegrated ; eqn [5]) as a given reservoir of reactant N is consumed (Figure 1): 15

15

d Nreactant ¼ d Ninitial eflnðf Þg 15

15

d Ninstantaneous ¼ d Nreactant e

½3

25

Rayleigh (closed system): 15Nreactant 15Ninstantaneous

20 15N (% vs. air)

subsequent conversion to N2 (often referred to as the ammonia ‘distillation’ and ‘diffusion’ methods). Recently, more sensitive isotope analysis methods (requiring only 5–10 nmol of N per analysis) have been developed for nitrate and nitrite in which these species are converted to nitrous oxide (N2O), followed by isotopic analysis of this gas (the ‘bacterial’ or ‘denitrifier’ method and the ‘chemical’ or ‘azide’ method). The N2O produced by these methods (or naturally occuring N2O) is analyzed by a purge and trap system, followed by gas chromatography and isotope ratio mass spectrometry. The N2O-based methods also allow for oxygen isotope analysis of nitrate and nitrite, a measurement not previously possible in seawater. In addition, they provide a cornerstone for isotopic analysis of other dissolved forms of N, such as dissolved organic N (DON) and ammonium (NH4 þ ), which can be converted to nitrate and/or nitrite. With respect to dissolved gases, methods of collection and isotopic analysis have been developed for N2 and N2O, with recent progress on isotopomer analysis of N2O (i.e., distinguishing 15N14N16O from 14N15N16O).

15Nintegrated 15

Steady-state (open system): 15Nreactant 15Nproduct

10 5 0 0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Fraction of reactant remaining

1

Figure 1 The d15N of reactant and product N pools of a single unidirectional reaction as a function of the fraction of the initial reactant supply that is left unconsumed, for two different models of reactant supply and consumption, following the approximate equations given in the text. The Rayleigh model (black lines) applies when a closed pool of reactant N is consumed. The steady-state model (gray lines) applies when reactant N is supplied continuously. The same isotopic parameters, an isotope effect (e) of 5% and a d15N of 5% for the initial reactant supply, are used for both the Rayleigh and steady-state models. e is approximately equal to the isotopic difference between reactant N and its product (the instantaneous product in the case of the Rayleigh model).

forms of the full expressions. They are typically adequate, but their error is greater for higher consumption (lower f ) and higher e. The Rayleigh model is often used to describe events in the ocean, such as the uptake of nitrate by phytoplankton during a bloom in a stratified surface layer. The end-member alternative to the Rayleigh model is the steady-state model, in which reactant N is continuously supplied and partially consumed, with residual reactant N being exported at a steady-state rate such that the gross supply of reactant N equals the sum of the product N and the residual reactant N exported. In this case, the following approximate expessions apply to the reactant N pool (d15 Nreactant ; eqn [6]) and the product N pool (d15 Nproduct ; eqn [7]) (Figure 1): d15 Nreactant ¼ d15 Ninitial þ eð1 f Þ

½6

d15 Nproduct ¼ d15 Ninitial eðf Þ

½7

½4

d15 Nintegrated ¼ d15 Ninitial þ eff =ð1 f Þglnðf Þ ½5 where f is the fraction of the reactant remaining, d15Ninitial is the d15N of the initial reactant N pool, and e is the kinetic isotope effect of the transformation. These equations are simplified, approximate

The steady-state model and modified forms of it, such as the more spatially complex ‘reaction– diffusion’ model, are used to quantify uptake processes where supply and uptake are simultaneous and relatively time-invariant, such as in the consumption of nitrate by denitrification in the ocean interior or in sediments.


42


Processes Inputs

N2 fixation is the major input of fixed N to the ocean (Figure 2). N2 fixation is carried out by N2 fixers, cyanobacteria and other microorganisms able to convert N2 into biomass N. Subsequent remineralization of this biomass supplies new N to the dissolved fixed N pools in the surface and subsurface ocean. Field collections of Trichodesmium colonies, the best-known genus of open ocean N2 fixer, have yielded a d15N of c. 2% to þ 0.5%. Taking into account the d15N of dissolved N2 (0.6% in the surface mixed layer), this range in d15N is consistent with, but perhaps less variable than, the range in isotope effects estimated from culture studies of marine and terrestrial N2 fixers, B0–4% (Table 1). An average d15N of 1% has been suggested for the fixed N input to the ocean from N2 fixation.

Other inputs of fixed N to the marine environment include terrestrial runoff and atmospheric precipitation, the N isotopic compositions of which are poorly constrained (Figure 2). Dissolved and particulate d15N in pristine river systems ranges mostly from 0% to 5%. However, biological processes along the flow path and in estuaries (in particular, by denitrification; see below) can alter the d15N of the final inputs from terrestrial runoff in complex ways. Anthropogenic inputs often increase the d15N of a system because they encourage denitrification. In atmospheric inputs, a wide range in the d15N of inorganic (c. 16% to 10%) and organic (c. 8% to 1%) N has been observed, with increasing evidence that at least some of this variability can provide insight into sources and processes. In the face of large uncertainties, a preindustrial mean d15N of 4% for terrestrial runoff and 2% for atmospheric precipitation has been suggested by some workers.

Sinking 15N ~ 3.5‰

N2 fixation

≤ 2‰

Atmospheric precipitation 15N ~ −2‰?

Terrestrial runoff 15N ~ 4‰?

Dissolved N2 15N = 0.6‰

~ 5‰

− supply and as si m ila t

NO 3

Thermocline NO3− 15N ~ 3‰

Particulate N 15N ~ 0‰

≤ 3‰?

≤ 3‰? DON 15N ~ 4‰

?

ion

Deep ocean

Recycling

NH4+

Remineralization

Surface ocean

Atmospheric N2 15N = 0‰

Water-column denitrification ~ 25‰

Sedimentary denitrification ~ 0‰

Deep NO3− 15N = 5‰

Figure 2 Processes affecting the distribution of nitrogen isotopes in the sea. The inputs and outputs (solid arrows) control the ocean’s inventory of fixed N, the majority of which is in the form of nitrate (NO3 ). Inputs are marine N2 fixation in the surface ocean, terrestrial runoff, and atmospheric precipitation. Outputs indicated are sedimentary and water column denitrification. As discussed in the text, water-column denitrification in low-oxygen regions of the ocean interior leads to elevated d15N of nitrate (dark area). Remineralization of newly fixed N can explain the low d15N of nitrate in the shallow subsurface, or thermocline, of low-latitude regions (light area). Internal cycling is represented with dashed arrows. Nitrate supplied from the deep ocean and thermocline is assimilated in the surface ocean. PN is recycled in the surface ocean, degraded to ammonium (NH4 þ ) that is subsequently assimilated; the role of DON in this recycling is not yet clear. Sinking and remineralization (via degradation and nitrification, see text) returns N from the particulate pool to nitrate. For simplicity, possible nitrification in the surface ocean is not shown. The isotope fractionation associated with nitrification in the ocean interior is also excluded because this process generally goes to completion (see text). The d15N of particulate suspended and sinking N, DON, and thermocline nitrate is taken from the subtropical North Atlantic. Question marks indicate the greatest uncertainties, due to variation in the available data and/or insufficient data.



Table 1

43

Representative estimates of isotope effects for N cycles processes

Process

Isotope effect (e)

Details

N2 fixation (N2-PN)

1.8–3.0% 0.2% 0.4–0.2% 5–30% 18–28% 25 30% r3% 5–17% 5–6% 5–9% 5% 5% 20% 8–27% 4–27% 6.5–8% 9% 18.5%

Trichodesmium spp. Western Tropical North Pacific Western Tropical Atlantic Pseudomonas stutzeri (marine) Paracoccus denitrificans (terrestrial) Eastern Tropical North Pacific and Arabian Sea Coastal sites, eastern N. Pacific margin, and deep Bering Sea Thalassiosira weissflogii Coastal Antarctic Open Antarctic and Subantarctic Subarctic Pacific Equatorial Pacific Thalassiosira pseudonana Skeletonema costatum Bacterial assemblage Chesapeake Bay Delaware Estuary Scheldt Esturary

14% 19% 35–38% 12–16% Unknown

Nitrosomonas marina (marine) Nitrosomonas C-113a (marine) Nitrosomonas europaea (terrestrial) Chesapeake Bay

Denitrification (NO3 -N2) Water column Sedimentary NO3 assimilation (NO3 -PN)

NH4 þ assimilation (NH4 þ -PN)

Nitrification Ammonia oxidation (NH3 -NO2 )

Nitrite oxidation (NO2 -NO3 )

Outputs

Denitrification, the bacterial reduction of nitrate to N2, is the major mechanism of fixed N loss from the ocean, occurring both in the water column and in sediments when the oxygen concentration is low (o5 mM) (Figure 2). Denitrification strongly discriminates against the heavier isotope, 15N, progressively enriching the remaining nitrate pool in 15N as nitrate consumption proceeds. Culture studies of denitrifying bacteria suggest (with some exceptions) an isotope effect of B20–30%, a range supported by water column estimates (Table 1). The isotopic discrimination during denitrification likely takes place as nitrate is reduced intracellularly to nitrite by the dissimilatory form of the enzyme nitrate reductase, such that unconsumed, 15N-enriched nitrate effluxing from the cell back into ambient waters allows the enzyme-level isotope effect to be expressed at the environmental scale. Where it occurs in low-oxygen regions of the mid-depth ocean, water column denitrification causes a clear elevation in the d15N of nitrate, and it is the reason that global ocean nitrate d15N is higher than that of the N source from N2 fixation, the dominant input. In contrast to water-column denitrification, denitrification in sediments leads to little increase in the d15N of water-column nitrate. The high d15N of nitrate within the pore waters of actively denitrifying sediments demonstrates that isotopic discrimination

occurs at the scale of the organism. However, expression of the organism-scale isotope effect at the scale of sediment/water exchange is minimized by nearly complete consumption of the nitrate at the site of denitrification within sediment pore waters, which prevents 15N-enriched residual nitrate from evading to the overlying water column, yielding an ‘effective’ isotope effect of 3% or less in most sedimentary environments studied so far (Table 1). Another mechanism of fixed N loss that occurs in sediments and the water column is anaerobic ammonium oxidation, or ‘anammox’, in which nitrite (from nitrate reduction or ammonium oxidation) is used to oxidize ammonium to N2 (NO2 þ NH4 þ -N2 þ 2H2 O). This process has unknown effects on isotope distributions in the ocean. The effects of anammox on N isotopes must depend on the organism-scale isotope effects, the sources of nitrite and ammonium substrates for the reaction, and the degree to which these substrates are consumed. For instance, if nitrate reduction by denitrifiers is the source of the nitrite, remineralization processes are the source of the ammonium, and both the nitrite and ammonium are completely consumed in the environment where anammox occurs, then the isotope discrimination would simplify to that of the nitrate reduction by denitrifiers averaged with any isotope discrimination during the remineralization that produces the needed ammonium. It should be noted that many water-column-derived


44


isotope effect estimates for ‘denitrification’ have inherently included the effect of anammox, in that they regress the nitrate d15N increase against the total nitrate deficit relative to phosphate, and ammonium is not observed to accumulate. Internal Cycling

The fluxes associated with internal cycling are neither sources nor sinks of fixed N but affect the distributions of N species and isotopes in the ocean. N assimilation In the surface ocean, phytoplankton assimilate fixed N (nitrate and ammonium, as well as nitrite, urea, and other organic N compounds) (Figure 3). Culture studies indicate that different forms of fixed N are assimilated with distinct isotope effects, although these isotope effects may vary with physiological conditions. For all studied forms, phytoplankton preferentially consume 14N relative to 15 N (Figures 3 and 4). Nitrate is the deep-water source of fixed N for phytoplankton growth, and the degree of its consumption varies across the surface ocean. The isotope effect of nitrate assimilation therefore has a major impact on the isotopic distributions of all N forms in the ocean. Field-based estimates of the isotope effect of nitrate assimilation range from 4% to 10%, with most estimates closer to 5–8% (Table 1).

PN

Culture-based estimates are more variable. Physiological studies suggest that isotopic fractionation associated with nitrate assimilation is imparted by the intracellular assimilatory nitrate reductase enzyme, which has an estimated intrinsic isotope effect of 15– 30%. The enzyme-level isotope effect is expressed by efflux of unconsumed nitrate out of the cell, as also appears to be the case with denitrifiers. The lower range of isotope effect estimates associated with algal nitrate assimilation than with denitrification suggests proportionally less nitrate efflux by algae, perhaps related to the fact that fixed N is often scarce in the surface ocean. Some studies suggest the degree of efflux and isotope effect of nitrate assimilation may vary with growth conditions; one set of studies of the diatom Thalassiosira weissflogii showed a higher isotope effect under light-limited growth than under growth limited by iron or temperature. The isotope effect of nitrate assimilation is an integrative characteristic of the upper ocean biota that can be measured without perturbing the system, and its magnitude is of broad significance in the application of N isotopes to various questions in the modern and past oceans. Thus, it is a priority to develop a predictive understanding of its controls. Other forms of fixed N assimilated by phytoplankton (ammonium, nitrite, and urea) are

Degradation ( ≤ 3‰ ?) DON

Assimilation (nitrate ~ 5‰)

NH4+

(nitrite ~ 1‰)

(ammonium 20‰) (urea ~ 1‰) NO2−

Nitrification (ammonia oxidation ~ 16‰) (nitrite oxidation ~ ?)

NO3−

Figure 3 Schematic diagram of the processes and pools central to the internal cycling of N in the ocean. The isotope effects shown here are based on laboratory studies. Dashed arrows represent assimilation of dissolved species into particulate matter, and solid arrows represent remineralization. Complete consumption of the ammonium pool by assimilation in the surface ocean or by nitrification in the ocean interior causes the relatively high isotope effects associated with these processes to have little effect on N isotope dynamics. However, in regions where ammonium assimilation and nitrification co-occur, their isotope effects will impact the d15N of their respective products, PN and nitrate. In nitrification, ammonia (NH3), rather than the protonated form ammonium (NH4þ ), is oxidized. However, ammonium is the dominant species in seawater, and there is isotope discrimination in the ammonium–ammonia interconversion. Thus, the isotope effects for ‘ammonia oxidation’ given here and elsewhere in the text refer specifically to consumption of ammonium. The processes surrounding DON production and utilization are not well understood from an isotopic perspective but are thought to play an important role in N cycling.



30 NO3−

150

25 Fluor.

20 15

100

10 50

5

15N-PN(‰)

4

5

4

5 0

0 (b)

(c) 6 Fluorescence

PN

15N-PN(‰)

Nitrogen (μM)

(a) 200

45

3

2

3 15N of source

2

1 15N = 4.4−4.5 F

1 0 0

30

40

50 Time (h)

60

70

0 1.0

0.5 F

0

Figure 4 Isotopic fractionation during nitrate (NO3 ) assimilation by a culture of the marine diatom Thalassiosira pseudonana. (a) Time series of NO3 concentration in the culture medium (filled circles), PN (open circles), and fluorescence, a measure of phytoplankton biomass (filled triangles); (b) d15N of PN accumulated over the same time period; (c) d15N of PN accumulated during log phase of growth plotted vs. F, a measure of nitrate utilization. F is [ f/(1 f )] ln f, where f is the fraction of initial NO3 remaining at the time the culture is sampled. e is calculated to be 4.5% from the slope of the regression in (c), according to the Rayleigh integrated product equation (see text, eqn [5]). Dashed horizontal lines in (b) and (c) represent the initial d15N of the source NO3 (3.8%). Reprinted from Waser NAD, Harrison PJ, Nielsen B, Calvert SE, and Turpin DH (1998) Nitrogen isotope fractionation during the uptake and assimilation of nitrate, nitrite, ammonium, and urea by a marine diatom. Limnology and Oceanography 43: 215–224.

produced and nearly completely consumed within the open ocean surface mixed layer (Figure 3). Culture studies suggest an isotope effect for ammonium assimilation of up to B20%, decreasing as ammonium concentration decreases, and minimal isotope effects (o1%) for assimilation of nitrite and urea. In estuaries, where ammonium can accumulate in the shallow subsurface and be entrained into the surface layer, ammonium assimilation causes a clear increase in the d15N of the remaining ammonium pool. Isotope effect estimates based on ammonium concentrations and d15N in these environments range from 6.5% to 18.5% (Table 1).

Remineralization The return of organic N to nitrate occurs in two steps, the degradation of organic N to ammonium and the bacterial oxidation of ammonium to nitrate, or ‘nitrification’ (Figure 3). Nitrification itself occurs in two steps, the oxidation of ammonium to nitrite and the oxidation of nitrite to nitrate, mediated by distinct groups of microorganisms. Isotopic discrimination may occur at all steps involved in remineralization. Field studies generally suggest that both bacteria and zooplankton preferentially degrade low-d15N PN to ammonium, yielding residual organic matter

relatively high in 15N. The wide spectrum of reactions involved in organic N degradation and the heterogeneous nature of organic matter (comprised of compounds with distinct d15N that degrade at various rates) make quantifying the isotope effect associated with degradation difficult. A few laboratory studies have quantified the isotope effects of individual processes such as thermal peptide bond cleavage, bacterial amino acid uptake and transamination, and zooplankton ammonium release. Laboratory studies attempting to mimic degradation as a whole suggest a net isotope effect of r3%. Culture studies indicate a large isotope effect for the conversion of ammonium to nitrite, the first step in nitrification (Table 1). Estimates of the isotope effect for marine ammonia-oxidizing bacteria (B14–19%) are lower than those for terrestrial ammonia-oxidizing bacteria (as high as B38%), possibly due to phylogenetic differences. The isotope effect of nitrification estimated from ammonium concentration and d15N measurements in the Chesapeake Bay is 12–16%, similar to the culture results for marine ammonia-oxidizing bacteria (Table 1). Thus far, no culture-based information is available regarding isotope effects for ammonia-oxidizing crenarchaea or nitrite-oxidizing bacteria.


46


20

Nitrogen Reservoirs Dissolved Nitrogen

15 15NO3− (‰ vs. air)

Nitrate Nitrate accounts for most of the fixed N in the ocean. The d15N of deep ocean nitrate is typically B5%. Regionally, the d15N of nitrate varies between 2% and 20% due to the effects of N2 fixation, nitrate assimilation, and denitrification (Figure 5). Nitrate d15N significantly lower than deep-ocean nitrate has been observed in the upper thermocline of the low-latitude oligotrophic ocean (Figure 6). This 15N depletion is most likely due to the oxidation of newly fixed N, which, as described above, has a d15N of c. 1%. Values higher than 5% result from discrimination associated with nitrate assimilation by phytoplankton at the ocean surface (Figure 7) or denitrification in oxygendeficient zones of the ocean interior (Figure 8). Nitrate assimilation by phytoplankton leads to elevated d15N of nitrate in regions of the ocean where nitrate is incompletely consumed in surface waters, such as the high-latitude, nutrient-rich regions of the Southern Ocean and the subarctic Pacific, and the low-latitude upwelling regions of the California Current and the Equatorial Pacific. In the surface waters of these regions, there is a strong correlation between the degree of nitrate consumption by phytoplankton and the d15N of the nitrate remaining in the water (Figure 7). However, while nitrate assimilation elevates the d15N of nitrate in the surface ocean and causes modest 15N enrichment in some newly formed thermocline waters, it does not appear to affect greatly the d15N of nitrate in the deep ocean. Below 2.5–3.0- km depth in the ocean, nitrate d15N is relatively constant at B5%, despite large inter-basin differences in nitrate concentration. The minimal degree of isotopic variation in the nitrate of the deep ocean is due to the fact that, in most surface waters, the nitrate supply from below is almost completely consumed by phytoplankton, such that the organic N exported from the surface ocean converges on the d15N of the nitrate supply. Because the sinking flux d15N is close to that of the nitrate supplied from the ocean interior, remineralization of the sinking flux in the ocean interior does not alter greatly the d15N of deep nitrate. In this respect, the oceanic cycling of N isotopes differs markedly from that of the carbon isotopes. Because water-column denitrification occurs in the subsurface and because it consumes only a fraction of the nitrate plus nitrite available, its isotope effect is more completely expressed in the d15N of subsurface nitrate. In denitrifying regions of the water column, the d15N of nitrate in the subsurface is commonly elevated to 15% or higher (Figure 8). The

Water-column denitrification ∼ 25‰

10

Nitrate uptake ∼ 5‰

5

Newly fixed N added 15N ~ −1‰

Sedimentary denitrification ∼ 0‰ 0 0

0.5

1

1.5

2

[NO3−] (factor of initial value) Figure 5 The effect of different marine N cycle processes on nitrate d15N and concentration, assuming an initial nitrate d15N of 5%. The trajectories are for reasonable estimates of the isotope effects, and they depend on the initial nitrate d15N as well as the relative amplitude of the changes in nitrate concentration (30% for each process in this figure). A solid arrow denotes a process that adds or removes fixed N from the ocean, while a dashed line denotes a component of the internal cycling of oceanic fixed N. The effects of these two types of processes can be distinguished in many cases by their effect on the concentration ratio of nitrate to phosphate in seawater. The actual impact of the different processes on the N isotopes varies with environment. For instance, if phytoplankton completely consume the available nitrate in a given environment, the isotope effect of nitrate uptake plays no major role in the d15N of the various N pools and fluxes; the effect of nitrate generation by organic matter degradation and nitrification, not shown here, will depend on this dynamic. Similarly, the lack of a large isotope effect for sedimentary denitrification is due to the fact that nitrate consumption by this process can approach completion within sedimentary pore waters.

subsurface d15N maximum occurs in the core of the oxygen minimum and is correlated with the estimated degree of nitrate consumption by water-column denitrification. Denitrification, both in the water column and sediments, exerts a direct control on the d15N of mean deep ocean nitrate. When the ocean N budget is at steady state, the d15N of the fixed N removed (through water-column and sedimentary denitrification) will equal the d15N of the fixed N added (c. 1%, approximating N2 fixation as the sole source) (Figure 9). If denitrification with an isotope effect of 20–30% were occurring homogenously in the ocean water column and responsible for all fixed N loss, the d15N of mean oceanic nitrate would be 19–29% to achieve a d15N of –1% for N loss. That the modern mean oceanic nitrate d15N is B5%, much lower than 19–29%, reflects at least two factors: (1)



(a)

(b)

[N] (μM) 0

5

10

15

20

2

3

15N (‰ vs. air) 4

5

47

6

0 TON 15N

[TON]

Depth (m)

200

400

600

[NO3−]

NO3− 15N

800

1000

Figure 6 Depth profiles of (a) [NO3 ] (open circles) and [TON] (total organic N, or DON plus the small pool of PN) (open squares) and (b) nitrate d15N (filled circles) and TON d15N (filled squares) at the Bermuda Atlantic Time-Series Study site in the oligotrophic Sargasso Sea. Low nitrate d15N in the thermocline has been proposed to reflect N2 fixation at the surface, the N from which sinks and is remineralized at depth. The increase in nitrate d15N above 200 m reflects fractionation associated with nitrate assimilation. [TON] increases slightly into the surface layer while TON d15N decreases, suggesting a possible source of low-15N DON in the surface. The d15N of the TON pool is lower than that of mean ocean nitrate and deep nitrate at Bermuda (B5%) but higher than that of thermocline nitrate at Bermuda (B2.5%). High nitrate concentrations below 250 m prevent accurate assessment of TON d15N using published methods (see text). The d15N of sinking particles collected at 100 m (3.7%) is indicated by the arrow at top of (b); surface suspended PN d15N is c. 0.2% (not shown). Nitrate and TON data are the means of monthly measurements between June 2000 and May 2001. Modified from Knapp AH, Sigman DM, and Lipschultz F (2005) N isotopic composition of dissolved organic nitrogen and nitrate at the Bermuda Atlantic time-series study site. Global Biogeochemical Cycles 19: GB1018 (doi:10.1029/2004GB002320). Sinking and suspended PN d15N data from Altabet MA (1988) Variations in nitrogen isotopic composition between sinking and suspended particles: Implications for nitrogen cycling and particle transformation in the open ocean. Deep Sea Research 35: 535–554.

the importance of sedimentary denitrification, which appears to express a minimal isotope effect; and (2) the localized nature of water-column denitrification. With regard to the second, because denitrification consumes a significant fraction of the ambient nitrate in the ocean’s suboxic zones and elevates its d15N above that of the mean ocean (Figures 5 and 8), the d15N of nitrate being removed by water column denitrification is higher than if the substrate for denitrification had the mean ocean d15N. Much as with sedimentary denitrification, this reduces the expression of the organism-level isotope effect of water column denitrification and thus lowers the mean d15N of nitrate required to achieve an isotope balance between inputs and outputs. With these considerations, one study estimates that water-column denitrification is responsible for 30% of fixed N loss from the modern ocean, with sedimentary denitrification responsible for the remainder. Still, this isotope-based budget for marine fixed N remains uncertain. One limitation of using the N isotopes to investigate N cycling in the ocean is their inability to separate co-occurring processes with competing N isotopic signatures, such as denitrification/N2 fixation and nitrate assimilation/nitrification. Coupled

analysis of N and O isotopes in nitrate promises to disentangle such otherwise overprinting processes. Culture studies have demonstrated that the two most important nitrate-consuming processes, nitrate assimilation and denitrification, fractionate the N and O in nitrate with a ratio close to 1:1 (Figure 10). Deviations in the ratio of d18O and d15N in nitrate from 1:1 can therefore provide information about the nitrate being added by nitrification, such as if it derives from newly fixed N or if there is otherwise undetected nitrate cycling within a water parcel. Nitrite Nitrite is an intermediate in oxidative and reductive processes such as nitrification and denitrification. It also serves as a substrate for anammox and can be assimilated by phytoplankton. Only when these processes become uncoupled, such as at the base of the euphotic zone and in oxygen minimum zones, does nitrite accumulate to significant levels (0.1–10 mM). In some oxygen-deficient regions of the water column in the eastern tropical North Pacific, eastern tropical South Pacific, and the Arabian Sea, nitrite can represent as much as 25% of the pool of nitrogen oxides (NO3 þ NO2 ).


48


[NO3−] (μmol kg−1)

(a) 20

25

30

15N (‰ vs. air)

(b)

35

4

40

4.5

5

5.5

6

6.5

7

7.5

0 500 1000

Depth (m)

1500 2000 2500 3000 3500 4000 (c) 35

14 13

30

+

+

11 10

+ +

+ 15

+

+

9 ++ +

15NO3−

8

15N (‰ vs. air)

[NO3−] (μM)

25 20

12

[NO3−]

7

10

+ +

5

+

Sediment

+

+

15N

6 +

+

+ +

0

5 4

40

42

44

46

48

50

52

54

56

58

60

Latitude along 115° E (S) Figure 7 Nitrate d15N data from the Southern Ocean show the effect of uptake by phytoplankton. A depth profile of nitrate concentration (a) and d15N (b) from the Antarctic region of the east Indian Ocean (53.21 S, 1151 E) shows a decrease in nitrate concentration into the surface layer and an associated increase in nitrate d15N, both resulting from nitrate uptake by phytoplankton. A meridional transect (c) of nitrate concentration (open circles) and d15N (filled circles) in the surface mixed layer of the Southern Ocean along 115 1 E shows that nitrate d15N is spatially correlated with variations in the utilization of nitrate by phytoplankton, as reflected by the equatorward decrease in nitrate concentration. Nitrate d15N is lowest where the nitrate concentration is highest (and nitrate utilization is lowest), in Antarctic waters, south of B511 S in this region. The meridional gradient in nitrate d15N is recorded in bulk sediment d15N along 115 1 E (c, crosses) through the sinking of PN out of the surface ocean. Across the transect, sediment d15N is B5% higher than expectations for and measurements of sinking N d15N, probably mainly due to isotopic alteration of this N at the seafloor. Nevertheless, the link between nitrate consumption and sediment d15N provides a possible avenue for paleoceanographic reconstruction of nitrate utilization by phytoplankton. The left axis in (c) is scaled to 35 mM, the approximate nitrate concentration of Upper Circumpolar Deep Water that upwells in the Antarctic (see (a)). (a, b) Reprinted from Sigman DM, Altabet MA, Michener RH, McCorkle DC, Fry B, and Holmes RM (1997) Natural abundance-level measurement of the nitrogen isotopic composition of oceanic nitrate: An adaptation of the ammonia diffusion method. Marine Chemistry 57: 227–242. (c) Modified from Sigman DM, Altabet MA, McCorkle DC, François R, and Fischer G (2000) The d15N of nitrate in the Southern Ocean: Nitrate consumption in surface waters. Global Biogeochemical Cycles 13: 1149–1166, with the sediment data taken from Altabet MA and François R (1994) Sedimentary nitrogen isotopic ratio records surface ocean nitrate utilization. Global Biogeochemical Cycles 8: 103–116.

The d15N of nitrite is expected to reflect the balance of isotopic fractionation during nitrite production and consumption processes. In the eastern tropical North Pacific, the d15N of nitrite is very low

( 18% to 7%). This d15N is B30% lower than the nitrate in the same environment and is lower than expected from denitrification alone, give the known isotope effects for nitrate and nitrite reduction. The



15N of NO3−

(a)

6.0

0

9.0

Nitrate (μM)

(b) 12.0 15.0

500 Depth (m)

Depth (m)

0 10 20 30 40

0

500

1000

1500

1000

1500

2000

2000

0.2

49

0.4

0.6 15N

0

0.8

10 Nitrate deficit (μM)

of N2

Figure 8 (a) The d15N of NO3 (filled circles) and N2 (open circles) in water column profiles through an intense denitrification zone in the eastern tropical North Pacific (221 N, 1071 W). The shaded interval indicates the depth range with dissolved O2 concentration o10 mM where denitrification is encouraged (b, concentration shown in filled squares) and leads to a characteristic NO3 deficit (b, open squares) relative to phosphate. Measurements indicate enrichment of NO3 in 15N and concurrent depletion of N2 in 15N, arising from isotope discrimination during denitrification, with the conversion of NO3 to N2. e for denitrification in this environment was estimated to be B25%. Modified from Brandes JA, Devol AH, Yoshinari T, Jayakumar DA, and Naqvi SWA (1998) Isotopic composition of nitrate in the central Arabian Sea and eastern tropical North Pacific: A tracer for mixing and nitrogen cycles. Limnology and Oceanography 43: 1680–1689.

20 15 15N (‰ vs. air) 10 5

15N ~ 5‰

Sedimentary 30

0 −5

Denitrification

−10 −15

Water column

−20 Figure 9 Simplified global ocean N isotope budget. The y-axis indicates the d15N of a given flux or pool. The d15N of N from oceanic N2 fixation, the dominant N input to the ocean, is c. 1% (‘N2 fixation’ on the left). At steady state, the total denitrification loss (‘denitrification’ on the right) must have the same d15N as the input. The d15N of mean ocean nitrate is B5%. Water column denitrification removes nitrate with a low d15N (‘water column’ at lower right), while sedimentary denitrification removes nitrate with a d15N similar to that of mean ocean nitrate (‘sedimentary’ at upper right). The need for the flux-weighted d15N of the denitrification loss to be c. –1% leads to estimates of partitioning between water-column and sedimentary denitrification in which sedimentary denitrification is found to drive the greater part of the total N loss.

meaning of this low d15N is still not well understood but likely points to other processes acting on the NO2 pool. In these regions, nitrite can represent a significant reservoir of N that is depleted in 15N.

Δ18O of NO3− (‰ vs. initial)

N2 fixation

Mean ocean NO3−

18:15

= 1 0.9, 1.1

20

10

T. weissflogii T. pseudonana T. oceanica E. huxleyi

0 0

20 10 Δ15N of NO3− (‰ vs. initial)

30

Figure 10 The d18O vs. d15N in nitrate as it is progressively assimilated by four eukaryotic species of marine phytoplankton. Both d15N and d18O in nitrate increase as nitrate is consumed, and they do so with an O:N ratio for isotopic discrimination (18e:15e) of B1. Dashed lines show slopes of 1.1 and 0.9 for comparison. Modified from Granger J, Sigman DM, Needoba JA, and Harrison PJ (2004) Coupled nitrogen and oxygen isotope fractionation of nitrate during assimilation by cultures of marine phytoplankton. Limnology and Oceanography 49: 1763–1773.


50


Until recently, the two species nitrate and nitrite have typically been combined in isotopic analysis, such that the presence of low d15N nitrite may have masked some of the 15N enrichment of nitrate in the oxygen-deficient zone. Ammonium The d15N of ammonium reflects the production of ammonium by the degradation of organic N and its consumption by nitrification, ammonium assimilation, and perhaps anammox (Figure 3). Analytical constraints have limited isotopic studies of ammonium to environments with ammonium concentrations greater than 1 mM, excluding studies in the open ocean. In estuarine systems, where ammonium can be abundant, its d15N is often high (commonly higher than þ 10%, with one observation of þ 70%) and it increases as the ammonium concentration decreases along transects from riverine to marine waters, due to discrimination associated with ammonium consumption by nitrification and/or ammonium assimilation. In the open ocean interior, below the depth of algal assimilation, essentially all ammonium generated from particles is oxidized to nitrite and then nitrate before it can be transported into or out of a given region. Thus, nitrification should be of limited importance for the isotope dynamics of both particulate and dissolved N once the former has sunk out of the upper ocean. In the open ocean surface mixed layer, it is generally assumed that ammonium generated by remineralization is quickly and entirely assimilated by plankton, in which case the isotope effect associated with its consumption would not play an important role in N isotope dynamics of the open ocean. However, in at least some regions of the upper ocean, ammonium oxidation and assimilation are likely to co-occur. If the isotope effect of ammonium oxidation is greater than that of ammonium assimilation, low-d15N N will preferentially be routed to the nitrate pool by oxidation and high-d15N N will be routed back to the PN pool by assimilation. If the isotope effect of oxidation is less than that of assimilation, the opposite will occur. Thus, with better constaints on the isotope effects of ammoniumconsuming processes, the isotopes of upper ocean N pools promise to provide an integrative constraint on the relative rate of nitrification in the upper ocean, especially when paired with the O isotopes of nitrate (see above). Dissolved organic nitrogen DON concentrations are significant in the open ocean, typically Z4 mM in surface waters, decreasing to B2 mM in deep water. Fluxes associated with the DON pool are among

the least constrained terms in the modern marine N budget and may be important. Studies to date of bulk DON have been in the subtropical ocean, where DON is by far the dominant N pool in the surface ocean. In the surface mixed layer at the Bermuda Atlantic Time-series site in the Sargasso Sea, the concentration and d15N of TON are B4 mM and B4% (TON being total organic N, or DON plus the small pool of PN) (Figure 6). This d15N is similar to or slightly higher than the shallow subsurface nitrate that is entrained into the euphotic zone during wintertime vertical mixing (Figure 6(b)). Minimal gradients in the concentration and d15N of DON in this region of the upper ocean hinder reconstruction of fluxes of DON or the d15N of those fluxes. There is a weak increase in the concentration of TON into the surface layer and an accompanying decrease in its d15N (Figures 6(a) and 6(b)). Thus, there may be an input of low-d15N N into the surface DON pool, which is remineralized at depth, but this requires further validation. Progress on DON d15N dynamics would be aided by a method to remove nitrate from samples without compromising the DON pool, which would make subsurface waters and high-nitrate surface waters more accessible to study. Ongoing work on separable fractions of the DON pool (e.g., the highmolecular-weight fraction and its components) is also promising. Dissolved gases Dissolved N2 in equilibrium with atmospheric N2 at the surface has a d15N of 0.6%. The isotopic composition of dissolved N2 does not vary greatly in ocean profiles, except in zones of denitrification. Production of low-d15N N2 in denitrification zones results in measured N2 d15N as low as 0.2% (Figure 8(a)). Since N2 is the main product of denitrification, its d15N provides a test of the nitrate-based estimates of the isotope effect for this process. Dissolved N2O is produced by nitrification and is both produced and consumed by denitrification. The marine flux of N2O is perhaps one-third of the global flux of this greenhouse gas to the atmosphere; therefore, an understanding of the mechanisms of N2O production and their regulation in the ocean is an important goal. Culture studies indicate that bacterial production of N2O by nitrification and denitrification produces gas depleted in 15N and 18O relative to the source material. Consumption of N2O by denitrification leaves the residual gas enriched in 15 N and 18O, with d15N of N2O as high as 40% measured in the Arabian Sea. In oxygenated waters of the open ocean, nitrification likely dominates N2O production and its isotopic profile. A depth profile in



the subtropical North Pacific shows three main features (Figure 11): (1) isotopic equilibrium with atmospheric N2O at the surface; (2) a subsurface d15N minimum attributed to nitrification; and (3) a broad d15N maximum in deeper waters probably due to N2O consumption, perhaps in the denitrifying waters of the eastern Pacific margin. In and near denitrification zones, a strong maximum in the d15N of N2O is observed, presumably due to isotope fractionation associated with N2O consumption (via reduction to N2).

to surface waters. The isotopic effect of N recycling originates from heterotrophic processes. Zooplankton appear to release ammonium which has a lower d15N than their food source, making their tissues and solid wastes B3% higher in d15N than their food source. The low-d15N ammonium is consumed by phytoplankton and thus retained in the surface ocean N pool, while the 15N-enriched PN is preferentially exported as sinking particles, leading to a lower d15N of surface PN in regions where recycled N is an important component of the gross N supply to phytoplankton. Low d15N observed in suspended PN from the Antarctic and other highlatitude regions, beyond what is expected from isotope discrimination during nitrate assimilation, is unlikely to be due to N2 fixation and thus likely reflects N recycling. In the low-latitude, lownutrient ocean surface, such as the Sargasso Sea and western tropical Pacific, the relative importance of N2 fixation and N recycling in producing lowd15N surface particles is uncertain. Because of its implications for the rates of N2 fixation and N recycling, this question deserves further study. The d15N of suspended particles in the subsurface is typically B6% higher than suspended particles in the surface ocean and B3% higher than the sinking flux (Figure 12). The d15N of deep

Particulate Nitrogen

Suspended particles A typical profile of suspended particles has its lowest d15N in the surface layer, increasing below the euphotic zone (Figure 12). The d15N of suspended particles reflects in part the d15N of nitrate supplied to the surface ocean and, in nutrient-rich surface waters, isotope discrimination associated with its incomplete assimilation. However, the low d15N in the surface layer is typically lower than what would be expected solely from nitrate assimilation. This low d15N has two competing explanations: N2 fixation and N recycling. As described earlier, N2 fixation is expected to add fixed N with a d15N of c. 1% (a)

51

(b)

(c)

0 200 400 600

Depth (m)

800 1000

2000

HOT-65 HOT-76 HOT-79 HOT-82

3000 4000 5000 10

20

30

40

[N2O] (nM)

50

6

7

8

9

10

15N-N2O (‰ vs. air-N2)

18

22

26

30

34

18O-N2O (‰ vs. air-O2)

Figure 11 Depth profiles of (a) N2O concentration and (b) d15N and (c) d18O of N2O at station ALOHA in the subtropical North Pacific (221 450 N, 1581 W) during four separate cruises. The solid line in (a) indicates theoretical saturation with atmospheric N2O at in situ temperatures and salinities. The minima in d15N and d18O around 200 m are thought to be due to significant in situ production of N2O from nitrification. The broad isotopic maxima at depth are likely due to N2O consumption, perhaps in the denitrifying waters along the eastern Pacific margin. The filled squares at the top of (b) and (c) represent measurements of d15N and d18O of atmospheric N2O during the Hawaii Ocean Time-series 76 cruise, and arrows indicate the range of historical measurements as of the late 1980s. Reprinted from Dore JE, Popp BN, Karl DM, and Sansone FJ (1998) A large source of atmospheric nitrous oxide from subtropical North Pacific surface waters. Nature 396: 63–66.


52


15N (‰) −1

0

1

2

3

4

5

6

7

8

9

0 500

Depth (m)

1000 1500 2000 2500 3000 3500 Sediment trap Suspended particles Figure 12 Nitrogen isotopic values of suspended particulate matter and sinking particles (as collected by sediment traps) in the subtropical North Atlantic Ocean near BATS (311 500 N, 641 100 W; see Figure 6). The profiles of suspended PN show the representative depth gradient in d15N, with lower d15N in the surface ocean than at depth. The d15N of the sinking flux shows a decrease with depth, which has also been observed in other regions. Reprinted from Altabet MA, Deuser WG, Honjo S, and Stienen C (1991) Seasonal and depth-related changes in the source of sinking particles in the North Atlantic. Nature 354: 136–139.

particles is consistent with the inference that deep particles are the breakdown products of material exported from the surface, and that bacteria preferentially remineralize low-d15N PN. Isotopic analysis of zooplankton and organisms at higher trophic levels can provide insights into the marine N cycle. The ‘trophic effect’, an observed B3% increase in d15N per trophic level that presumably results from isotopic discrimination during metabolism of N-bearing organic matter, is used widely in food-web studies. N isotopic analysis of specific amino acids within organisms and organic matter promises new insights, as some amino acids increase in d15N with trophic level while others preserve the d15N of the food source. Sinking Particulate Nitrogen and sedimentary Nitrogen Because vertical sinking is an important mode of N export from the surface ocean, the d15N of the sinking flux is one of the most valuable N isotopic constraints on modern ocean processes. Combined with other isotopic data, sinking flux d15N data can provide information on the routes and mechanisms of nitrate supply and can be used to constrain other sources of N to the surface. The sinking flux also transfers the isotopic signal from

the surface ocean to the seafloor, providing the link through which the sediment column records the history of surface ocean processes (Figure 7(c)). Sinking particles collected in depth arrays of sediment traps often show a modest decrease in d15N with depth (Figure 12). This trend runs contrary to our expectations for the isotopic change of particulate matter as it degrades, and it currently lacks a compelling explanation. There is generally a good correlation between the d15N of surface sediments and sinking particulate d15N from the overlying water column. In regions of the ocean where a relatively large fraction of the organic rain is preserved in the sediment column, as occurs along continental margins, this correlation is excellent. In open ocean sediments where only a very small fraction of N is preserved, spatial patterns in the d15N of sediment core tops mirror those in the water column above (Figure 7(c)), but a significant 15 N enrichment (of B2–5%) is observed in the sediment N relative to sinking particles. Upon burial, reactions in the shallow sediment column known collectively as ‘diagenesis’ can cause a clear increase in the d15N of PN as it is incorporated into the sediment mixed layer. While some studies have found that sedimentary diagenesis has not greatly affected the paleoceanographic information provided by specific sedimentary records, it cannot be assumed that changes in the ‘diagenetic offset’ are unimportant. To address concerns regarding alteration of both sinking and sedimentary bulk d15N, studies are increasingly focusing on isolating specific N components, the d15N of which is insensitive to diagenesis, such as N bound within the mineral matrix of microfossils, or that does not change in d15N as it is degraded, such as chlorophyll degradation products. Nitrogen isotopes in the sedimentary record The isotopes of sedimentary N are used to investigate past changes in the marine N budget and the internal cycling of N within the ocean. The processes and parameters reflected by the d15N of sedimentary N include (1) mean ocean nitrate d15N, (2) regional subsurface nitrate 15N depletion or enrichment relative to the global ocean owing to N2 fixation or denitrification, (3) regional isotope dynamics associated with partial nitrate consumption in surface waters, and (4) possible direct contribution of newly fixed N to sinking PN. Paleoceanographers have focused on sediment d15N records underlying three environments where a single process or parameter is thought to dominate changes in sinking d15N. In oligotrophic regions, sediment d15N is assumed to reflect the d15N of mean ocean nitrate and therefore the global ocean



(a)

15N (‰ vs. air)

2

3

4

5

6

15N (‰ vs. air)

(b) 7

5

6

7

8

9

15N (‰ vs. air)

(c)

10 11 1

53

2

3

4 5

6

7 8

0 Stage 1 20

Age (kyr)

40

Stage 2

Stage 3

60 Stage 4 80

100

Stage 5

120

140

Stage 6

Figure 13 Sedimentary d15N records spanning the past 140 ky, which encompass recent ice ages (shaded, marine oxygen isotope stages 2, 4, and 6, with 2 and 6 being the most extreme) and interglacials (stages 1, 3, and 5, with 1 and 5 being the most extreme). (a) Sediment record underlying the oligotrophic South China Sea (8 130.4 N, 112 119.9 E), where sedimentary d15N is expected to track, with some offset, the d15N of nitrate in the western Pacific thermocline. The western Pacific thermocline nitrate, in turn, is assumed to have maintained a constant isotopic relationship with deep ocean nitrate. The small magnitude of variation in d15N (o1.5%) and lack of correlation with glacial/interglacial transitions suggest that mean ocean nitrate d15N remained unchanged through shifts in Earth’s climate. (b) Sedimentary d15N record underlying the eastern tropical North Pacific (22 123.3 N, 107 104.5 W), a major region of denitrification. Interglacials are characterized by high d15N (8–9%), with d15N 2–3% lower during glacials. Low d15N, along with coincident evidence for decreased productivity and higher oxygen content in the mid-depth water-column, indicates decreased water column denitrification during glacial periods. (c) Sedimentary d15N record from the high-nitrate Antarctic Zone of the Southern Ocean (54 155 S, 73 150 E) shows higher d15N during the period spanning glacial stages 2–4, suggesting greater algal utilization of nitrate in the surface ocean. Coupled with evidence of lower glacial productivity, the glacial 15N enrichment suggests reduced nutrient supply from below. All data shown are of bulk sediment. The sedimentary d15N record shown in (b) is from a region of high organic matter preservation in the sediments, where bulk sedimentary d15N correlates well with sinking d15N. The records in (a) and (c) are from regions where a diagenetically driven difference is observed between sinking and sedimentary N, which introduces uncertainties in interpretation. (a) Core 17961 from Kienast M (2000) Unchanged nitrogen isotopic composition of organic matter in the South China Sea during the last climatic cycle: Global implications. Paleoceanography 15: 244–253. (b) Core NH8P from Ganeshram RS, Pederson TF, Calvert SE, and Murray JW (1995) Evidence from nitrogen isotopes for large changes in glacial–interglacial oceanic nutrient inventories. Nature 376: 755–758. (c) Core MD84-552 from François R, Altabet MA, Yu E-F, et al. (1997) Contribution of Southern Ocean surface-water stratification to low atmospheric CO2 concentrations during the last glacial period. Nature 389: 929–935.

balance of inputs and outputs of fixed N (Figure 13(a)). In denitrifying regions, sediment d15N has been taken to largely reflect changes in regional 15N enrichment due to water column denitrification (Figure 13(b)). In high-nutrient regions, sediment d15N primarily records the degree of nitrate consumption by algal assimilation, providing insight into changes in balance between gross nitrate supply to surface waters and export of organic N from the surface (Figure 13(c)). However, it must be kept in mind that multiple processes may affect the d15N of sediments in any given region. For example, even in low-nutrient regions far from denitrification zones, the sediment d15N record may record changes not only in mean ocean nitrate d15N

but also in the d15N of nitrate imported from subpolar regions and in the isotopic imprint of N2 fixation.

Concluding Remarks The study of the N isotopes in the ocean is young relative to that of the other light isotopes (e.g., carbon, oxygen, and sulfur), with much of the work to date developing the methods needed to measure different forms of oceanic N and establishing the isotope systematics of N cycle processes that are necessary to interpret observed patterns. Over the previous decades, the N isotopes have had perhaps their greatest impact on foodweb studies and in paleoceanographic


54


work. In the case of the latter, this reflects the ability of the N isotopes to provide basic constraints on environmental conditions when there are few other indicators available. Recent and ongoing method development is greatly improving our ability to measure diverse N pools in the ocean. This is yielding a new generation of N isotope studies that promise to provide geochemical estimates for the rates and distributions of N fluxes in the modern ocean, complementing instantaneous ‘bottle’ measurements of these fluxes as well as other geochemical approaches. Fundamental aspects of the oceanic N cycle are still poorly understood, and the N isotopes provide an important tool for their study.

See also Nitrogen Cycle. Redfield Ratio. Sedimentary Record, Reconstruction of Productivity from the.

Further Reading Altabet MA (2005) Isotopic tracers of the marine nitrogen cycle: Present and past. In: Hutzinger O (ed.) The Handbook of Environmental Chemistry, Vol. 2: Marine Organic Matter: Chemical and Biological Markers, pp. 251--294. Berlin: Springer. Altabet MA and Francois R (1994) The use of nitrogen isotopic ratio for reconstruction of past changes in surface ocean nutrient utilization. In: Zahn R, Kaminski M, Labeyrie L, and Pederson TF (eds.) Carbon Cycling in the Glacial Ocean: Constraints on the Ocean’s Role in Global Change, pp. 281--306. Berlin: Springer. Brandes JA and Devol AH (2002) A global marine fixed nitrogen isotopic budget: Implications for Holoene nitrogen cycling. Global Biogeochemical Cycles 16: 1120 (doi:10.1029/2001GB001856). Casciotti KL and McIlvin MR (2007) Isotopic analyses of nitrate and nitrite from references mixtures and application to Eastern Tropical North Pacific waters. Marine Chemistry 107: 184--201. Chang CCY, Silva SR, Kendall C, Michalski G, Casciotti KL, and Wankel S (2004) Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio Measurement. In: deGroot PA (ed.) Handbook of Stable Isotope Analytical Techniques, vol. 1, pp. 305--354. Amsterdam: Elsevier.

Fogel ML and Cifuentes LA (1993) Isotope fractionation during primary production. In: Engel MH and Macko SA (eds.) Organic Geochemistry, pp. 73--98. New York: Plenum. Fry B (2006) Stable Isotope Ecology. New York: Springer. Galbraith ED, Sigman DM, Robinson RS, and Pedersen TF (in press) Nitrogen in past marine environments. In: Bronk DA, Mulholland MR, and Capone DG (eds.) Nitrogen in the Marine Environment. Goericke R, Montoya JP, and Fry B (1994) Physiology of isotope fractionation in algae and cyanobacteria, In: Lajtha K and Michener R (eds.) Stable Isotopes in Ecology and Environmental Science, 1st edn., pp. 187--221. Oxford: Blackwell Scientific Publications. Macko SA, Engel MH, and Parker PL (1993) Early diagenesis of organic matter in sediments: Assessment of mechanisms and preservation by the use of isotopic molecular approaches. In: Engel MH and Macko SA (eds.) Organic Geochemistry, pp. 211--224. New York: Plenum. Michener R and Lajtha K (2007) Stable Isotopes in Ecology and Environmental Science, 2nd edn., New York: Wiley-Blackwell. Montoya JP (1994) Nitrogen isotope fractionation in the modern ocean: Implications for the sedimentary record. In: Zahn R, Kaminski M, Labeyrie L, and Pederson TF (eds.) Carbon Cycling in the Glacial Ocean: Constraints on the Ocean’s Role in Global Change, pp. 259--279. Berlin: Springer. Needoba JA, Sigman DM, and Harrison PJ (2004) The mechanism of isotope fractionation during algal nitrate assimilation as illuminated by the 15N/14N of intracellular nitrate. Journal of Phycology 40: 517--522. Owens NJP (1987) Natural variations in 15N in the marine environment. Advances in Marine Biology 24: 390--451. Peterson BJ and Fry B (1987) Stable isotopes in ecosystem studies. Annual Review of Ecology and Systematics 18: 293--320. Sigman DM, Granger J, DiFiore PJ, Lehmann MF, Ho R, Cane G, and van Geen A (2005) Coupled nitrogen and oxygen isotope measurements of nitrate along the eastern North Pacific margin. Global Biogeochemical Cycles 19: GB4022 (doi:10.1029/2005GB002458). Wada E (1980) Nitrogen isotope fractionation and its significance in biogeochemical processes occurring in marine environments. In: Goldberg ED, Horibe Y, and Saruhashi K (eds.) Isotope Marine Chemistry, pp. 375--398. Tokyo: Uchida Rokakudo.


NOBLE GASES AND THE CRYOSPHERE M. Hood, Intergovernmental Oceanographic Commission, Paris, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1894–1898, & 2001, Elsevier Ltd.

Introduction Ice formation and melting strongly influence a wide range of water properties and processes, such as dissolved gas concentrations, exchange of gases between the atmosphere and the ocean, and dense water formation. As water freezes, salt and gases dissolved in the water are expelled from the growing ice lattice and become concentrated in the residual water. As a result of the increased salt content, this residual water becomes more dense than underlying waters and sinks to a level of neutral buoyancy, carrying with it the dissolved gas load. Dense water formation is one of the primary mechanisms by which atmospheric and surface water properties are transported into the interior and deep ocean, and observation of the effects of this process can answer fundamental questions about ocean circulation and the ocean–atmosphere cycling of biogeochemically important gases such as oxygen and carbon dioxide. Because it is not possible to determine exactly when and where dense water formation will occur, it is not an easy process to observe directly, and thus information about the rates of dense water formation and circulation is obtained largely through the observation of tracers. However, when dense water formation is triggered by ice formation, interaction of surface water properties with the ice and the lack of full equilibration between the atmosphere and the water beneath the growing ice can significantly modify the concentrations of the tracers in ways that are not yet fully understood. Consequently, the information provided by tracers in these ice formation areas is often ambiguous. A suite of three noble gases, helium, neon, and argon, have the potential to be excellent tracers in the marine cryosphere, providing new information about the interactions of dissolved gases and ice, the cycling of gases between the atmosphere and ocean, and mixing and circulation pathways in high latitude regions of the world’s oceans and marginal seas. The physical chemistry properties of these three gases span a wide range of values, and these differences cause them to respond to varying degrees to physical

processes such as ice formation and melting or the transfer of gas between the water and air. By observing the changes of the three tracers as they respond to these processes, it is possible to quantify the effect the process has on the gases as a function of the physical chemistry of the gases. Subsequently, this ‘template’ of behavior can be used to determine the physical response of any gas to the process, using known information about the physical chemistry of the gas. Although this tracer technique is still being developed, results from laboratory experiments and field programs have demonstrated the exciting potential of the nobel gases to provide unique, quantitative information on a range of processes that it is not possible to obtain using conventional tracers.

Noble Gases in the Marine Environment The noble gases are naturally occurring gases found in the atmosphere. Table 1 shows the abundance of the noble gases in the atmosphere as a percentage of the total air composition, and the concentrations of the gases in surface sea water when in equilibrium with the atmosphere. Other sources of these gases in sea water include the radioactive decay of uranium and thorium to helium-4 (4He), and the radioactive decay of potassium (40K) to argon (40Ar). For most areas of the surface ocean, these radiogenic sources of the noble gases are negligible, and thus the only significant source for these gases is the atmosphere. The noble gases are biogeochemically inert and are not altered through chemical or biological reactions, making them considerably easier to trace and quantify as they move through a system than other gases whose concentrations are modified through reactions. The behavior of the noble gases is largely determined by the size of the molecule of each gas and the natural affinity of each gas to reside in a Table 1

Noble gases in the atmosphere and sea water

Gas

Abundance in the atmosphere (%)

Concentration in seawater (cm3 g 1)

Helium Neon Argon

0.0005 0.002 0.9

3.75 108 1.53 107 2.49 104


55

56

NOBLE GASES AND THE CRYOSPHERE

gaseous or liquid state. The main physical chemistry parameters of interest are the solubility of the gas in liquid, the temperature dependence of this solubility, and the molecular diffusivity of the gas. The suite of noble gases have a broad range of these properties, and the behavior of the noble gases determined by these properties, can serve as a model for the behavior of most other gases. One unique characteristic of the noble gases that makes them ideally suited as tracers of the interactions between gases and ice is that helium and neon are soluble in ice as well as in liquids. It has been recognized since the mid-1960s that helium and neon, and possibly hydrogen, should be soluble in ice because of the small size of the molecules, whereas gases having larger atomic radii are unable to reside in the ice lattice. These findings, however, were based on theoretical treatises and carefully controlled laboratory studies in idealized conditions. It was not until the mid-1980s that this process was shown to occur on observable scales in nature, when anomalies in the concentrations of helium and neon were observed in the Arctic. The solubility of gases in ice can be described by the same principles governing solubility of gases in liquids. Solubility of gases in liquids or ice occurs to establish equilibrium, where the affinities of the gas to reside in the gaseous, liquid, and solid state are balanced. The solubility process can be described by two principle mechanisms: 1. creation of a cavity in the solvent large enough to accommodate a solute molecule; 2. introduction of the solute molecule into the liquid or solid surface through the cavity. In applying this approach to the solubility of gases in ice, it follows that if the atomic radius of the solute gas molecule is smaller than the cavities naturally present in the lattice structure of ice, then the energy required to make a cavity in the solvent is zero, and the energy required for the solubility process is then only a function of the energy required to introduce the solute molecule into the cavity. For this reason, the solubility of a gas molecule capable of fitting in the ice lattice is greater than its solubility in a liquid. The solubilities of helium and neon in ice have been determined in two separate laboratory studies, and although the values agree for the solubility of helium in ice, the values for neon disagree. The size of neon is very similar to the size of a cavity in the ice lattice, and the discrepancies between the two reported values for the solubility of neon in ice may result from small differences in the experimental procedure. During ice formation, most gases partition between the water and air phases to try to establish

Table 2

Noble gas partitioning in three phases

Partition phases

Helium

Neon

Argon

Bubble to water Bubble to ice Ice to water

106.8 56.9 1.9

81.0 90.0, 56.3 0.9, 1.4

18.7 N 0

equilibrium under the changing conditions, whereas helium and neon additionally partition into the ice phase. As water freezes, salt and gases are rejected from the growing ice lattice, increasing the concentrations of salt and gas in the residual water. Helium and neon partition between the water and ice reservoirs according to their solubility in water and ice. The concentrations of the gases in the residual water that have been expelled from the ice lattice, predominantly oxygen and nitrogen, can become so elevated through this process that the pressure of the dissolved gases in the water exceeds the in situ hydrostatic pressure and gas bubbles form. The gases then partition between the water, the gas bubble, and the ice according to the solubilities of the gases in each phase. This three-phase partitioning process can occur either at the edge of the growing ice sheet at the ice–water interface, or in small liquid water pockets, called ‘brine pockets’ in salt water systems, entrained in the ice during rapid ice formation. Table 2 quantitatively describes how the noble gases partition between the three phases when a system containing these three phases is in equilibrium in fresh water at 01C. The numbers represent the amount of the gas found in one phase relative to the other. For example, the first row describes the amount of each gas that would reside in the gaseous bubble phase relative to the liquid phase; thus for helium, there would be 106.8 times more helium present in the bubble than in the water. This illustrates the small solubility of helium in water and its strong affinity for the gas phase. Because helium is 1.9 times more soluble in ice than in water, helium partitions less strongly between the bubble and ice phases compared to the partition between the bubble and water phases. The two numbers shown for neon represent the two different estimates for the solubility of neon in the ice phase. One estimate suggests that neon is less soluble in ice than in water, whereas the other suggests that it is more soluble in ice.

Application of the Noble Gases as Tracers The noble gases have been used as tracers of air–sea gas exchange processes for more than 20 years.


NOBLE GASES AND THE CRYOSPHERE

Typically, the noble gases are observed over time at a single location in the ocean along with other meteorological and hydrodynamic parameters to characterize and quantify the behavior of each of the gases in response to the driving forces of gas exchange such as water temperature, wind speed, wave characteristics, and bubbles injected from breaking waves. Because both the amount and rate of a gas transferred between the atmosphere and ocean depend on the solubility and diffusivities of the gas, the noble gases have long been recognized as ideal tracers for these processes. In addition, argon and oxygen have very similar molecular diffusivities and solubilities, making argon an excellent tracer of the physical behavior of oxygen. By comparing the relative concentration changes of argon and oxygen over time, it is possible to account for the relative contributions of physical and biological processes (such as photosynthesis by phytoplankton in the surface ocean) to the overall concentrations, thus constraining the biological signal and allowing for estimates of the biological productivity of the surface ocean. The observations of anomalous helium and neon concentrations in ice formation areas and the suggestion that these anomalies could be the result of solubility of these gases in the ice were made in 1983, and since that time, a number of laboratory and field studies have been conducted to characterize and quantify these interactions. The partitioning of the noble gases among the three phases of gas, water, and ice creates a very distinctive ‘signature’ of the noble gas concentrations left behind in the residual water. Noble gas concentrations are typically expressed in terms of ‘saturation’, which is the concentration of a gas dissolved in the water relative to its equilibrium with the atmosphere at a given temperature. For example, a parcel of water at standard temperature and pressure containing the concentrations of noble gases shown in column 2 of Table 1 would be said to have a saturation of 100%. Saturations that deviate from this 100% can arise when equilibration with the atmosphere is incomplete, either because the equilibration process is slow relative to some other dynamic process acting on the system (for example, rapid heating or cooling, or injection of bubbles from breaking waves), or because full equilibration between the water and atmosphere is prevented, as in the case of ice formation. Typical saturations for the noble gases in the surface ocean range from 100 to 110% of atmospheric equilibrium, due mostly to the influx of gas from bubbles. Ice formation, however, can lead to quite striking saturations of 70 to 60% for helium and neon and þ 230% for argon in the relatively

57

undiluted residual water. Ice melting can also lead to large anomalous saturations of the noble gases, showing the reverse of the freezing pattern for the gas saturations, where helium and neon are supersaturated while argon is undersaturated with respect to the atmosphere. The interactions of noble gases and ice have been well-documented and quantified in relatively simple freshwater systems. Observations of large noble gas anomalies in a permanently ice-covered antarctic lake were quantitatively explained using the current understanding of the solubility of helium and neon in ice and the partitioning of the gases in a three-phase system. Characteristics of ice formed from salt water are more complex than ice formed from fresh water, and the modeling of the system more complex. Using a set of equations developed in 1983 and measurements of the ice temperature, salinity, and density, it is possible to calculate the volume of the brine pockets in the ice and the volume of bubbles in the ice. With this type of information, a model of the ice and the dissolved gas balance in the various phases in the ice and residual water can be constructed. Such an ice model was developed during a field study of gas–ice interactions in a seasonally ice-covered lagoon, and the model predicted the amount of argon, nitrogen, and oxygen measured in bubbles in similar types of sea ice. No measurements are available for the amount of helium and neon in the bubbles of sea ice to verify the results for these gases. It is also possible to predict the relative saturations of the noble gases in the undiluted residual water at the ice– water interface, and this unique fingerprint of the noble gases can then serve as a tracer of the mixing and circulation of this water parcel as it leaves the surface and enters the interior and deep ocean. In this manner, the supersaturations of helium from meltwater have been successfully used as a tracer of water mass mixing and circulation in the Antarctic, and the estimated sensitivity of helium as a tracer for these processes is similar to the use of the conventional tracer, salinity, for these processes. As an illustration of the ways in which the noble gases can be used to distinguish between the effects of ice formation, melting, injection of air bubbles from breaking waves, or temperature changes on dissolved gases, Figure 1 shows a vector diagram of the characteristic changes of helium compared to argon resulting from each of these processes. From a starting point of equilibrium with the atmosphere (100% saturation), both helium and argon saturations increase as a result of bubbles injected from breaking waves. Ice formation increases the saturation of argon and decreases the saturation of helium, whereas ice melting has the opposite effect.


58

NOBLE GASES AND THE CRYOSPHERE 114

Helium saturation (%)

112 110

Air injection

Melting

108 106 104

mass was formed and the transformations it has experience since leaving the surface ocean. These issues are important for our understanding of the global cycling of gases between the atmosphere and the ocean and for revealing the circulation pathways of water in the Arctic, Antartic, and high latitude marginal seas. The noble gases could represent a significant addition to the set of tracers typically used to study these processes.

102 100

Rapid cooling

98 97 97.5

Freezing

See also 98 98.5 99

99.5 100 100.5 101 101.5 102

Argon saturation (%)

Figure 1 Vector diagram of helium and argon saturation changes in response to upper ocean processes.

Changes in temperature with no gas exchange with the atmosphere to balance this change can lead to modest changes in the saturations of the gases, where the gas saturations decrease with decreasing temperature and increase with increasing temperature. The trends presented here are largely qualitative indicators, since quantitative assessment of the changes depend on the exact nature of the system being studied. However, this diagram does illustrate the general magnitude of the changes that these processes have on the noble gases and conversely, the ability of the noble gases to differentiate between these effects.

Conclusions The use of the noble gases as tracers in the marine cryosphere is in its infancy. Our understanding of the interactions of the noble gases and ice have progressed from controlled, idealized laboratory conditions to natural freshwater systems and simple salt water systems, and the initial results from these studies are extremely encouraging. This technique is currently being developed more fully to provide quantitative information about the interactions of dissolved gases and ice, and to utilize the resulting effects of these interactions to trace water mass mixing and circulation in the range of dynamic ice formation environments. Water masses in the interior and deep ocean originating in ice formation and melting areas have been shown to have distinct noble gas ratios, which are largely imparted to the water mass at the time of its formation in the surface ocean. By understanding and quantifying the processes responsible for these distinct ratios, we will be able to learn much about where and how the water

Air–Sea Gas Exchange. Arctic Ocean Circulation. Bottom Water Formation. Bubbles. CFCs in the Ocean. Ice–ocean interaction. Long-Term Tracer Changes. Oxygen Isotopes in the Ocean. Polynyas. Sea Ice: Overview. Stable Carbon Isotope Variations in the Ocean. Sub Ice-Shelf Circulation and Processes. Tritium–Helium Dating. Water Types and Water Masses.

Further Reading Bieri RH (1971) Dissolved noble gases in marine waters. Earth and Planetary Science Letters 10: 329--333. Cox GFN and Weeks WF (1982) Equations for determining the gas and brine volumes in sea ice samples, USA Cold Regions Research and Engineering Laboratory Report 82-30, Hanover, New Hampshire. Craig H and Hayward T (1987) Oxygen supersaturations in the ocean: biological vs. physical contributions. Science 235: 199--202. Hood EM, Howes BL, and Jenkins WJ (1998) Dissolved gas dynamics in perennially ice-covered Lake Fryxell, Antarctica. Limnology and Oceanography 43(2): 265--272. Hood EM (1998) Characterization of Air–sea Gas Exchange Processes and Dissolved Gas/ice Interactions Using Noble Gases. PhD thesis, MIT/WHOI, 98–101. Kahane A, Klinger J, and Philippe M (1969) Dopage selectif de la glace monocristalline avec de l’helium et du neon. Solid State Communications 7: 1055--1056. Namoit A and Bukhgalter EB (1965) Clathrates formed by gases in ice. Journal of Structural Chemistry 6: 911--912. Schlosser P (1986) Helium: a new tracer in Antarctic oceanography. Nature 321: 233--235. Schlosser P, Bayer R, Flodvik A, et al. (1990) Oxygen-18 and helium as tracers of ice shelf water and water/ice interaction in the Weddell Sea. Journal of Geophysical Research 95: 3253--3263. Top Z, Martin S, and Becker P (1988) A laboratory study of dissolved noble gas anomaly due to ice formation. Geophysical Research Letters 15: 796--799. Top Z, Clarke WB, and Moore RM (1983) Anomalous neon–helium ratios in the Arctic Ocean. Geophysical Research Letters 10: 1168--1171.


NON-ROTATING GRAVITY CURRENTS P. G. Baines, CSIRO Atmospheric Research, Aspendale, VIC, Australia Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1898–1903, & 2001, Elsevier Ltd.

Gravity currents (also known as ‘density currents’) in the ocean are flows of water that are principally due to differences in density (due to differing temperature and/or salinity) between neighboring water bodies. These density differences cause lateral pressure gradients that produce the horizontal motion. They may occur on length scales ranging from centimeters to hundreds of kilometers, and they have a characteristic structure that is described below. In the oceanic environment there are three main types, as follows. Firstly, gravity currents may move along the ocean bottom, with denser (colder or saltier) water moving under a lighter water mass. Secondly, a body of lighter water may move along the ocean surface, above denser water, and thirdly, a homogeneous body of water of intermediate density may penetrate a larger stratified water body of varying density, with lighter fluid above and denser fluid below. The latter process is termed an intrusion. If these flows last for more than a significant fraction of the inertial period, the Earth’s rotation (via the Coriolis force) has an important effect on the flow, and these rotational effects are discussed elsewhere (see Rotating Gravity Currents). This article gives some examples of these flows, and then describes their basic dynamical properties for flows that move horizontally in one direction, and that move radially outward in two dimensions. It then proceeds to discuss gravity currents that flow down slopes, taking into account the effects of environmental stratification.

Examples Gravity currents exist in the ocean in a wide variety of forms. On the smallest scale of interest here, they are man-made. Prominent examples are sewage outfalls, in which effluent is piped to some distance offshore and is then released into the ocean. If the effluent is denser than the environment it may spread laterally over the local bottom, but if lighter, it may rise in a plume to the surface and spread there. Another example is effluent from power stations; this may be released on the surface, but it then behaves in a similar fashion to the sewage outfall. Oil released

on the surface (as from a shipwrecked tanker) will initially spread as a buoyant gravity current, although being immiscible with water, its mixing properties will be quite different from the previous examples. Moving to larger scale, gravity currents may be found in rivers, lakes, and estuaries. A sudden surge of water down a river into a lake or estuary can form a gravity current on the bottom or at the surface. If the lake is deep and density-stratified, and the flow into the lake is denser (e.g., colder) than the surface water, most of this inflow may begin as a downslope current, but end up as an intrusion. Several examples in the coastal environment are due to tides, which can cause the shoreward flow of buoyant fluid into a region on a broad front, or alternatively, a salt wedge of dense fluid moving inward on the bottom. In the open ocean, fronts (see Shelf Sea and Shelf Slope Fronts) that separate water of different densities can exist for periods of days, weeks, and months in much the same location. These fronts are maintained in dynamical balance by the Coriolis force, but they may exhibit gravity current character locally at times when the flow is disturbed and the geostrophic balance is broken. Sometimes turbulence within a current will cause the sediment on the bottom to become suspended in the water. When this occurs, the sediment can add to the fluid density, and the result constitutes a turbidity current. This may cause a redistribution of sediment, often to deeper water. Earthquakes may trigger submarine landslides that transport and redistribute sediment over large distances. Over timescales of millions of years, these may become conspicuous layers of sedimentary rock (turbidites) that leave a notable signature in the geological record.

Unidirectional Density Currents The dynamical essentials of the above phenomena can be encapsulated by looking at simple idealized flows that represent them. These can best be seen from simple laboratory experiments. A density current can be created in the laboratory by releasing cold water into relatively warmer water, but it is usually simpler to work instead with isothermal fresh water, and use dissolved salt to make the water denser. In Figure 1 a body of dense, dyed salty water has been released into a tank of fresh water, producing a density current moving from left to right over a horizontal surface, viewed from the side. The leading part of the current consists of a ‘head’ with an overhanging leading nose. This head may be


59

60

NON-ROTATING GRAVITY CURRENTS

In order to quantify this process, an idealized twodimensional model of a density current can be considered, in which the volume flux Q and density r1 of the dense fluid are specified at a given source location. The Reynolds number (¼ Q=n where n is kinematic viscosity) is assumed to be large. If the density of the environmental fluid is r0 , the buoyancy of the dense fluid is then g0 ¼ ðr1 r0 Þg=r0 Behind the head, the velocity v1 and thickness d1 of the dense layer are approximately constant. We then have Q ¼ v1 d1 , and from dimensional analysis alone we have: d1 BðQ2 =g0 Þ1=3 Figure 1 A vertical cross-section showing the structure of the flow in and behind the head of a gravity current advancing over a rigid bottom in a laboratory experiment, from left to right. The dense fluid has been marked by fluorescein dye. Note the dense fluid layer on bottom at the left, and the mixed layer above it. (Dyed salt water inflow density 1.05 g cm3 into fresh water, current depth of order 1–2 cm.)

regarded as a limiting form of hydraulic jump in a cold layer of dense fluid (see Overflows and Cascades), in which the depth upstream of the jump is zero. Behind the head the fluid has a three-layer structure, of which the first layer of dense fluid constitutes the main part of the current, at the bottom. This fluid moves faster than the head, and catches up with it. Inside the head it rises and is mixed with the surrounding lighter environmental fluid, and forms a density-stratified layer, spread out above the bottom current. Above this is the third layer of unmixed environmental fluid. Vorticity is produced in the upper part of the head by shear between the head and the ambient fluid, and this is also deposited in the stratified layer. This mixed stratified layer moves slowly in the direction of the current. It is highly turbulent immediately behind the head, where most of the mixing takes place, and this turbulence decays with distance from the head. Here the interfaces between the mixed layer and the dense fluid below and the environmental fluid above are generally stable, so that apart from the decaying turbulence little further significant mixing occurs. Density current heads have three-dimensional structure containing many lumps and bumps, and these shapes change continually as the head propagates. Some small parts of the head are more advanced than others, but these then disappear and are overtaken by others. This lobe and cleft structure is due to the drag on the current by the rigid lower surface over which it propagates. This retards the lowest levels of dense fluid, causing the overhanging noses. The lighter fluid beneath these noses then causes the unstable three-dimensional lobe and cleft structure.

v1 BðQg0 Þ1=3 Bðg0 d1 Þ1=2

½1

The speed vH of the head also scales with v1, with vH ov1 . Laboratory observations of flow into a deep homogeneous environment at large Reynolds numbers give: vH ¼ 1:2ðg0 d1 Þ1=2 v1 ¼ 1:4ðg0 d1 Þ1=2

½2

which are approximately constant with time for a sustained source of fluid. The same observations also show that the total height of the head is typically 3d1, and the height of the nose is about 0.15d1. These expressions apply to density current heads in general, when the buoyancy and thickness of the dense fluid approaching the head are given by g0 and d1. This steady pattern is maintained because the driving pressure gradient is retarded by loss of momentum through mixing in the head, and the following current is restrained by friction with the bottom surface and the fluid above. When the current becomes long, the thickness of the dense layer decreases gradually away from the source, providing a small pressure gradient to overcome friction. If buoyant fluid is released into denser fluid causing a gravity current that flows along the surface, the lack of stress on the surface implies that the leading nose is smooth rather than overhanging, and the lobe and cleft structure is largely absent (see Figure 2). Experiments show that these flows may be significantly affected by surface tension unless the latter is very small compared with buoyancy, but where buoyancy dominates (as in most cases in the ocean) the above dynamical relationships in eqn [2] apply with slightly different constants. Unidirectional Density Currents due to Collapse of a Dense Body of Fluid

Density currents may also be created by suddenly releasing a large body of dense fluid within a stationary environment of lighter fluid. This models a finite-sized source, and is readily simulated in a laboratory tank by raising a vertical barrier that



61

< L0 >

D0

x

Figure 2 As for Figure 1, but showing a section along a gravity current advancing to the left beneath a free surface. Note the (relative) absence of an overhanging nose. (Dyed freshwater inflow salt water of density 1.1 g cm3, current depth of order 1– 2 cm.)

separates the dense and lighter fluid. In a twodimensional situation where the fluid spreads in the x-direction only, the flow passes successively through three stages: the slumping, inertia–buoyancy, and viscous stages. In the first stage of collapse of the dense fluid (of initial depth d0, length L0, and area A ¼ A0 ¼ d0 L0 in fluid of total depth D0), termed the slumping stage, the front travels as a density current of constant speed and depth. However, the surface of the dense fluid behind this front is not horizontal, as large amplitude internal waves propagate on it (see Figure 3). These waves reflect from the left-hand end (or center, for a symmetric collapsing body), and the character of the flow changes when they catch up to the front, or density current head. If the total depth is sufficiently small, or if the upper level fluid is sufficiently strongly stratified to (partially) simulate a rigid lid, the resulting motion of the ambient fluid may affect the waves on the interface with the dense fluid, and hence affect the details of the slumping behavior. This effect may be represented in experiments by a finite total depth, as shown schematically in Figure 3. In particular, the presence of a reversed flow in the upper layer causes a corresponding reduction in the speed over the ground of the collapsing front, relative to eqn [2]. The distance xs at which the slumping stage ends and the waves reach the head, measured from the end of the tank, is then observed to be: xs =L0 ¼ 3 þ 7:4d0 =D0

½3

After the reflected waves have caught up with the density current head the flow enters the second selfsimilar stage, which is dominated by inertia,

Figure 3 Schematic diagrams showing various stages at successive times in the initial slumping phase of the collapse of a rectangular volume (shown dashed) of dense fluid with depth d0 equal to the total depth D0. The mixed region has been omitted from this sketch.

buoyancy, and mixing. Here the dense fluid collapses in the form of a rectangle of approximately uniform area, with increasing length and uniformly decreasing height d1. This rectangular uniformity is maintained by internal waves propagating outwards on the dense fluid interface. Mixing in the main body of the collapsing rectangular current is generally very small. This is because the mean gradient Richardson number at its upper boundary (on which mixing depends) is generally 40.25, implying that the mean flow is stable and does not generate local mixing (see Upper Ocean Mixing Processes). Hence the density within the collapsing rectangular body of dense fluid remains largely unaltered until the fluid enters the head, where most of the mixing takes place. The total rectangular area AðtÞ of unmixed dense fluid continually decreases because v1 > vH in eqn [2]. During this stage, the length of the dense fluid increases as ðAg0 t2 Þ1=3 , so that the velocity of the density current head decreases with time as ðAg0 t2 Þ1=3 because of the steady decrease of d1. This continues until the third stage is reached, where the dense layer becomes sufficiently thin and slow for viscous effects to become important, a regime that is not relevant here. In the second stage, dense fluid enters the head from behind where it is mixed with environmental fluid. The resulting mixture is spread out behind, in a layer over the dense layer. The rate of mixing within the head is roughly proportional to its mean height, and hence decreases with it. At early times in this stage ðx > xs Þ, the detrained mixed fluid consists mostly of the dense fluid with a small part of environmental fluid. However, as the current proceeds, this proportion reverses and near the end (ðxcxs ), nearly all the mixed fluid is environmental. When the end of the second stage has been reached (at 1=2 x ¼ xs þ 29A0 ), and mixing has effectively ceased, the total volume of mixed fluid that has been


62


produced is slightly more than twice the initial volume of dense fluid. The volume of the remaining unmixed dense fluid is quite small, so that overall, the dense fluid and environmental fluid have been mixed in approximately equal proportions.

Radial Density Currents Collapsing localized bodies of dense fluid may be constrained to spread in one direction if they are spatially confined, such as in a submarine valley, but often they are free to spread horizontally without confinement. These have a curved front or head, expanding radially away from a central source. Provided that the curvature of the front is not too large, the speed of the front and of the fluid behind it are given approximately by eqn [2]. Such flows may be modeled by axisymmetric collapsing bodies of dense fluid, which have some properties that resemble those of one-dimensional spreading. Again, there are three stages of spreading: the slumping, inertia–buoyancy balance, and viscous stages. For a body of dense fluid of initial height d0, radius R0, and volume V0, in fluid of overall depth D0, the initial collapse occurs in the ‘slumping’ stage, governed by wave propagation and adverse flow of the ambient fluid. Here the radial position R of the front increases at a constant speed in the range R0oRoRs, where Rs corresponds to xs and denotes the limit of the slumping stage. The shear between the front of dense fluid and the ambient fluid produces vorticity that is initially contained within the mixing head of the current, and as the front expands this vorticity increases by vortex stretching. This initial vortex intensifies as it expands, and may be much deeper than the following fluid. For a small initial volume of dense fluid, this expanding vortex ring of dense mixed fluid is all that is produced. But, for a larger initial volume, as the dense fluid expands and the head moves forward, the initial vortex progressively breaks up and is subsumed into the mixed layer. New vorticity is continually created at the head, as for the unidirectional currents, but this is associated with newly mixed fluid and the stretching is much less than that in the initial vortex. In this second inertial–buoyancy phase (where R>Rs), gravity waves maintain approximately uniform depth of the body of dense fluid, and it collapses in the form of an axisymmetric pillbox, with the volume slowly decreasing due to mixing and detrainment behind the circular head. If the volume of dense fluid at any given time is V ¼ pR2 d1 , and the properties of the front are given by eqn [2], the radius R increases as: RBðg0 VÞ1=4 t1=2

½4

V slowly decreases as dense fluid entering the head is mixed, and this dynamical regime continues until (in the laboratory) the viscous regime is reached. If the source of dense fluid is maintained with constant volume flux Q1, the initial vortex and head form as above, but the flow behind it evolves differently, as the depth d1 and velocity v1 of the following radial outflow are not uniform and decrease with radial distance r. The only significant mixing occurs behind the head, and for steady flow, dimensional analysis gives: 2 1=3 Q Q1 g0 1=3 v1 B ½5 d1 B 0 12 gr r In the inertio-gravity range, the radial speed of the head is then given by: Q1 g0 1=4 ½6 vH ¼ 0:63 t

Density Currents Down Slopes When the terrain is not horizontal but slopes downwards at angles greater than about 51, buoyancy acts directly to drive the current downward, and this provides stronger forcing than the indirect effect of establishing a horizontal pressure gradient as described above. The downslope buoyancy force is now g0 siny per unit mass of dense fluid, where y is the angle between the slope of the terrain and the horizontal. The onset of a steady source of dense fluid at the top of the slope leads to the formation of a density current head similar to that described above. In the current following the head the physics is different because the flow is now unstable, unless the slope angle is very small, and mixing now occurs along the whole length of the density current, not just behind the head. After the passage of the head an approximately steady flow is established, with the buoyancy force being balanced primarily by the entrainment of environmental fluid into the current from above and, to a much lesser extent, by the drag on the bottom surface. For a homogeneous environment, both the size of the head and the thickness of the following current increase with downslope distance. Behavior of the Head

For a constant supply of dense fluid of buoyancy g00 and flow rate Q0 at the top of the slope, the speed of the head vH is almost independent of slope angle and is given approximately by: vH ¼ ð1:570:2Þðg00 Q0 Þ1=3


53 r y r 903

½7


However, in spite of this uniform speed, the size of the head increases as it moves downslope, at the rate: dH=dx ¼ 0:72y=p

½8

where H is the height (or thickness) of the head, x is downslope distance, and y is in radians. The increased buoyancy at steeper slopes is balanced by increased entrainment that acts to keep the head speed constant, regardless of downslope distance and slope angle. Entrainment in Downslope Flows

Behind the head the flow is unstable, and mixing with the overlying fluid results. This process becomes stronger with increasing slope angle y. The mixed fluid above the dense layer is also denser than the environment, and hence it also moves downslope under gravity but at a reduced speed. The thickness and volume of this mixed layer both increase with downslope distance. This combined flow may now be regarded as a single entity, and the net downslope buoyancy flux is constant with x and t, and equal to g0 0Q0. The net entrainment of environmental fluid into this overall downslope flow may be described by an entrainment coefficient E which is a function of the bulk Richardson number Ri, defined by: Ri ¼ g0 d cosy=U2

½9

where U is the mean velocity and d the mean thickness of the total flow (see Upper Ocean Mixing Processes). At each level, the velocity of inflow we of the environmental fluid into the downflow is given by: we ¼ EðRi ÞU

½10

63

buoyancy frequency (see Upper Ocean Vertical Structure) of the ambient stratification, a depth D may be identified below the source, defined by N2 ¼ g0 0/D, where the ambient density equals the inflowing density. If there were no mixing, all of the inflowing fluid would be expected to reach and spread horizontally at this level. The speed of the head of a downslope gravity current into a stratified environment is again approximately constant over most of its distance travelled, and scales with eqn [7] above. The main differences from the homogeneous case concern the following current, which is discussed next. For slope angles less than about 201, which covers most cases of interest in the ocean, it is appropriate to regard the main dense current and the mixed layer above it as separate entities. In laboratory experiments a clear interface is visible between them, as seen in Figure 4, although there are turbulent fluxes across it. These flows are governed by two dimensionless parameters – a bulk Richardson number Ri defined as in eqn [9] but based on the dense layer only, and a parameter M, defined by: M ¼ QN3 =g02

½11

which is a measure of the effect of the ambient stratification. The dense layer is observed to have approximately uniform thickness over most of its length, but its velocity mostly decreases with downslope distance. It loses fluid to the mixed layer by detrainment, but also entrains fluid from it so that its density progressively decreases, and the fluid that remains in the dense current reaches its ambient level where it spreads out, at a height somewhere above the level of D below the source. Entrainment into this dense layer may be expressed in terms of an

where U is the mean velocity of the downflow. E decreases monotonically with increasing Ri, from E ¼ 0.075 at Ri ¼ 0 to very small values for Ri 40.8. In these flows, Ri is approximately constant with downslope distance, and decreases with increasing slope angle. Experiments show that U, and hence we, are also constant, so that the downslope flux Q and the lateral spreading increase linearly with distance. Downslope Flows into Stratified Environments

If the environmental fluid is density stratified, the effects of this stratification on density currents flowing over a horizontal surface are mostly limited to its effects on the surrounding flow, including the generation of internal waves (see Internal Waves). However, for flows down slopes, if ambient stratification is present it is a major parameter. If N is the

Figure 4 An example of flow of a layer of dense fluid down a slope at 61 to the horizontal, in a density-stratified environment, showing the structure of the interface. The dense fluid is the lightly colored dyed layer close to the boundary. Note the filaments extending from it, and the mixed fluid above. CM ¼ 0.015, with current depth of order 1 cm, buoyancy frequency of stratification N ¼ 1.24 rad s1, Q ¼ 2.9 cm2 s1, g0 ¼ 19.4 cm s2.)


64


entrainment coefficient that depends on Ri, in a manner similar to eqn [10]. There is also a loss of dense fluid to the mixed layer, and an exchange between the mixed layer and the environment. Fluid may leave the mixed layer to find its own neutral level, and this occurs continuously along the path of the current. Hence dense fluid from the source may be distributed over a range of depths, and not all at one level. This detrainment into the environment depends on both Ri and M – larger M implies larger detrainment, to the extent that all the dense fluid may be detrained before it reaches its ambient level. The above sections describe the ‘pure’ dynamical form of a gravity current, in various different situations. But when gravity currents occur in the ocean, they are often strongly influenced by other factors over most of their life cycles. In particular, the flow of dense fluid down continental slopes is usually affected by the Earth’s rotation, which means that the flow is at least partly in geostrophic balance, with an Ekman layer next to the bottom boundary (see Rotating Gravity Currents). This applies in particular to the deep flows through Denmark Strait and around Antarctica, that drive the overturning thermohaline circulation. Also, sediment suspension is a common cause of submarine gravity currents, and it may also be a factor in these deep flows.

See also Internal Waves. Rotating Gravity Currents. Shelf Sea and Shelf Slope Fronts. Upper Ocean Mixing Processes. Upper Ocean Vertical Structure.

Further Reading Baines PG (1995) Topographic Effects in Stratified Flows. Cambridge: Cambridge University Press. Baines PG and Condie S (1998) Observations and modeling of Antarctic downslope flows: a review. In: Jacobs SS and Weiss R (eds.) Ocean, Ice and Atmosphere: Interactions at the Antarctic Continental Margin, AGU Antarctic Research Series, vol. 75, pp. 29--49. Washington, DC: American Geophysical Union. Simpson JE (1997) Gravity Currents. Cambridge: Cambridge University Press. Turner JS (1986) Turbulent entrainment: the development of the entrainment assumption, and its application to geophysical flows. Journal of Fluid Mechanics 173: 431--471.


NORTH ATLANTIC OSCILLATION (NAO) J. W. Hurrell, National Center for Atmospheric Research, Boulder, CO, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1904–1911, & 2001, Elsevier Ltd.

Introduction Simultaneous variations in weather and climate over widely separated points on Earth have long been noted in the meteorological literature. Such variations are commonly referred to as ‘teleconnections’. In the extratropics, teleconnections link neighboring regions mainly through the transient behavior of atmospheric planetary-scale waves. Consequently, some regions may be cooler than average, while thousands of kilometers away warmer conditions prevail. Though the precise nature and shape of these structures vary to some extent according to the statistical methodology and the data set employed in the analysis, consistent regional characteristics that identify the most conspicuous patterns emerge. Over the middle and high latitudes of the northern hemisphere, a dozen or so distinct teleconnection patterns can be identified during boreal winter. One of the most prominent is the North Atlantic Oscillation (NAO). The NAO dictates climate variability from the eastern seaboard of the USA to Siberia and from the Arctic to the subtropical Atlantic. This widespread influence indicates that the NAO is more than just a North Atlantic phenomenon. In fact, it has been suggested that the NAO is the regional manifestation of a larger scale (hemispheric) mode of variability known as the Arctic Oscillation (see below). Regardless of terminology, meteorologists for more than two centuries have noted the pronounced influence of the NAO on the climate of the Atlantic basin. Variations in the NAO are important to society and the environment. Through its control over regional temperature and precipitation variability, the NAO directly impacts agricultural yields, water management activities, and fish inventories among other things. The NAO accounts for much of the interannual and longer-term variability evident in northern hemisphere surface temperature, which has exhibited a warming trend over the past several decades to values that are perhaps unprecedented over the past 1000 years. Understanding the processes that govern variability of the NAO is therefore of high priority,

especially in the context of global climate change. This article defines the NAO and describes its relationship to variations in surface temperature and precipitation, as well as its impact on variability in the North Atlantic Ocean and on the regional ecology. It concludes with a discussion of the mechanisms that might influence the amplitude and timescales of the NAO, including the possible roles of the stratosphere and the ocean.

What is the North Atlantic Oscillation? Like all atmospheric teleconnection patterns, the NAO is most clearly identified when time averaged data (monthly or seasonal) are examined, since time averaging reduces the ‘noise’ of small-scale and transient meteorological phenomena not related to large-scale climate variability. Its spatial signature and temporal variability are most often defined through the regional sea level pressure field, for which some of the longest instrumental records exist. The NAO refers to a north–south oscillation in atmospheric mass with centers of action near Iceland and over the subtropical Atlantic from the Azores across the Iberian Peninsula. Although it is the only teleconnection pattern evident throughout the year in the northern hemisphere, its amplitude is largest during boreal winter when the atmosphere is dynamically the most active. During the months December through March, for instance, the NAO accounts for more than one-third of the total variance in sea level pressure over the North Atlantic. A time series (or index) of more than 100 years of wintertime NAO variability and the spatial signature of the oscillation are shown in Figures 1 and 21 Differences of 415 hPa occur across the North Atlantic between the two phases of the NAO. In the socalled positive phase, higher than normal surface pressures south of 551N combine with a broad region of anomalously low pressure throughout the Arctic. Because air flows counterclockwise around low pressure and clockwise around high pressure in the northern hemisphere, this phase of the oscillation is associated with stronger than average westerly winds across the middle latitudes of the Atlantic onto

1 More sophisticated and objective statistical techniques, such as eigenvector analysis, yield time series and spatial patterns of average winter sea level pressure variability very similar to those shown in Figures 1 and 2.


65

66

NORTH ATLANTIC OSCILLATION (NAO)

6

4

(Ln _ Sn)

2

0 _2 _4 _6 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Figure 1 Winter (December–March) index of the NAO based on the difference of normalized sea level pressure between Lisbon, Portugal, and Stykkisholmur/Reykjavik, Iceland from 1864 to 2000. The average winter sea level pressure data at each station were normalized by division of each seasonal pressure by the long-term mean (1864–1983) standard deviation. The heavy solid line represents the index smoothed to remove fluctuations with periods o4 years.

December _ March

SLP

(hPa)

150˚W

150˚E

120˚E

120˚W

0 _2 0 90˚W 2

_2

_6

0

2

60˚W

90˚E

0

_2

60˚E

2

4 2

4

30˚W

30˚E 0˚

Figure 2 Difference in sea level pressure between years with an NAO index value 41.0 and those with an index value o minus low index winters) since 1899. The contour increment is 2 hPa and negative values are dashed.

Europe, with anomalous southerly flow over the eastern USA and anomalous northerly flow across western Greenland, the Canadian Arctic, and the Mediterranean.

1.0 (high

The NAO is also readily apparent in meteorological data throughout the depth of the troposphere, and its variability is significantly correlated with changes in the strength of the winter polar vortex in



the stratosphere of the northern hemisphere. Within the lower stratosphere, the leading pattern of geopotential height variability is also characterized by a seesaw in mass between the polar cap and the middle latitudes, but with a much more zonally symmetric (or annular) structure than in the troposphere. When heights over the polar region are lower than normal, heights at nearly all longitudes in middle latitudes are higher than normal. In this phase, the stratospheric westerly winds that encircle the pole are enhanced and the polar vortex is ‘strong’ and anomalously cold. It is this annular mode of variability that has been termed the Arctic Oscillation. However, the signature of the stratospheric Arctic Oscillation in winter sea level pressure data looks very much like the anomalies associated with the NAO, with centers of action over the Arctic and the Atlantic (Figure 2). The ‘annular’ character of the Arctic Oscillation in the troposphere, therefore, reflects the vertically coherent fluctuations throughout the Arctic more than any coordinated behavior in the middle latitudes outside of the Atlantic basin. That the NAO and Arctic Oscillation reflect essentially the same mode of tropospheric variability is emphasized by the fact that their time series are nearly identical, with differences depending mostly on the details of the analysis procedure. There is little evidence for the NAO to vary on any preferred timescale (Figure 1). Large changes can occur from one winter to the next, and there is also a considerable amount of variability within a given winter season. This is consistent with the notion that much of the atmospheric circulation variability in the form of the NAO arises from processes internal to the atmosphere, in which various scales of motion interact with one another to produce random (and thus unpredictable) variations. On the other hand, there are also periods when anomalous NAO-like circulation patterns persist over many consecutive winters. In the Icelandic region, for instance, sea level pressure tended to be anomalously low during winter from the turn of the century until about 1930 (positive NAO index), while the 1960s were characterized by unusually high surface pressure and severe winters from Greenland across northern Europe (negative NAO index). A sharp reversal has occurred over the past 30 years, with strongly positive NAO index values since 1980 and sea level pressure anomalies across the North Atlantic and Arctic that resemble those in Figure 2. In fact, the magnitude of the recent upward trend is unprecedented in the observational record and, based on reconstructions using paleoclimate and model data, perhaps over the past several centuries as well. Whether such low frequency (interdecadal) NAO variability arises from

67

interactions of the North Atlantic atmosphere with other, more slowly varying components of the climate system (such as the ocean), whether the recent upward trend reflects a human influence on climate, or whether the longer timescale variations in the relatively short instrumental record simply reflect finite sampling of a purely random process, are topics of considerable current interest.

Impacts of the North Atlantic Oscillation Temperature

The NAO exerts a dominant influence on wintertime temperatures across much of the northern hemisphere. Surface air temperature and sea surface temperature (SST) across wide regions of the North Atlantic Ocean, North America, the Arctic, Eurasia, and the Mediterranean are significantly correlated with NAO variability.2 Such changes in surface temperature (and related changes in rainfall and storminess) can have significant impacts on a wide range of human activities, as well as on marine and terrestrial ecosystems. When the NAO index is positive, enhanced westerly flow across the North Atlantic during winter moves relatively warm (and moist) maritime air over much of Europe and far downstream across Asia, while stronger northerlies over Greenland and north-eastern Canada carry cold air southward and decrease land temperatures and SST over the northwest Atlantic (Figure 3). Temperature variations over North Africa and the Middle East (cooling), as well as North America (warming), associated with the stronger clockwise flow around the subtropical Atlantic high-pressure center are also notable. The pattern of temperature change associated with the NAO is important. Because the heat storage capacity of the ocean is much greater than that of land, changes in continental surface temperatures are much larger than those over the oceans, so they tend to dominate average northern hemisphere (and global) temperature variability. Especially given the large and coherent NAO signal across the Eurasian continent from the Atlantic to the Pacific (Figure 3), it is not surprising that NAO variability explains about one-third of the northern hemisphere interannual surface temperature variance during winter. The strength of the link between the NAO and northern hemisphere temperature variability has also

2

Sea surface temperatures (SSTs) are used to monitor surface air temperature over the oceans because intermittent sampling is a major problem and SSTs have much greater persistence.


68

NORTH ATLANTIC OSCILLATION (NAO) _

December _ March

(×10 1°C)

150°W 0

150°W

0 120°W 0

0

4 6

4

2

8

6 4

90°W

0 2

120°W

2

0

90°E

10 _10 _4

14

2

8 12

0 60°W

0

4

_2 60°E

6 _4

0

30°E

30°W 0

Figure 3 Changes in land surface and sea surface temperatures ( 10 11C) corresponding to unit deviations of the NAO index for the winter months (December–March) from 1935 to 1999. The contour increment is 0.21C. Temperature changes 40.21C are indicated by dark shading, and those o 0.21C are indicated by light shading. Regions with insufficient data are not contoured.

added to the debate over our ability to detect and distinguish between natural and anthropogenic climate change. Since the early 1980s, winter temperatures over much of North America and Eurasia have been considerably warmer than average, while temperatures over the northern oceans have been slightly colder than average. This pattern, which has contributed substantially to the well-documented warming trend in northern hemisphere and global temperatures over recent decades, is quite similar to winter surface temperature changes projected by computer models forced with increasing concentrations of greenhouse gases and aerosols. Yet, it is clear from the above discussion that a significant fraction of this temperature change signal is associated with the recent upward trend in the NAO (Figures 1 and 3). How anthropogenic climate change might influence modes of natural climate variability such as the NAO, and the nature of the relationships between increased radiative forcing and interdecadal variability of these modes, remain central research questions. Precipitation and Storms

Changes in the mean circulation patterns over the North Atlantic are accompanied by changes in the

intensity and number of storms, their paths, and their associated weather. Here, the term ‘storms’ refers to atmospheric disturbances operating on timescales of about a week or less. During winter, a well-defined storm track connects the North Pacific and North Atlantic basins, with maximum storm activity over the oceans. The details of changes in storminess differ depending on the analysis method and whether the focus is on surface or upper-air features. Generally, however, positive NAO index winters are associated with a northward shift in the Atlantic storm activity, with enhanced activity from southern Greenland across Iceland into northern Europe and a modest decrease in activity to the south. The latter is most noticeable from the Azores across the Iberian Peninsula and the Mediterranean. Positive NAO winters are also typified by more intense and frequent storms in the vicinity of Iceland and the Norwegian Sea. The ocean integrates the effects of storms in the form of surface waves, so that it exhibits a marked response to long-lasting shifts in the storm climate. The recent upward trend toward more positive NAO index winters has been associated with increased wave heights over the north-east Atlantic and decreased wave heights south of 401N. Such



changes have consequences for the operation and safety of shipping, offshore industries, and coastal development. Changes in the mean flow and storminess associated with swings in the NAO are also reflected in pronounced changes in the transport and convergence of atmospheric moisture and, thus, the distribution of evaporation and precipitation. Evaporation (E) exceeds precipitation (P) over much of Greenland and the Canadian Arctic during high NAO index winters (Figure 4), where changes between high and low NAO index states are on the order of 1 mm d 1. Drier conditions of the same magnitude also occur over much of central and southern Europe, the Mediterranean, and parts of the Middle East, whereas more precipitation than normal falls from Iceland through Scandinavia. This pattern, together with the upward trend in the NAO index since the late 1960s (Figure 1), is consistent with recent observed changes in precipitation over much of the Atlantic basin. One of the few regions of the world where glaciers have not exhibited a retreat over the past several decades is in Scandinavia, where more than average amounts of precipitation have been typical of many winters since the early 1980s. In contrast, over the Alps, snow depth and duration in recent winters have been among the lowest recorded this century, and the retreat of Alpine glaciers has been widespread. Severe drought has persisted throughout parts of Spain and Portugal as well. As far east as Turkey, river runoff is significantly correlated with NAO variability. There is also some observational and modeling evidence

E_P

69

of a declining precipitation rate over much of the Greenland Ice Sheet over the past two decades, although measurement uncertainties are large. Ocean Variability

It has long been recognized that fluctuations in SST and the strength of the NAO are related, and there are clear indications that the North Atlantic Ocean varies significantly with the overlying atmosphere. The leading pattern of SST variability during winter consists of a tripolar structure marked, in one phase, by a cold anomaly in the subpolar North Atlantic, a warm anomaly in the middle latitudes centered off Cape Hatteras, and a cold subtropical anomaly between the equator and 301N. This structure suggests that the SST anomalies are driven by changes in the surface wind and air–sea heat exchanges associated with NAO variations. The relationship is strongest when the NAO index leads an index of the SST variability by several weeks, which highlights the well-known result that extratropical SST responds to atmospheric forcing on monthly and seasonal timescales. Over longer periods, persistent SST anomalies also appear to be related to persistent anomalous patterns of SLP (including the NAO), although the mechanisms which produce SST changes on decadal and longer time-scales remain unclear. Such fluctuations could primarily be the local oceanic response to atmospheric decadal variability. On the other hand, nonlocal dynamical processes in the ocean could also be contributing to the SST variations.

_

December_ March

(mm d 1)

80˚N

0.6

_ 1.2

_ 0.6

60˚N

_ 0.6 0

0

0.6 40˚N

0

0

0.6

_ 0.6

0

0.6 0 20˚N 90˚W

60˚W

30˚W

0˚

30˚E

60˚E

Figure 4 Difference in evaporation (E) minus precipitation (P) between years with an NAO index value 41.0 and those with an index value o 1.0 (high minus low index winters) since 1958. The contour increment is 0.3 mm d 1. Differences 40.3 mm d 1 are indicated by dark shading, and those o0.3 mm d 1 are indicated by light shading.


70


Subsurface ocean observations more clearly depict long-term climate variability, because the effect of the annual cycle and month-to-month variability in the atmospheric circulation decays rapidly with depth. These measurements are much more limited than surface observations, but over the North Atlantic they too indicate fluctuations that are coherent with the winter NAO index to depths of 400 m. The ocean’s response to NAO variability also appears to be evident in changes in the distribution and intensity of winter convective activity in the North Atlantic. The convective renewal of intermediate and deep waters in the Labrador Sea and the Greenland/ Icelandorwegian Seas contribute significantly to the production and export of North Atlantic Deep Water and, thus, help to drive the global thermohaline circulation. The intensity of winter convection at these sites is not only characterized by large interannual variability, but also interdecadal variations that appear to be synchronized with variations in the NAO (Figure 1). Deep convection over the Labrador Sea, for instance, was at its weakest and shallowest in the postwar instrumental record during the late 1960s. Since then, Labrador Sea Water has become progressively colder and fresher, with intense convective activity to unprecedented ocean depths (42300 m) in the early 1990s. In contrast, warmer and saltier deep waters in recent years are the result of suppressed convection in the Greeland/Icelandorwegian Seas, whereas intense convection was observed there during the late 1960s. There has also been considerable interest in the occurrence of low salinity anomalies that propagate around the subpolar gyre of the North Atlantic. The most famous example is the Great Salinity Anomaly (GSA). The GSA formed during the extreme negative phase of the NAO in the late 1960s, when clockwise flow around anomalously high pressure over Greenland fed record amounts of fresh water through the Denmark Strait into the subpolar North Atlantic ocean gyre. There have been other similar instances as well, and statistical analyses have revealed that these propagating salinity modes are closely connected to a pattern of atmospheric variability strongly resembling the NAO. Sea Ice

The strongest interannual variability of Arctic sea ice occurs in the North Atlantic sector. The sea ice fluctuations display an oscillation in ice extent between the Labrador and Greenland Seas. Strong interannual variability is evident in the sea ice

changes over the North Atlantic, as are longer-term fluctuations, including a trend over the past 30 years of diminishing (increasing) ice concentration during boreal winter east (west) of Greenland. Associated with the sea ice fluctuations are large-scale changes in sea level pressure that closely resemble the NAO. When the NAO is in its positive phase, the Labrador Sea ice boundary extends farther south, while the Greenland Sea ice boundary is north of its climatological extent. Given the implied surface wind changes (Figure 2), this is qualitatively consistent with the notion that sea ice anomalies are directly forced by the atmosphere, either dynamically via wind-driven ice drift anomalies, or thermodynamically through surface air temperature anomalies (Figure 3). The relationship between the NAO index (Figure 1) and an index of the North Atlantic ice variations is indeed strong, although it does not hold for all individual winters. This last point illustrates the importance of the regional atmospheric circulation in forcing the extent of sea ice. Ecology

Changes in the NAO have a wide range of effects on North Atlantic ecosystems. Temperature is one of the primary factors, along with food availability and spawning grounds, in determining the large-scale distribution pattern of fish and shellfish. Changes in SST and winds associated with changes in the NAO have been linked to variations in the production of zooplankton, as well as to fluctuations in several of the most important fish stocks across the North Atlantic. This includes not only longer-term changes associated with interdecadal NAO variability, but also interannual signals as well. Over land, fluctuations in the strength of the NAO have been linked to variations in plant phenology. In Norway, for example, most plant species have been blooming from 2–4 weeks earlier in recent years because of increasingly warm and wet winters, and many species have been blooming longer. Winter climate directly impacts the growth, reproduction, and demography of many animals, as well. European amphibians and birds have been breeding earlier over the past two to three decades, and these trends have been attributed to earlier growing seasons and increased forage availability. Variations in the NAO index are also significantly correlated with the growth, development, fertility, and demographic trends of large mammals from North America to northern Europe, such as northern ungulates and Canadian lynx.



71

What are the Mechanisms that Govern influence on the Earth’s surface climate. Volcanic North Atlantic Oscillation Variability? aerosols act to enhance north–south temperature Although the NAO is an internal mode of variability of the atmosphere, surface, stratospheric or anthropogenic processes may influence its phase and amplitude. At present there is no consensus on the process or processes that most influence the NAO, especially on long (interdecadal) timescales. It is quite possible that NAO variability is affected by one or more of the mechanisms described below. Atmospheric Processes

Atmospheric general circulation model (AGCMs)3 provide strong evidence that the basic structure of the NAO results from the internal, nonlinear dynamics of the atmosphere. The observed spatial pattern and amplitude of the NAO are well simulated in AGCMs forced with fixed climatological annual cycles of solar insolation and SST, as well as fixed atmospheric trace gas composition. The governing dynamical mechanisms are interactions between the time-mean flow and the departures from that flow. Such intrinsic atmospheric variability exhibits little temporal coherence and, indeed, the timescales of observed NAO variability (Figure 1) do not differ significantly from this reference. A possible exception is the interdecadal NAO variability, especially the strong trend toward the positive index polarity of the oscillation over the past 30 years. This trend exhibits a high degree of statistical significance relative to the background interannual variability in the observed record; moreover, multi-century AGCM experiments like those described above do not reproduce interdecadal changes of comparable magnitude. A possible source of the trend could be through a connection to processes that have affected the strength of the atmospheric circulation in the lower stratosphere. During winters when the stratospheric westerlies are enhanced, the NAO tends to be in its positive phase. There is a considerable body of evidence to support the notion that variability in the troposphere can drive variability in the stratosphere, but it also appears that some stratospheric control of the troposphere may also occur. The atmospheric response to strong tropical volcanic eruptions provides some evidence for a stratospheric

3 Atmospheric general circulation models consist of a system of equations that describe the large-scale atmospheric balances of momentum, heat, and moisture, with schemes that approximate small-scale processes such as cloud formation, precipitation, and heat exchange with the sea surface and land.

gradients in the lower stratosphere by absorbing solar radiation in lower latitudes. In the troposphere, the aerosols exert only a very small direct influence. Yet, the observed response following eruptions is not only lower geopotential heights over the pole with stronger stratospheric westerlies, but also a positive NAO-like signal in the tropospheric circulation. Reductions in stratospheric ozone and increases in greenhouse gas concentrations also appear to enhance the meridional temperature gradient in the lower stratosphere, leading to a stronger polar vortex. It is possible, therefore, that the upward trend in the NAO index in recent decades (Figure 1) is associated with trends in either or both of these quantities. Indeed, a decline in the amount of ozone poleward of 401N has been observed during the last two decades, and the stratospheric polar vortex has become colder and stronger. Ocean Forcing of the Atmosphere

In the extratropics, it is clear that the atmospheric circulation is the dominant driver of upper ocean thermal anomalies. A long-standing issue, however, has been the extent to which anomalous extratropical SST feeds back to affect the atmosphere. Most evidence suggests that this effect is quite small compared with internal atmospheric variability. Nevertheless, the interaction between the ocean and atmosphere could be important for understanding the details of the observed amplitude of the NAO, its interdecadal variability, and the prospects for meaningful predictability. While intrinsic atmospheric variability is random in time, theoretical and modeling evidence suggest that the ocean can respond to it with marked persistence or even oscillatory behavior. On seasonal and interannual timescales, for example, the large heat capacity of the upper ocean can lead to slower changes in SST relative to the faster, stochastic atmospheric forcing. On longer timescales, SST observations display a myriad of variations. Middle and high latitude SSTs over the North Atlantic, for example, were colder than average during the 1970s and 1980s, but warmer than average from the 1930s through the 1950s. Within these periods, shorterterm variations in SST are also apparent, such as a dipole pattern that fluctuates on approximately decadal time scales with anomalies of one sign east of Newfoundland, and anomalies of opposite polarity off the south-east coast of the USA. A key to whether or not changes in the state of the NAO reflect these variations in the state of the ocean


72


surface is the sensitivity of the atmosphere to middle and high latitude SST and upper ocean heat content anomalies. Most AGCM studies show weak responses to extratropical SST anomalies, with sometimes contradictory results. Yet, some AGCMs, when forced with the time history of observed, global SSTs and sea ice concentrations over the past 50 years or so, show modest skill in reproducing aspects of the observed NAO behavior, especially its interdecadal fluctuations (Figure 1). Such results do not necessarily imply, however, that the extratropical ocean is behaving in anything other than a passive manner. It could be, for instance, that long-term changes in tropical SSTs force a remote atmospheric response over the North Atlantic, which in turn drives changes in extratropical SSTs and sea ice. Some model studies indicate a sensitivity of the North Atlantic atmosphere to tropical SST variations, including variations over the tropical Atlantic which are substantial on both interannual and interdecadal timescales. The response of the extratropical North Atlantic atmosphere to changes in tropical and extratropical SST distributions, and the role of land processes and sea ice in producing atmospheric variability, are problems which are currently being addressed. Until these are better understood, it is difficult to evaluate the realism of more complicated scenarios that rely on truly coupled interactions between the atmosphere, ocean, land, and sea ice to produce North Atlantic climate variability. It is also difficult to evaluate the extent to which interannual and longerterm variations of the NAO might be predictable.

Glossary Great salinity anomaly A widespread freshening of the upper 500–800 m layer of the far northern North Atlantic Ocean, traceable around the subpolar gyre from its origins north of Iceland in the mid-to-late 1960s until it return to the Greenland Sea in the early 1980s.

Fisheries and Climate. Heat and Momentum Fluxes at the Sea Surface. North Sea Circulation. Ocean Circulation. Open Ocean Convection. Plankton and Climate. Sea Ice. Sea Ice: Overview. Wave Generation by Wind. Wind Driven Circulation.

Further Reading Appenzeller C, Stocker TF, and Anklin M (1998) North Atlantic oscillation dynamics recorded in Greenland ice cores. Science 282: 446--449. Dickson B (1999) All change in the Arctic. Nature 397: 389--391. Hurrell JW (1995) Decadal trends in the North Atlantic Oscillation regional temperatures and precipitation. Science 269: 676--679. Kerr RA (1997) A new driver for the Atlantic’s moods and Europe’s weather? Science 275: 754--755. National Research Council (1998) Decade-to-century scale climate variability and change. A science strategy. Washington: National Academy Press. Rodwell MJ, Rowell DP, and Folland CK (1999) Oceanic forcing of the wintertime North Atlantic Oscillation and European climate. Nature 398: 320--323. Shindell DT, Miller RL, Schmidt G, and Pandolfo L (1999) Simulation of recent northern winter climate trends by greenhouse gas forcing. Nature 399: 452--455. Stenseth NC and Co-authors 1999 (1999) Common dynamic structure of Canadian lynx populations within three climatic regions. Science 285: 1071--1073. Sutton R and Allen MR (1997) Decadal predictability of North Atlantic sea surface temperature and climate. Nature 388: 563--567. Thompson DWJ and Wallace JM (1998) The Arctic oscillation signature in the wintertime geopotential height and temperature fields. Geophysical Research Letters 25: 1297--1300. World Climate Research Programme (1998) The North Atlantic Oscillation. Climate Variability and Predictability (CLIVAR) Initial Implementation Plan. WCRP no. 103, WMO/TD no. 869, ICPO no. 14, 163–192.

See also Abrupt Climate Change. Arctic Ocean Circulation. Current Systems in the Atlantic Ocean. Elemental Distribution: Overview. Evaporation and Humidity.


NORTH SEA CIRCULATION M. J. Howarth, 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 1912–1921, & 2001, Elsevier Ltd.

Introduction Currents in the North Sea, as in any continental shelf sea, occur in response to forcing by tides, winds, density gradients (arising from freshwater input), and pressure gradients. For most of the North Sea the dominant motion is tidal, with the next most significant motion generated by wind forcing. The response is determined by the sea’s topography and bathymetry and is modified by, and modifies, its density distribution. Currents occur over a range of scales – a practical minimum is indicated in each case, ignoring waves and turbulence which are outside the scope of this article. For time the range is from minutes to years, horizontally in space from o1 km to basin-wide (1000 km), and vertically from 1 m to the full water depth. Current amplitudes range generally from 0.01 m s 1 to 2 m s 1 (in a very few places to in excess of 5 m s 1). The North Sea has been extensively studied for more than 100 years and its physical oceanography has been widely reviewed (see Further Reading section).

Topography The North Sea is a semi-enclosed, wide continental shelf sea (Figure 1). It stretches southward for about 1000 km from the Atlantic Ocean in the north, and is about 500 km broad, from west to east. The northern boundary is composed of two connections. Firstly, the main connection with the ocean at the shelf edge is between the Shetland Islands and Norway. Secondly, from the mainland of Scotland to the Orkney and Shetland Islands via the Pentland Firth and the Fair Isle Channel, the connection is with the continental shelf seas to the west and north of Scotland. The North Sea has two other, narrow, connections. One is at its southern end, via the Dover Strait to the relatively salty English Channel (and ultimately again Atlantic) water. The second, on its eastern side, is via the Skagerrak and Kattegat to fresh Baltic water. Topographically the North Sea can be split into three (Figure 2). In the south, bounded by about 541N and the shallow Dogger Bank, depths are o40 m.

Secondly, to the north of this the water depth increases from 40 m to 200 m at the shelf edge. Thirdly, the deep Norwegian Trench is in the north east, penetrating from the Norwegian Sea southward along the coast of Norway. The deepest depths in the trench, about 750 m, occur in the Skagerrak, off southern Norway and western Sweden, whilst the minimum axial depth is about 225 m off western Norway. This has a significant impact on the sea’s dynamics.

Meteorology The North Sea lies at temperate latitudes, between 511N and 621N, and so experiences pronounced seasonal changes in meteorological conditions. From September to April, in particular, it is exposed to the effect of a series of storms. These usually track eastward to the north of the British Isles, taking about a day to pass by the North Sea, accompanied by strong winds from the south west, west, and north west. Although these are the dominant wind directions, strong winds can blow from any direction. To some extent the British Isles shelter the North Sea from the wind’s full effect since the only wind direction uninterrupted by land is from the north. (This is more significant for waves than currents.) The size of the storms is generally of the same order as that of the North Sea, or larger, and they can recur every few days. On average gale force winds blow on 30 days per year. The extreme ‘one-in-50-year’ hourly mean wind speeds are estimated to decrease from 39 m s 1 in the north to 32 m s 1 in the south, slightly stronger in the center of the North Sea compared with near the coast.

River Discharges The main input of fresh water into the North Sea is from rivers, including the Rhine and Elbe, discharging along its southern coast (Belgium, The Netherlands, Germany). On average this input is 4000 m3 s 1, but it is highly variable from day to day and also from year to year. This fresh water has a significant impact on the salinity and density distribution (lower salinities along the continental coast and in the German Bight) and ultimately on the longterm circulation. A further 1500 m3 s 1 of fresh water on average is discharged from Scottish and English rivers. More fresh water flows into the Kattegat from the Baltic, on average 15 000 m3 s 1 in the form of a low salinity (8.7 PSU) flow of


73

NORTH SEA CIRCULATION

tla

74

Norway

e _ Sh e o r

h enc gian Tr Norwe

Fa

nd

el nn a Ch

Shetland

Fai r

Isle

Ch

ann

el

Orkney

60°N

Pentland Firth Skagerrak

Scotland Denmark

55°N Sellafield

German Bight

Dogger Bank

Flamborough head Humber

s

Elbe

la er

Weser Germany

Ne

th

England

nd

Southern Bight

Thames

ov

er

ra it

Rhine Meuse Scheldt

D

St

Be

lgiu

m

Cap de la Hague

5°W

50°N

0°

5°E

Figure 1 Map of the North Sea.


10°E


Depth, m 0_30 30_40 40_50 50_100 100_200 200_300 300_400 400_500 500_600 600_700 700_800 800_900 900_1000 1000_1100 1100_1200 1200_1300 1300_1400 Below 1400

Figure 2 Bathymetry of the North Sea. (Reproduced with permission from North Sea Task Force, 1993.)


75

76


30 000 m3 s 1 in a surface layer partially compensated by a flow of 15 000 m3 s 1 of higher salinity (17.4 PSU) North Sea water into the Baltic in a nearbed layer. However, the instantaneous exchange between the Kattegat and the Baltic Sea is by no means a steady two-layer flow, but is predominantly barotropic (in response to wind forcing and the resulting sea level slopes) and can be up to 300 000 m3 s 1 in either direction. The fresher water flows into the Norwegian Coastal Current, a surface current following the Norwegian coast, and exits the North Sea. Generally there is little exchange with water in the rest of the North Sea, although on occasion a surface layer of low salinity water can spread westward across much of the northern North Sea. The boundary between the Norwegian Coastal Current and Atlantic water in the northern North Sea (see Figure 3) is not smooth but is subject to meanders with a wavelength of 50–100 km which can break off into gyres and vortex pairs generally propagating northward. Orbital speeds within the gyres are asymmetric; up to 2 m s 1 has been reported in the upper layer.

Stratification The behavior of currents depends on whether the water column is well mixed or stratified. For some processes, for instance the response to wind forcing, stratification leads to the surface layer becoming decoupled from the bed layer – the sharper the transition from surface to bed layers (called the thermocline, for temperature) the greater the degree of decoupling. Well-mixed and stratified regions are separated by sharp fronts. Stratification arises when mixing is insufficient to mix down lighter water at the sea surface – the water might be lighter either because of solar heat input during summer or because of freshwater river discharge. The energy available for mixing is primarily derived from the tides and its effect on the water column depends on the water depth. Hence the southern North Sea, where depths are shallow and tidal currents strong, tends to remain well mixed throughout the year, apart from regions close to the coast affected by river plumes. Solar heat input causes most of the northern North Sea to stratify between April/May and October/ December, with a well-mixed surface layer of about 30–40 m deep. In autumn heat loss at the surface leads to the surface mixed layer deepening and cooling until the bottom is reached. Stratification caused by river discharge can occur at any time of the year, the fresher water tending to

form a thin surface layer in a plume about 30 km wide which stays close to the coast. The nature of the plume is very variable, depending on freshwater discharge, wind, and tide (both during the semi–diurnal and the spring-neap cycles). Mean currents in the plume can be up to 0.2 m s 1.

Measurements (see Drifters and Floats; Moorings; Nuclear Fuel Reprocessing and Related Discharges; Profiling Current Meters; Single Point Current Meters) The long-term circulation (Figure 3) has been investigated since before 1900 by studying the returns of surface and seabed drifters and the distribution of tracers, initially salinity. More recently dissolved radionuclides have been used as tracers, primarily cesium-137 and technetium-99 discharged into the sea at very low levels in waste from nuclear fuel reprocessing plants, mainly at Sellafield on the west coast of the UK and Cap de la Hague on the north coast of France. These have given unambiguous confirmation of circulation paths. Since the 1960s variations in currents on short time scales have been measured at fixed points by deploying moorings, often for about a month, with current meters, fitted with rotors or propellers, recording data every few minutes. In the 1990s a new instrument, acoustic Doppler current profiler (ADCP), became available which can measure the current profile throughout the majority of the water column in continental shelf seas. Frequently deployed in a seabed frame the current is determined from the Doppler shift in the back-scattered signal transmitted at 201 or 301 to the vertical. Currents at different depths are determined by chopping the return signal into segments. Another new technique is shore-based hf radar, where again the Doppler shift in a back-scattered signal is used to determine surface currents over a region, typically in 1 km cells, out to a range of about 30 km.

Numerical Models At any one time measurements can give only an estimate of currents with very limited spatial coverage compared with the extent of the North Sea. However, these can be complemented by numerical models to give fuller spatial coverage. The basic hydrodynamic equations are well known and have been solved on regular finite difference or irregular finite element grids. Currents vary in two horizontal directions and in the vertical, but the models are particularly quick if the equations are depth-averaged,


78


which is especially useful for the estimation of tide and storm surge elevations. Such models are now run operationally in association with meteorological models (the only sufficiently detailed source of meteorological forcing data to drive realistic sea models). Three-dimensional models are being developed which predict the tidal and wind-driven current profile and the effects of horizontal and vertical density gradients. Given the difficulty and expense of making accurate current measurements the future for prediction is via numerical models tested against a few critical measurements. Numerical models are also an integral component of process studies, since the significance of terms in the equations can be isolated and determined.

Tides The dominant motion for most of the shelf seas around the UK is the semi-diurnal tides in response to the gravitational attraction of the moon and the sun. The tides are regular, repetitive, and predictable. Tidal currents control many physical and nonphysical processes and determine mixing, even though they are predominantly cyclical (to and fro or describing an ellipse) and tend to have small net movement, typically 0.01–0.03 m s 1, except near irregularities in the coastline such as headlands. In addition to the familiar basic M2 12.4 hour cycle, the tides experience fortnightly spring–neap (on average spring tidal currents are twice as big as neap tidal currents), monthly, and 6 monthly cycles (larger spring tides usually occur near the spring and autumn equinoxes). Away from coasts the tidal currents can be accurately approximated (better than 0.1 m s 1) by just four constituents, or frequencies – M2, S2, N2, and K2, with periods between 11.97 and 12.66 hours. Near the shelf edge there can be localized amplification of the diurnal constituents O1 and K1. Only close to the coasts, in shallow water and near headlands, and in estuaries are distortions to the tides (represented by higher harmonics) significant. Clearly this is a very important region for predicting sea levels for navigation, but represents a small proportion of the area of the North Sea. The tides enter the North Sea from the Atlantic Ocean north of Scotland and sweep round it in anticlockwise sense as a progressive (Kelvin) wave gradually losing energy by working against bottom friction. By the time Norway is reached tidal currents are relatively weak, o 0.1 m s 1 at spring tides. Maximum currents occur in localized coastal constrictions, such as the Pentland Firth (in excess of 5 m s 1 at spring tides) and the Dover Strait (2 m s 1). Elsewhere currents at spring tides rarely exceed 1 m s 1 (Figure 4).

The amplitude of tidal currents does not vary significantly with depth, except in a boundary layer near to the bed generated by friction, of the order of a few meters (up to 10 m) thick. Here the current profile at maximum flow is approximately logarithmic in distance from the bottom. Also due to the effects of bottom friction tidal currents near the bed are in advance of those near the surface, so that the tide turns earlier, by up to half an hour, near the bed compared with near the surface.

Wind Forcing Wind forcing is responsible for the second largest currents after tides over most of the region, and in areas where tidal currents are weak (for instance off Norway), the largest. By the very nature of storms the forcing is intermittent and seasonal. The largest storms tend to occur in winter, when the water column is homogeneous. Two effects are important. Firstly, sea level gradients are set up against a coast counterbalancing the wind stress, since the North Sea is semi-enclosed and so has no long approximately straight coasts, in contrast with many other continental shelf seas. Secondly there is the direct action of the wind stress at the sea surface. The sea level gradients lead to storm surges at the coast, with the threat of coastal flooding. Sizeable currents can be generated during the setting up and relaxing of the sea level gradients. Measurement of these currents at depth is scarce not only because the environment is harsh during storms, although less severe than at the surface, but also because long deployments are required – at least over a 6 month period including winter to ensure an adequate sampling of storms. Seabed-mounted ADCPs are helpful, both because the storm environment is less likely to impinge on data quality and also because longer measurements are possible. The only feasible way of predicting extreme currents, for instance the ‘one-in50-year’ current for the design of offshore structures, is via numerical models, either run for long periods or forced by a series of extreme events. The accuracy of extreme sea level estimates is better than that for extreme currents, because models can be validated against long records (many years) of coastal elevations from tide gauges. The direct current response at the sea surface to the action of the wind stress is very difficult to measure, since waves tend to be large when winds are strong, but it appears to be limited to at most the top 25 m of the water column. (A very useful remote sensing technique here is hf radar which measures surface currents.) The response is rapid, less than a



60°N

55°N

5°N

°W

0°

5°E

Figure 4 Estimate of the maximum depth-averaged currents for an average spring tide, in m s 1.


10°E

79

80


few hours, and dependent on the roughness of the sea surface (i.e. on the waves and whitecapping). The surface current amplitude used, for instance in oil spill studies, is usually taken to be in the range 1–6% of the wind speed, with a typical value of 3%. The higher ratio appears to be more appropriate for developing seas. The direction of the wind drift is a few degrees to the right of the wind. Since there are very few accurate measurements to test models critically, the accuracy of models in this region is uncertain.

Circulation Circulation is the long-term movement of water. Within the North Sea it is forced by the net tides (for instance, responsible for 50% of water transport in the western North Sea), the mean winds, and density gradients, principally caused by freshwater input from rivers and the Baltic. All these forces tend to act in the same sense, described below. There are seasonal variations, as storms mainly occur in winter, river discharge has an annual cycle, and in summer the water column stratifies in some regions. Over short periods the net movement of water is likely to be determined by the last storm or storms. The circulation pattern outlined below is enhanced by winds from the south west and can be reversed by strong winds from south and east. Although of considerable and long-standing interest to oceanographers – to physicists, fisheries biologists, and to those studying the movement of suspended sediments and pollution – it is difficult to measure and quantify since in most places its magnitude is o 0.1 m s 1, much smaller than tidal or extreme wind-driven currents. It frequently has very short space scales, changing significantly in magnitude over a few kilometers and changing direction by 1801 from the surface to the seabed. There is, however, an overall long-term mean pattern (Figure 3), shown up clearly in the distributions of tracers whose major features in terms of direction have been known since the early 1900s. There are large uncertainties on estimates of amplitudes. The flow is broadly anticlockwise round the coasts of the North Sea, with weak and varied circulation in its center. The mean coastal flow is southward past Scotland and England into the Southern Bight, where there are inputs of salty water through the Dover Straits and of fresh water down the main rivers, all of which pass through large industrialized regions (Humber, Thames, Meuse, Scheldt, Rhine, Weser, Elbe) and on into the German Bight, flowing northward past Denmark in the Jutland current to join the Norwegian Coastal Current

in the Skagerrak. The average input through the Dover Strait is about 0.1 106 m3 s 1, but the flow there is strongly influenced by the wind and can reverse, moving south-westward out of the North Sea. There is also some flow from the coast of East Anglia to the north Dutch coast. There are major inflows (in excess of 106 m3 s 1) of water of Atlantic origin across the northern boundary, but very little penetrates far into the North Sea. The larger portion flows in on the western side of the Norwegian Trench and recirculates in the Skagerrak, flowing out along the eastern side of the Norwegian Trench underneath the Norwegian Coastal Current. A smaller inflow of mixed Atlantic and shelf water flows in on either side of the Shetland Islands and turns eastward to follow the 100 m contour across the North Sea, eventually to flow out along the eastern side of the Norwegian trench. The majority of water passing through the North Sea goes through the Skagerrak. The only major outflow is along the eastern side of the Norwegian Trench, with magnitude of about (1.3–1.8) 106 m3 s 1.

Effects of Stratification Stratification profoundly affects the nature of currents in two ways. Firstly currents are driven by the density gradients associated with fronts – the sharp boundaries between stratified and well-mixed water. The currents tend to be along the fronts with transverse exchanges inhibited. Secondly, the structure of motion in the water column is affected by the region of vertical density change – either because the surface and bed layers are decoupled or because internal waves can propagate along the density interface. Theoretically there is no difference between stratification caused by heat input and that caused by fresh water; dynamically all that matters is the density difference. In practice the consequences of thermal stratification in summer can be easier to study because the extent of the surface mixed layer and of the water depth are more manageable and because regions of thermal stratification are predictable and dependable. Fronts (see Shelf Sea and Shelf Slope Fronts)

The main front separating the thermally stratified water in summer to the north from the well-mixed water to the south starts from Flamborough Head, bifurcates around the Dogger Bank and passes to the north of the Frisian Islands. Tidal current speeds and water depths determine the front’s position. Currents along the front in a west to east sense are typically about 0.05 m s 1 but can be up to 0.15 m s 1.



81

Inertial Currents

Shelf Edge

Inertial currents are generated by pulses of wind. The currents rotate clockwise in the Northern Hemisphere with a period in hours of 12/sine (latitude), which, in the North sea, is from 13.6 h at 621N to 14.6 h at 551N. In stratified water their vertical structure is primarily the first baroclinic mode; the currents are 1801 out of phase above and below the thermocline and integrate to zero over the water column. Because of the phase reversal large shears can be generated across the thermocline – the largest in the water column. Their decay time, if not reinvigorated by another wind pulse, is of the order of days. ADCPs covering the majority of the water column are ideal instruments for measuring inertial currents, which can show all the characteristics indicated above. In the North Sea inertial current speeds up to 0.25 m s 1 have been observed in the surface well-mixed layer, which for most of the period of stratification is the shallower and hence the layer where inertial currents are stronger. During stratification some energy at or near inertial periods is present most of the time, but not always with a simple first baroclinic mode structure. The situation is different from the open ocean – in shelf seas the water depth is limited and tidal friction/mixing is ever present.

The main distinguishing feature of the shelf edge is the pronounced change in water depth over a short distance, from shelf depths (o 200 m) to oceanic (usually several thousand meters). At the shelf break between the North Sea and the Faroe–Shetland Channel water depths deepen from 200 m to 1000– 1500 m over a distance of tens of kilometers. Compared with other shelf edges, even around the UK, this is relatively gentle and the gradients are relatively smooth. A mixture of dynamical processes can exist at the shelf edge involving a combination of stratification and steep bottom slope. Three relevant to the North Sea are internal tides (see above), the slope current flowing north-westward, and coastal trapped waves with super-inertial periods which can, for instance, lead to enhanced diurnal tides (see above).

Internal waves (see Internal Waves)

Internal waves are ubiquitous in stratified water. They propagate along the thermocline density interface as a cyclical vertical movement of the density interface, which can be quite large, over 10 m, with only a 0 (1 cm) movement of the sea surface. The restoring force is primarily the buoyancy force. Internal interfacial waves are analogous to surface waves but their phase speed is less. They can have any period between the local inertial (see above) and the Brunt-Vaïssa¨la¨ or buoyancy period, which depends on the degree of stratification and can be as short as a few minutes for sharp stratification. The corresponding currents are out of phase above and below the density interface and, for small amplitude waves, sum to zero in the vertical. It is only possible to give generalities since each situation is unique. Because vertical accelerations are significant, nonhydrostatic numerical models are required to reproduce the dynamics. These are rare for shelf-wide models. Internal tides are often generated as the tides cross sharp changes in bathymetry, such as the shelf break. The internal tide then propagates both into the shelf sea and into the ocean, usually in the form of a pulse of waves with periods of order 30 minutes, repeated with a tidal periodicity.

Conclusions The North Sea is a typical semi-enclosed continental shelf sea, with a wide range of tidal conditions, winds, circulation, and effects of stratification. To interpret and understand the dynamics it is not sufficient just to study the currents but also seabed pressures/coastal elevations, the (horizontal and vertical) density field which is not static, meteorological forcing (wind, atmospheric pressure, heat input at sea surface), and river discharges.

See also Drifters and Floats. Moorings. Nuclear Fuel Reprocessing and Related Discharges. Regional and Shelf Sea Models. Shelf Sea and Shelf Slope Fronts. Single Point Current Meters. Storm Surges. Tides.

Further Reading Charnock H, Dyer KR, and Huthnance JM (eds.) (1994) Understanding the North Sea System. London: Chapman and Hall. Eisma D (1987) The North Sea: an overview. Philosophical Transactions of the Royal Society B 316: 461--485. Lee AJ (1980) North Seap: hysical oceanography. Elsevier Oceanography Series 24B: 467--493. North Sea Task Force (1993) North Sea Quality Status Report 1993. Oslo and Paris Commissions, London. Fredensborg, Denmark: Olsen Olsen. Otto L, Zimmerman JTF, and Furnes GK (1990) Review of the physical oceanography of the North Sea. Netherlands Journal of Sea Research 26: 161--238. Rodhe J (1998) The Baltic and North Seasa: processoriented review of the physical oceanography. In: Robinson AR and Brink KH (eds.) The Sea, 11, pp. 699--732. Chichester: John Wiley Sons.


NUCLEAR FUEL REPROCESSING AND RELATED DISCHARGES H. N. Edmonds, University of Texas at Austin, Port Aransas, TX, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1921–1928, & 2001, Elsevier Ltd.

Introduction Oceanographers have for some decades made use of human perturbations to the environment as tracers of ocean circulation and of the behavior of similar substances in the oceans. The study of the releases of anthropogenic radionuclides – by-products of the nuclear industry – by nuclear fuel reprocessing plants is an example of such an application. Other examples of the use of anthropogenic substances as oceanographic tracers addressed in this Encyclopedia include tritium and radiocarbon, largely resulting from atmospheric nuclear weapons testing in the 1950s and 1960s, and chlorofluorocarbons (CFCs), a family of gases which have been used in a variety of applications such as refrigeration and polymer manufacture since the 1920s. As outlined in this article, the source function of nuclear fuel reprocessing discharges, which is very different from that of either weapons test fallout or CFCs, is both a blessing and a curse, defining the unique applicability of these tracers but also requiring much additional and careful quantification. The two nuclear fuel reprocessing plants of most interest to oceanographers are located at Sellafield, in the UK, which has been discharging wastes into the Irish Sea since 1952, and at Cap de la Hague, in north-western France, which has been discharging wastes into the English Channel since 1966. These releases have been well documented and monitored in most cases. It should be noted, however, that the releases of some isotopes of interest to oceanographers have not been as well monitored throughout the history of the plants because they were not of radiological concern, were difficult to measure, or both. This is particularly true for 99Tc and 129I, which are the isotopes currently of greatest interest to the oceanographic community but for which specific, official release data are only available for the last two decades or less. The radionuclides most widely studied by ocean scientists have been 137 Cs, 134Cs, 90Sr, 125Sb, 99Tc, 129I, and Pu isotopes.

82

The Sellafield plant has dominated the releases of most of these, with the exception of 125Sb and, particularly in the 1990s, 129I. In general, the Sellafield releases have been particularly well documented and are the easiest for interested scientists and other parties to obtain. The figures of releases presented in this article are based primarily on the Sellafield data, except where sufficient data are available for Cap de la Hague. The location of the Sellafield and Cap de la Hague plants is important to the oceanographic application of their releases. As discussed below, the releases enter the surface circulation of the high latitude North Atlantic and are transported northwards into the Norwegian and Greenland Seas and the Arctic Ocean. They have been very useful as tracers of ventilation and deep water formation in these regions, which are of great importance to global thermohaline circulation and climate. In addition, the study of the dispersion of radionuclides from the reprocessing plants at Sellafield and Cap de la Hague serves as an analog for understanding the ultimate distribution and fate of other wastes arising from industrial activities in north-western Europe. Much of the literature on reprocessing releases in the oceans has focused on, or been driven by, concerns relating to contamination and radiological effects, and they have not found the same broad or general oceanographic application as some other anthropogenic tracers despite their great utility and the excellent quality of published work. This may be attributable to a variety of factors, including perhaps (1) the contaminant-based focus of much of the work and its publication outside the mainstream oceanographic literature, (2) the complications of the source function as compared with some other tracers, and (3) the comparative difficulty of measuring many of these tracers and the large volumes of water that have typically been required. This article addresses the historic and future oceanographic applications of these releases. A Note on Units

In most of the literature concerning nuclear fuel reprocessing releases, amounts (both releases and ensuing environmental concentrations) are expressed in activity units. The SI unit for activity is the becquerel (Bq), which replaced the previously common curie


NUCLEAR FUEL REPROCESSING AND RELATED DISCHARGES

(Ci), a unit which was based on the activity of one gram of radium. These two units are related as follows: 1Bq ¼ 1disintegration per second 1Ci ¼ 3:7 1010 Bq The conversion from activity to mass or molar units is dependent on the half-life of the isotope in question. Activity (A) is the product of the number of atoms of the isotope present (N) and its decay constant (l): A ¼ lN. The decay constant, l, is related to the half-life by t1=2 ¼ ðln2Þ=l. Using this equation, it is simple to convert from Bq to mol, given the decay constants in units of s1, and Avogadro’s number of 6.023 1023 atoms mol1. It is also worth noting that the total activity of naturally occurring radionuclides in sea water is approximately 12 kBq m3. Most of this activity is from the long-lived naturally occuring isotope, 40 K(t1/2 ¼ 1.25 109 years). Activities contributed from anthropogenic radionuclides greatly exceed this (millions of Bq m3) in the immediate vicinity of the Sellafield and La Hague outfall pipes. However, most oceanographic studies of these releases involve measurements of much lower activities: up to 10s of Bq m3 in the case of 137Cs, and generally even lower for other radionuclides, for instance mBq m3 or less for plutonium isotopes and 129I.

Nuclear Fuel Reprocessing and Resulting Tracer Releases Origin and Description of Reprocessing Tracers

Nuclear fuel reprocessing involves the recovery of fissile material (plutonium and enriched uranium) and the separation of waste products from ‘spent’ (used) fuel rods from nuclear reactors. In the process, fuel rods, which have been stored for a time to allow short-lived radionuclides to decay, are dissolved and the resulting solution is chemically purified and separated into wastes of different composition and activity. Routine releases from the plants to the environment occur under controlled conditions and are limited to discharge totals dictated by the overseeing authorities of each country. The discharges have varied over the years as a function of the amount and type of fuel processed and changes in the reprocessing technology. The discharge limits themselves have changed, in response to monitoring efforts and also spurred by technological advances in waste-treatment capabilities. As a general rule, these changes have resulted in decreasing releases for most nuclides (most notably cesium and the actinide elements), but

83

there are exceptions. For instance, the Enhanced Actinide Removal Plant (EARP) was constructed at the Sellafield site in the 1990s in order to enable the additional treatment and subsequent discharge of a backlog of previously stored wastes. Although the new technology enabled the removal of actinide elements from these wastes, it is not effective at removing 99Tc. Therefore an allowance was made for increased discharge of 99Tc, up to 200 TBq per year. The resulting pulse of increased 99Tc discharges from Sellafield beginning in 1994 is currently being followed with great interest, as is discussed in more detail below. The end result of these processes is the availability of a suite of oceanographic tracers with different discharge histories (e.g., in terms of the timing and magnitude of spikes) and a range of half-lives, and thus with a range of utility and applicability to studies of oceanographic processes at a variety of spatial and temporal scales. A brief summary of the reprocessing radionuclides most widely applied in oceanography is shown in Table 1, and examples of discharge histories are given in Figure 1. It has also been particularly useful in some cases to measure the ratio of a pair of tracers, for instance 134Cs/137Cs, 137 Cs/90Sr, or potentially, 99Tc/129I. The use of isotope ratios can (1) provide temporal information and aid the estimation of rates of circulation, (2) mitigate the effects of mixing which will often alter individual concentrations more than ratios, and (3) aid in distinguishing the relative contributions of different sources of the nuclides in question. Finally, differences in the chemical behavior of the different elements released can be exploited to study different processes. Although most of the widely applied tracers are largely conservative in sea water and therefore serve as tracers of water movement, the actinides, particularly Pu, have a high affinity for particulate material and accumulate in the sediments. These tracers are then useful for studying sedimentary processes. The Reprocessing Tracer Source Function

The primary difference between the north-western European reprocessing releases and other anthropogenic tracers used in oceanography is the nature of their introduction to the oceans. Reprocessing releases enter the oceans essentially at a point source, rather than in a more globally distributed fashion as is the case for weapons test fallout or the chlorofluorocarbons. Thus, reprocessing tracers are excellent, specific tracers for waters originating from north-western Europe. Because these waters are transported to the north into the Nordic Seas and


84


Summary of the major reprocessing tracers and their applications

Table 1 Isotope

Half-life (years)

Sources

Applications

137

30

Weapons testing, reprocessing (mostly Sellafield), Chernobyl

90

28

Weapons testing, reprocessing (mostly Sellafield)

134

2.06

Reprocessing and Chernobyl

125

2.7

Reprocessing, mostly Cap de la Hague

99

213 000

Reproducing – large pulse from Sellafield beginning in 1994. Some weapons testing

129

15.7 106

Reprocessing, mostly Cap de la Hague since 1990, some weapons testing

241

14 24 000 6000 88

Weapons testing, reprocessing (mostly Sellafield)

The ‘signature’ reprocessing tracer, used in the earliest studies of Sellafield releases. Has been applied in European coastal waters, the Nordic Seas, Arctic Ocean, and deep North Atlantic In combination with 137Cs, early tracer for Sellafield discharges. The 137Cs/90Sr ratio was used to distinguish reprocessing and weapons sources In combination with 137Cs, this short-lived isotope has been used in the estimation of transit times, e.g., from Sellafield to the northern exit of the Irish Sea Circulation in the English Channel and North Sea, into the Skaggerak and Norwegian coastal waters Surface circulation of European coastal waters, the Nordic Seas, Arctic Ocean, and East and West Greenland Currents Deep water circulation in the Nordic Seas, Arctic Ocean, and Deep Western Boundary Current of the Atlantic Ocean. It has also been measured in European coastal waters and the Gulf of Mexico Not widely used as circulation tracers but often measured in conjunction with other reprocessing radionuclides. Other transuranic elements that have been measured include 241Am and 237Np

Cs

Sr

Cs

Sb

Tc

I

Pu Pu 240 Pu 238 Pu 239

1200

5000

1000

4000

800 3000

600

2000

400 99

1000

Tc

200 0

0 1950

1960

1970

1980

1990

2000

129

I

_ Tc (TBq year 1 ),

_ Cs (TBq year 1)

Cs

137

1400

129

137

99

6000

_ I (TGq year 1 )

Adapted in part from Dahlgaard (1995).

Year Figure 1 Examples of the source functions of reprocessing tracers to the oceans, illustrating some of the differences in magnitude and timing of the releases. 137Cs data, from Gray et al. (1995) and 99Tc data, courtesy of Peter Kershaw, are for liquid discharges from Sellafield. 129I release information is courtesy of G. Raisbeck and F. Yiou, and combines available information for Sellafield from 1966 and Cap de la Hague from 1975. Note the different scales used for the releases of the three isotopes.

Arctic Ocean, reprocessing releases are particularly sensitive tracers of the climatically important deep water formation processes that occur in these regions.

There are several complications associated with this point source tracer introduction, however. Comparison with CFCs indicates some of these difficulties. Within each hemisphere (northern or southern), CFCs are well mixed throughout the troposphere, and the time history of their concentrations is well known. Their entry into the oceans occurs by equilibration with surface waters, and the details of their solubilities as a function of temperature and salinity have been well characterized. The primary complication is that in some areas equilibrium saturation is not reached, and so assumptions must be made about the degree of equilibration. Where they have been necessary, these assumptions appear to be fairly robust. In the case of reprocessing tracers, the discharge amounts have not always been as well known, although this situation has improved continuously. One major difficulty arises in translating a discharge amount, in kg or Bq per year, or per month, to a concentration in sea water some time later. In order to do this, the surface circulation of the coastal regions must be very well known. This circulation is highly variable, on daily, seasonal, interannual and decadal timescales, further



complicating the problem. Also, other sources of the radionuclides under consideration must often be taken into account. Depending on the tracer, these sources may be important and/or numerous, and due to the secrecy involved in many aspects of the operation of nuclear installations the necessary information may not always be available. Other Sources of Anthropogenic Radionuclides

Other sources of radionuclides to the oceans complicate the interpretation of Sellafield and Cap de la Hague tracers to varying extents, depending on the isotope under consideration, the location, and the time. First there is the need to understand the mixing of signals from these two plants. Other sources may include fallout from nuclear weapons tests, uncontrolled releases due to nuclear accidents, dumping on the seabed, other reprocessing plants, atmospheric releases from Sellafield and Cap de la Hague, and unknown sources. Many of these other sources are small compared to the Sellafield and Cap de la Hague releases. The primary complicating source for many isotopes is nuclear weapons test fallout, which peaked in the 1950s and again, more strongly, in 1962–63. This source has been particularly important for 137Cs and 90Sr. The Chernobyl accident in 1986 also released significant amounts of radioactive materials which have themselves found use as oceanographic tracers. In terms of comparison to releases from Sellafield, the Chernobyl accident has been most important with respect to 134Cs and 137Cs. In addition, it has recently been noted that in the case of some nuclides, particularly 137Cs and Pu isotopes, the sediments of the Irish Sea have become a significant ongoing source of tracers to the North Atlantic. Although generally considered a conservative tracer, 137Cs exhibits some affinity for particulate material and a significant amount has accumulated in the sediments around Sellafield. With the continuing reduction of 137Cs activities in liquid effluents from Sellafield, release from the sediments, either by resuspension of the sediments or by reequilibration with the reduced seawater concentrations, has become a relatively large (though still small in an absolute sense compared to the liquid discharges of the 1970s) contributor to the current flux out of the Irish Sea, and will continue to be such for years to come. A further complication of the use of some tracers derived from Sellafield and Cap de la Hague, but one that cannot be lamented, is the continuing reduction of the releases, as well as the radioactive decay of those with the shorter half-lives, such as 137Cs which was released in large quantities over 20 years ago. With respect to many of these complications, and for

85

several other reasons, the reprocessing radionuclides attracting the most attention in the oceanographic community today are 99Tc and 129I. Both are long lived and have fairly small relative contributions from weapons testing and other sources. Unlike most radionuclides, their releases increased in the 1990s, and the recent releases of each are largely dominated by a single source: 99Tc by Sellafield and 129I by Cap de la Hague. Significant advances have been and are being made in the measurement of these isotopes, allowing their measurement on smaller sample volumes. The activity of 129I is so low that it is measured by accelerator mass spectrometry. This technique allows measurement of 129I on 1 liter seawater samples. Advances in 99Tc measurement, using both radiochemical and mass-spectrometric (ICP-MS) techniques, are continuing. The primary limitation of 99 Tc studies continues to be the comparative difficulty of its measurement. This is also true to some extent for 129I, for although the sample sizes have been greatly reduced the measurement requires highly specialized technology. A further complication for 129I is its volatility, and the fact that as much as 10% of the reprocessing discharges have been released directly to the atmosphere. Nevertheless, the promise of both tracers is such that these difficulties are likely to be overcome. Regional Setting and Circulation of Reprocessing Discharges

A summary of the regional circulation into which the liquid effluents from Sellafield and Cap de la Hague are released is presented in Figure 2. It is particularly important to note that studies of the reprocessing discharges have contributed greatly to the development of this detailed picture of the regional circulation. Briefly, from the Sellafield site the waste stream is carried north out of the Irish Sea, around the coast of Scotland, through the North Sea, and into the northward-flowing Norwegian Coastal Current (NCC). Transport across the North Sea occurs at various latitudes: some fraction of the reprocessing releases ‘short-circuits’ across the northern part of the North Sea, while some flows farther south along the eastern coast of the UK before turning east and north. Recent studies of the EARP 99Tc pulse from Sellafield have suggested that the rate and preferred transport path of Sellafield releases across the North Sea into the NCC may vary in relation to climatic conditions in the North Atlantic such as the North Atlantic Oscillation (NAO). Radionuclides discharged from the Cap de la Hague reprocessing plant flow north-east through the English Channel and into the North Sea,


86


_ 300

30

80

60

_ 60

60

Arctic Ocean

80 Labrador Sea

Greenland Sea

WSC

surface currents area of deep convection eastern overflow Denmark Strait Overflow Water _ DWBC Labrador Sea Water

EGC DSOW

_ 60

LSW Norwegian Sea

NEADW ISOW

NCC

FBC

60

40 North Sea

Sellafield

km 0

500

1000

Cap de la Hague

40

_ 30

0

Figure 2 Map of the northern North Atlantic and Nordic Seas, indicating the locations of the Sellafield and Cap de la Hague reprocessing plants (stars), and the major circulation pathways relevant to the discussion of the dispersal of reprocessing wastes. Surface currents are indicated with thin solid lines, and deep waters, (deep overflows from the Nordic Seas, the resulting Deep Western Boundary Current (DWBC) and Labrador Sea Water (LSW)), with heavy dashed lines. Curled arrows indicate the two major areas where deep waters are formed by convective processes. Additional abbreviations: NCC, Norwegian Coastal Current; WSC, West Spitsbergen Current; EGC, East Greenland Current; FBC, Faroe Banks Channel; ISOW, Iceland Scotland Overflow Water; NEADW, Northeast Atlantic Deep Water; DSOW, Denmark Strait Overflow Water.

following the coast and joining the Sellafield releases in the NCC. A small amount of the Sellafield releases and some of the Cap de la Hague releases, which flow closer to the coast, flow east through the Skaggerak and Kattegat to enter the Baltic Sea. The NCC is formed of a mixture of coastal waters (containing the reprocessing tracers) and warm, saline North Atlantic surface waters. Dilution of the reprocessing signal with Atlantic water continues along the northward flow path of the NCC. There is evidence from reprocessing tracers that turbulent eddies between the NCC and Atlantic water result in episodic transport westward into the surface waters of the Norwegian Sea, in addition to a fairly welldefined westward advective transport towards Jan Mayen Island. The NCC branches north of Norway, with one branch, the North Cape (or Nordkap) Current, flowing eastward through the Barents Sea and thence into the Kara Sea and Arctic Ocean, and the remainder flowing north and west as part of the West Spitsbergen Current (WSC). This latter flow branches in the Fram Strait west of Spitsbergen, with some recirculation to the west and south joining the

southward flowing East Greenland Current (EGC), and the remainder entering the Arctic Ocean. The majority of surface outflow from the Arctic is through the Fram Strait into the EGC, thus the bulk of the reprocessing nuclides entering the Arctic Ocean will eventually exit to the Nordic (Greenland, Iceland, and Norwegian) Seas and the North Atlantic. The presence of high concentrations of radionuclides derived from reprocessing in the surface waters of the Barents and Nordic Seas is also an indication of their utility as tracers of deep-water ventilation and formation in these regions, as discussed below.

The Use of Reprocessing Releases in Oceanography Historical Background

The first papers reporting measurements of Sellafield-derived 137Cs in coastal waters appeared in the early 1970s. Much of this early work arose from monitoring efforts by the Division of Fisheries



Research of the UK’s Ministry of Agriculture, Fisheries, and Food, which holds a joint oversight role on the Sellafield discharges and which continues to this day, now as the Centre for Environment, Fisheries and Aquaculture Science, to be a leader in studies of the distribution and oceanographic application of reprocessing radionuclides. In the late 1970s scientists at the Woods Hole Oceanographic Institution published the first of a number of papers detailing the unique utility of the reprocessing tracers and demonstrating their application. Leading studies of reprocessing tracers in the oceans have been undertaken by researchers in numerous other countries affected by the radiological implications of the releases, including France, Germany, Denmark, Canada, Norway, and the former Soviet Union. Coastal and Surface Circulation

A great deal of work has been published using the documented releases of radioisotopes from Sellafield and Cap de la Hague to study the local circulations of the Irish Sea, North Sea, and English Channel. Early studies of the Sellafield releases examined a variety of isotopes, including 134Cs, 137Cs, 90Sr, and Pu isotopes. Most attention focused on 137Cs and 90Sr, and their activity ratio, because: (1) the releases of these two isotopes were well documented, (2) they had been studied extensively since the 1950s and 1960s in weapons test fallout; and (3) the 137Cs/90Sr ratio in reprocessing releases (particularly those from Sellafield) was significantly higher than in global fallout and thus could be used to distinguish the sources of these isotopes in a given water sample. The use of the 134 Cs/137Cs ratio enabled estimates of transit times, assuming the initial ratio in the releases was constant and making use of the short half-life of 134Cs. The short-lived isotope 125Sb has been used as a specific tracer of the circulation of Cap de la Hague discharges through the English Channel and North Sea, into the Baltic and the Norwegian Coastal Current. Summaries of transit times and dilution factors for the transport of Sellafield and Cap de la Hague discharges to points throughout the North Sea, Norwegian Coastal Current, Barents and Kara Seas, Greenland Sea, and East and West Greenland Currents have been published in recent reviews. Numbers are not included in this article because they are currently under revision. In terms of transport to the NCC, where the two waste streams are generally considered to merge, the consensus has been that the transit time from Sellafield to about 601N is three to four years, and that from Cap de la Hague to the same area is one to two years. Compilations of ‘transfer factors,’ which relate observed concentrations to the discharge

87

amounts, and factor in the transit times, suggest that the two reprocessing waste streams meet in approximately equal proportions in the NCC. In other words, if Sellafield and Cap de la Hague released equal amounts of a radionuclide, with the Cap de la Hague release two years later, they would make equal contributions to the NCC. Recently, detailed studies have been undertaken of the dispersal of the EARP 99Tc pulse from Sellafield, which began in 1994. These studies have suggested substantially shorter circulation times for Sellafield releases to northern Scottish coastal waters and across the North Sea to the NCC than previously accepted. For instance, the 99Tc pulse reached the NCC within 2.5 years, rather than three to four. It has been suggested that this is a real difference between sampling periods, resulting from climatically induced circulation changes in the North Atlantic. Some of the difference may also be related to the fact that much more detailed data, both in terms of seawater sampling and regarding the releases, are available on this recent event. The continuing passage of the EARP 99Tc signal promises to be very useful in the Nordic Seas and Arctic Ocean as well. Deep-water Formation in the North Atlantic and Arctic Oceans

In addition to surface water flows, deep-water formation processes within the Arctic Ocean and Nordic Seas have been elucidated through the study of reprocessing releases. In a classic presentation of reprocessing tracer data, Livingston showed that the surface water distribution of 137Cs in the Nordic Seas in the early 1980s was marked by high concentrations at the margins of the seas, highlighting the delivery of the isotope in the northward-flowing NCC and WSC to the west and the return flow in the southward-flowing EGC to the east. The opposite distribution was found in the deep waters, with higher concentrations in the center of the Greenland Basin than at the margins, as a result of the ventilation of the Greenland Sea Deep Water by deep convective processes in the center of the gyre. In the Arctic Ocean, elevated 137Cs and 90Sr concentrations and 137Cs/90Sr ratios at 1500 m at the LOREX ice station near the North Pole in 1979 indicated that deep layers of the Arctic Ocean were ventilated from the shelves. Similar observations in deep water north of Fram Strait provided early evidence suggesting a contribution of dense brines from the Barents Sea shelf to the bottom waters of the Nansen Basin. Reprocessing tracers, particularly those, like 129I and 99Tc, which have only a small contribution from other sources such as weapons


88


fallout, hold great promise for illuminating eastern Arctic Ocean deep water ventilation processes from the Barents and Kara Sea shelves. The NCC delivers reprocessing tracers to the Barents and Kara relatively rapidly (B 5 years) and more importantly in high concentration. With high tracer concentrations in the area of interest as a source water, the reprocessing tracers may be particularly sensitive tracers of a contribution of dense shelf waters to the deep Arctic Ocean. In addition to the deep waters formed through convection in the polar regions (most notably in the Greenland Sea) which fill the deep basins of the Nordic Seas, intermediate waters are formed which subsequently overflow the sills between Greenland, Iceland, and Scotland and ventilate the deep North Atlantic. The presence of 137Cs and 90Sr from Sellafield was reported in the overflow waters immediately south of the Denmark Straits sampled during the Transient Tracers in the Ocean (TTO) program in 1981. Later, it was demonstrated that reprocessing cesium and strontium could be distinguished in the deep waters as far as TTO Station 214, off the Grand Banks of Newfoundland. No samples were taken for reprocessing radionuclides further south as part of that study, but it was clear in retrospect that the reprocessing signal had traveled even further in the Deep Western Boundary Current (DWBC). Recent work on 129I has highlighted the utility of reprocessing radionuclides as tracers of northern source water masses and the DWBC of the Atlantic. Profiles in stations south of the overflows show much clearer tracer signals for 129I than for the CFCs, and 129 I has been detected in the DWBC as far south as Cape Hatteras. As with the cesium studies of the 1980s, it is likely that sampling further south will reveal the tracer there as well.

Conclusions Releases of radionuclides from the nuclear fuel-reprocessing plants at Sellafield and Cap de la Hague have provided tracers for detailed studies of the circulations of the local environment into which they are released, namely the Irish Sea, English Channel, and North Sea, and for the larger-scale circulation processes of the North Atlantic and Arctic Oceans. These tracers have very different source functions for their introduction into the oceans compared to other widely used anthropogenic tracers, and in some cases compared to each other. The fact that they are released at point sources makes them highly specific tracers of several interesting processes in ocean

circulation, but the nature of the releases has complicated their quantitative interpretation to some extent. Recent advances have been made in the measurement of 129I and 99Tc, long-lived tracers whose releases have increased in recent years and which experience little complication from other sources. These two tracers hold great promise for elucidating deep water formation and ventilation processes in the North Atlantic, Nordic Seas, and Arctic Ocean in the years to come.

See also Arctic Ocean Circulation. CFCs in the Ocean. North Sea Circulation. Radioactive Wastes. Water Types and Water Masses.

Further Reading Aarkrog A, Dahlgaard H, Hallstadius L, Hansen H, and Hohm E (1983) Radiocaesium from Sellafield effluents in Greenland waters. Nature 304: 49--51. Dahlgaard H (1995) Transfer of European coastal pollution to the Arctic: radioactive tracers. Marine Pollution Bulletin 31: 3--7. Gray J, Jones SR, and Smith AD (1995) Discharges to the environment from the Sellafield Site, 1951–1992. Journal of Radiological Protection 15: 99--131. Jefferies DF, Preston A, and Steele AK (1973) Distribution of caesium-137 in British coastal waters. Marine Pollution Bulletin 4: 118--122. Kershaw P and Baxter A (1995) The transfer of reprocessing wastes from north-west Europe to the Arctic. Deep-Sea Research II 42: 1413--1448. Kershaw PJ, McCubbin D, and Leonard KS (1999) Continuing contamination of north Atlantic and Arctic waters by Sellafield radionuclides. Science of Total Environment 237/238: 119--132. Livingston HD (1988) The use of Cs and Sr isotopes as tracers in the Arctic Mediterranean Seas. Philosophical Transactions of the Royal Society of London A 325: 161--176. Livingston HD, Bowen VT, and Kupferman SL (1982) Radionuclides from Windscale discharges I: non-equilibrium tracer experiments in high-latitude oceanography. Journal of Marine Research 40: 253--272. Livingston HD, Bowen VT, and Kupferman SL (1982) Radionuclides from Windscale discharges II: their dispersion in Scottish and Norwegian coastal circulation. Journal of Marine Research 40: 1227--1258. ¨ stlund HG (1985) Livingston HD, Swift JH, and O Artificial radionuclide tracer supply to the Denmark Strait Overflow between 1972 and 1981. Journal of Geophysical Research 90: 6971--6982.


OCEAN BIOGEOCHEMISTRY AND ECOLOGY, MODELING OF N. Gruber, Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, Switzerland S. C. Doney, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Modeling has emerged in the last few decades as a central approach for the study of biogeochemical and ecological processes in the sea. While this development was facilitated by the fast development of computer power, the main driver is the need to analyze and synthesize the rapidly expanding observations, to formulate and test hypotheses, and to make predictions how ocean ecology and biogeochemistry respond to perturbations. The final aim, prediction, has gained in importance recently as scientists are increasingly asked by society to investigate and assess the impact of past, current, and future human actions on ocean ecology and biogeochemistry. The impact of the carbon dioxide (CO2) that humankind has emitted and will continue to emit into the atmosphere for the foreseeable future is currently, perhaps, the dominant question facing the marine biogeochemical/ecological research community. This impact is multifaceted, and includes both direct (such as ocean acidification) and indirect effects that are associated with the CO2-induced climate change. Of particular concern is the possibility that global climate change will lead to a reduced capacity of the ocean to absorb CO2 from the atmosphere, so that a larger fraction of the CO2 emitted into the atmosphere remains there, further enhancing global warming. In such a positive feedback case, the expected climate change for a given CO2 emission will be larger relative to a case without feedbacks. Marine biogeochemical/ecological models have played a crucial role in elucidating and evaluating these processes, and they are increasingly used for making quantitative predictions with direct implications for climate policy. There are many other marine biogeochemical and/ or ecological problems related to human activities, for which models play a crucial role assessing their importance and magnitude and devising possible

solutions. These include, for example, coastal eutrophication, overfishing, and dispersion of invasive species. The use of marine biogeochemical/ecological models is now so pervasive that practically every field of oceanography is on this list. The aim of this article is to provide an introduction and overview of marine biogeochemical and ecological modeling. Given the breadth of modeling approaches in use today, this overview can by design not be inclusive and authoritative. We rather focus on some basic concepts and provide a few illustrative applications. We start with a broader description of the marine biogeochemical/ecological challenge at hand, and then introduce basic concepts used for biogeochemical/ecological modeling. In the final section, we use a number of examples to illustrate some of the core modeling approaches.

The Marine Ecology and Biogeochemistry Challenge The complexity of the ocean biogeochemical/ecological problem is daunting, as it involves a complex interplay among biology, physical variability of the oceanic environment, and the interconnected cycles of a large number of bioactive elements, particularly those of carbon, nitrogen, phosphorus, oxygen, silicon, and iron (Figure 1). Furthermore, the ocean is an open system that exchanges mass and many elements with the surrounding realms, such as the atmosphere, the land, and the sediments. The engine that sets nearly all of these cycles into motion is the photosynthetic fixation of dissolved inorganic carbon and many other nutrient elements into organic matter by phytoplankton in the illuminated upper layers of the ocean (euphotic zone). The net rate of this process (i.e., net primary production) is distributed heterogeneously in the ocean, primarily as a result of the combined limitation of nutrients and light. This results in similar heterogeneity in surface chlorophyll, a direct indicator of the amount of phytoplankton biomass (Figure 2(a)). The large-scale surface nutrient distributions, in turn, reflect the balance between biological removal and the physical processes of upwelling and mixing that transport subsurface nutrient pools upward into the euphotic zone (Figure 2(b)). In addition to the traditional macronutrients (nitrate, phosphate, silicate etc.), growing evidence shows that iron limitation is


89

90

OCEAN BIOGEOCHEMISTRY AND ECOLOGY, MODELING OF

Atmosphere N2

O2

CO2 Air−sea gas exchange

N2O

Air−sea gas exchange

Ocean Euphotic zone

Calcification

O2

DIC

N2-fixation

NO3

Org. matter

CaCO3

N2

Silification 4-

SiO 4

N2O

PO4

Settling

SiO 2

Mixing/circulation Mixing/circulation

Settling

O2

DIC

Settling

Aphotic zone

Mixing/circulation

4-

NO3

Org. matter

CaCO3 SiO 4

N2 Denitrification

N2O

SiO 2

PO44 PO DIC

SiO 2

CaCO3

Org. matter

NO3

Sediments

Figure 1 Schematic diagram of a number of key biogeochemical cycles in the ocean and their coupling. Shown are the cycles of carbon, oxygen, phosphorus, nitrogen, and silicon. The main engine of most biogeochemical cycles in the ocean is the biological production of organic matter in the illuminated upper ocean (euphotic zone), part of which sinks down into the ocean’s interior and is then degraded back to inorganic constituents. This organic matter cycle involves not only carbon, but also nitrogen, phosphorus, and oxygen, causing a tight linkage between the cycles of these four elements. Since some phytoplankton, such as diatoms and coccolithophorids, produce shells made out of amorphous silicon and solid calcium carbonate, the silicon and CaCO3 cycles also tend to be closely associated with the organic matter cycle. These ocean interior cycles are also connected to the atmosphere through the exchange of a couple of important gases, such as CO2, oxygen, and nitrous oxide (N2O). In fact, on timescales longer than a few decades, the ocean is the main controlling agent for the atmospheric CO2 content.

a key factor governing net primary production in many parts of the ocean away from the main external sources for iron, such as atmospheric dust deposition and continental margin sediments. The photosynthesized organic matter is the basis for a complex food web that involves both a transfer of the organic matter toward higher trophic levels as well as a microbial loop that is responsible for most of the breakdown of this organic matter back to its inorganic constituents. A fraction of the synthesized organic matter (about 10–20%) escapes degradation in the euphotic zone and sinks down into the dark aphotic zone, where it fuels the growth of microbes and zooplankton that eventually also remineralize most of this organic matter back to its inorganic constituents. The nutrient elements and the dissolved inorganic carbon are then eventually returned back to the surface ocean by ocean circulation and mixing, closing this ‘great biogeochemical loop’. This loop is slightly leaky in that a

small fraction of the sinking organic matter fails to get remineralized in the water column and is deposited onto the sediments. Very little organic matter escapes remineralization in the sediments though, so that only a tiny fraction of the organic matter produced in the surface is permanently removed from the ocean by sediment burial. In the long-term steady state, this loss of carbon and other nutrient elements from the ocean is replaced by the input from land by rivers and through the atmosphere. Several marine phytoplankton and zooplankton groups produce hard shells consisting of either mineral calcium carbonate (CaCO3) or amorphous silica, referred to as ‘opal’ (SiO2). The most important producers of CaCO3 are coccolithophorids, while most marine opal stems from diatoms. Both groups are photosynthetic phytoplankton, highlighting the extraordinary importance of this trophic group for marine ecology and biogeochemistry. Upon the death of the mineral-forming organisms, these minerals


92


while respiration and bacterial degradation releases CO2 and consumes O2. As a result, one tends to find an inverse relationship in the oceanic distribution of these two gases. CO2 also sets itself apart from O2 in that it reacts readily with seawater. In fact, due to the high content of alkaline substances in the ocean, this reaction is nearly complete, so that only around 1% of the total dissolved inorganic carbon in the ocean exists in the form of CO2. The exchange of CO2 across the air–sea interface is a prime question facing ocean biogeochemical/ ecological research, particularly with regard to the magnitude of the oceanic sink for anthropogenic CO2. The anthropogenic CO2 sink occurs on top of the natural air–sea CO2 fluxes characterized by oceanic uptake in mid-latitudes and some high latitudes, and outgassing in the low latitudes and the Southern Ocean. This distribution of the natural CO2 flux is the result of an interaction between the exchange of heat between the ocean and the atmosphere, which affects the solubility of CO2, and the great biogeochemical loop, which causes an uptake of CO2 from the atmosphere in regions where the downward flux of organic carbon exceeds the upward supply of dissolved inorganic carbon, and an outgassing where the balance is the opposite. This flux has changed considerably over the last two centuries in response to the anthropogenically driven increase in atmospheric CO2 that pushes additional CO2 from the atmosphere into the ocean. Nevertheless, the pattern of the resulting contemporary air–sea CO2 flux (Figure 2(c)) primarily still reflects the flux pattern of the natural CO2 fluxes, albeit with a global integral flux into the ocean reflecting the oceanic uptake of anthropogenic CO2. Given the central role of the great biogeochemical loop, any modeling of marine biogeochemical/ecological processes invariably revolves around the modeling of all the processes that make up this loop. As this loop starts with the photosynthetic production of organic matter by phytoplankton, biogeochemical modeling is always tightly interwoven with the modeling of marine ecology, especially that of the lower trophic levels. Core questions that challenge marine biogeochemistry and ecology are as follows: 1. What controls the mean concentration and threedimensional (3-D) distribution of bioreactive elements in the ocean? 2. What controls the air–sea balance of climatically important gases, that is, CO2, N2O, and O2? 3. What controls ocean productivity, the downward export of organic matter, and the transfer of organic matter to higher trophic levels?

4. How do ocean biogeochemistry and ecology change in time in response to climate dynamics and human perturbations? Modeling represents a powerful approach to studying and addressing these core questions. Modeling is by no means the sole approach. In fact, integrated approaches that combine observational, experimental, and modeling approaches are often necessary to tackle this set of complex problems.

What Is a Biogeochemical/Ecological Model? At its most fundamental level, a model is an abstract description of how some aspect of nature functions, most often consisting of a set of mathematical expressions. In the biogeochemical/ecological modeling context, a model usually consists of a number of partial differential equations, which describe the time and space evolution of a (limited) number of ecological/biogeochemical state variables. As few of these equations can be solved analytically, they are often solved numerically using a computer, which requires the discretization of these equations, that is, they are converted into difference equations on a predefined spatial and temporal grid.

The Art of Biogeochemical/Ecological Modeling A model can never fully represent reality. Rather, it aims to represent an aspect of reality in the context of a particular problem. The art of modeling is to find the right level of abstraction, while keeping enough complexity to resolve the problem at hand. That is, marine modelers often follow the strategy of Occam’s razor, which states that given two competing explanations, the one that is simpler and makes fewer assumptions is the more likely to be correct. Therefore, a typical model can be used only for a limited set of applications, and great care must be used when a model is applied to a problem for which it was not designed. This is especially true for biogeochemical/ecological models, since their underlying mathematical descriptions are for the most part not based on first principles, but often derived from empirical relationships. In fact, marine biogeochemical/ecological modeling is at present a data-limited activity because we lack data to formulate and parametrize key processes and/or to evaluate the model predictions. Further, significant simplifications are often made to make the problem more tractable. For example, rather than treating



individual organisms or even species, model variables often aggregate entire functional groups into single boxes (e.g., photosynthetic organisms, grazers, and detritus decomposers), which are then simulated as bulk concentrations (e.g., mol C m 3 of phytoplankton). Despite these limitations, models allow us to ask questions about the ocean inaccessible from data or experiments alone. In particular, models help researchers quantify the interactions among multiple processes, synthesize diverse observations, test hypotheses, extrapolate across time – space scales, and predict past and future behavior. A well-posed model encapsulates our understanding of the ocean in a mathematically consistent form.

Biogeochemical/Ecological Modeling Equations and Approaches In contrast with their terrestrial counterparts, models of marine biogeochemical/ecological processes must be coupled to a physical circulation model of some sort to take into consideration that nearly all relevant biological and biogeochemical processes occur either in the dissolved or suspended phase, and thus are subject to mixing and transport by ocean currents. Thus, a typical coupled physical–biogeochemical/ ecological model consists of a set of time-dependent advection, diffusion, and reaction equations: @C þ AdvðCÞ þ DiffðCÞ ¼ SMSðCÞ @t

½1

where C is the state variable to be modeled, such as the concentration of phytoplankton, nutrients, or dissolved inorganic carbon, often in units of mass per unit volume (e.g., mol m 3). Adv(C) and Diff(C) are the contributions to the temporal change in C by advection and eddy-diffusion (mixing), respectively, derived from the physical model component. The term SMS(C) refers to the ‘sources minus sinks’ of C driven by ecological/biogeochemical processes. The SMS term often involves complex interactions among a number of state variables and is provided by the biogeochemical/ecological model component. Marine biogeochemical/ecological models are diverse, covering a wide range of complexities and applications from simple box models to globally 4-D(space and time) coupled physical–biogeochemical simulations, and from strict research tools to climate change projections with direct societal implications. Model development and usage are strongly shaped by the motivating scientific or policy problems as well as the dynamics and time–space scales considered. The

93

complexity of marine biogeochemical/ecological models can be organized along two major axes (Figure 3): the physical complexity, which determines how the left-hand side of [1] is computed, and the biogeochemical/ecological complexity, which determines how the right-hand side of [1] is evaluated. Due to computational and analytical limitations, there is often a trade-off between the physical and biogeochemical/ecological complexity, so that models of the highest physical complexity are often using relatively simple biogeochemical/ecological models and vice versa (Figure 4). Additional constraints arise from the temporal domain of the integrations. Applications of coupled physical–biogeochemical/ ecological models to paleoceanographic questions require integrations of several thousand years. This can only be achieved by reducing both the physical and the biogeochemical/ecological complexity (Figure 4). At the same time, the continuously increasing computational power has permitted researchers to push forward along both complexity axes. Nevertheless, the fundamental tradeoff between physical and biogeochemical/ecological complexity remains. In addition to the physical and biogeochemical/ ecological complexity, models can also be categorized with regard to their interaction with observations. In the case of ‘forward models’, a set of equations in the form of [1] is integrated forward in time given initial and boundary conditions. A typical forward problem is the prediction of the future state of ocean biogeochemistry and ecology for a certain evolution of the Earth’s climate. The solutions of such forward models are the time–space distribution of the state variables as well as the implied fluxes. The expression ‘inverse models’ refers to a broad palette of modeling approaches, but all of them share the goal of optimally combining observations with knowledge about the workings of a system as embodied in the model. Solutions to such inverse models can be improved estimates of the current state of the system (state estimation), improved estimates of the initial or boundary conditions, or an optimal set of parameters. A typical example of an inverse model is the optimal determination of ecological parameters, such as growth and grazing rates, given, for example, the observed distribution of phytoplankton, zooplankton, and nutrients.

Examples Given the large diversity of marine biogeochemical/ ecological models and approaches, no review can do


94


Higher complexity

Biogeochemical/ecological complexity Upper ocean ecology (lower trophic levels)

NP

N

Upper ocean ecology (higher trophic levels)

Aphotic zone processes

Implicit remineralization

Fixed stoichiometry

Other reservoirs

Ocean only

NiPj Zk Bl Dm

ZF with life cycles

ZF

Elemental coupling

NPZBD

NPZD

ZFn with life cycles and coupled to lower trophic level model

Sinking particles with parametrized remineralization

Sinking particles with aphotic ecosystem determining remineralization

Variable stoichiometry

Sediments passive

Sediments active

Physical complexity

Spatial resolution

System coupling

Physiological modeling

Higher complexity

box

ocean only

1-D

2-D

3-D

ocean/ sea-ice

3-D (eddy resolving)

ocean/ sea-ice/ atmosphere/ land surface

Figure 3 Schematic diagram summarizing the development of complexity in coupled physical–biogeochemical/ecological models. The two major axes of complexity are biogeochemistry/ecology and physics, but each major axis consists of many subaxes that describe the complexity of various subcomponents. Currently existing models fill a large portion of the multidimensional space opened by these axes. In addition, the evolution of models is not always necessarily straight along any given axis, but depends on the nature of the particular problem investigated. N, nutrient; P, phytoplankton; Z, zooplankton; B, bacteria; D, detritus; F, fish.

full justice. We restrict our article here to the discussion of four examples, which span the range of complexities as well as have been important milestones in the evolution of biogeochemical/ecological modeling. Box Models or What Controls Atmospheric Carbon Dioxide?

Our aim here is to develop a model that explains how the great biogeochemical loop controls atmospheric CO2. The key to answering this question is a quantitative prediction of the surface ocean concentration of

CO2, as it is this surface concentration that controls the atmosphere–ocean balance of CO2. The simplest models used for such a purpose are box models, where the spatial dimension is reduced to a very limited number of discrete boxes. Such box models have played an important role in ocean biogeochemical/ ecological modeling, mostly because their solutions can be readily explored and understood. However, due to the dramatic reduction of complexity, there are also important limitations, whose consequences one must keep in mind when interpreting the results. Box models can be formally derived from the tracer conservation eq [1] by integrating over



95

n

Higher complexity

ai

m

o ld

a

or

e mp tim d te U an CP int g o in p as id re r gr c In pe

Biogeochemical/ecological complexity

Ecological case studies (e.g., test-beds) short-term integrations

Climate-change studies century-timescale integrations

Temporal evolution

Physical−biological coupling studies short-term integrations

Geochemical applications long-term integrations

Physical complexity

Higher complexity

Figure 4 Schematic illustrating the relationship between the physical (abscissa) and the biogeochemical/ecological complexity (ordinate) of different typical applications of coupled physical–biogeochemical/ecological models. Given computational and analytical constraints, there is often a trade-off between the two main complexities. The thin lines in the background indicate the CPU time required for the computation of a problem over a given time period.

the volume of the box, V, and by applying Gauss’ (divergence) theorem, which states that the volume integral of a flux is equal to the flux in normal direction across the boundary surfaces, S. The resulting integral form of [1] is: Z

I

@C dV þ ðAdvn ðCÞ þ Diff n ðCÞÞ dS @t Z ¼ SMSðCÞ dV

½2

where Advn(C) and Diffn(C) are the advective and diffusive transports in the normal direction across S. If we assume that the concentration of tracer C as well as the SMS term within the box are uniform, [2]

can be rewritten as V

dC X þ Ti;out C Ti;in Ci þ ni ðC Ci Þ dt i ¼ V SMSðCÞ

½3

where C is the mean concentration within the box. We represented the advective contributions as products of a mass transport, T (dimensions volume time 1), with the respective upstream concentration, separately considering the mass transport into the box and out of the box across each interface i, that is, Ti,in and Ti,out. The eddy-diffusion contributions are parametrized as products of a mixing coefficient, ni (dimension volume time 1), with the gradient


96


across each interface ðC Ci Þ. For simplicity, we subsequently drop the overbar, that is, use C instead of C. The simplest box-model representation of the great biogeochemical loop is a two-box model, wherein the ocean is divided into a surface and a deep box, representing the euphotic and aphotic zones, respectively (Figure 5(a)). Mixing and transport between the two boxes is modeled with a mixing term, n. The entire complexity of marine ecology and biogeochemistry in the euphotic zone is reduced to a single term, F, which represents the flux of organic matter out of the surface box and into the deep box, where the organic matter is degraded back to inorganic constituents. Let us consider the deep-box balance for a dissolved inorganic bioreactive element, Cd (such as nitrate, phosphate, or dissolved inorganic carbon): Vd

dCd þ n ðCd Cs Þ ¼ F dt

½4

where Vd is the volume of the deep box and Cs is the dissolved inorganic concentration of the surface box. The steady-state solution of [4], that is, the solution when the time derivative of Cd vanishes (dCd/dt ¼ 0), n ðCd Cs Þ ¼ F

½5

represents the most fundamental balance of the great biogeochemical loop. It states that the net upward transport of an inorganic bioreactive element by physical mixing is balanced by the downward transport of this element in organic matter. This steady-state balance [5] has several important applications. For example, it permits us to estimate the downward export of organic matter by simply

analyzing the vertical gradient and by estimating the vertical exchange. The balance also states that the magnitude of the vertical gradient in bioreactive elements is proportional to the strength of the downward flux of organic matter and inversely proportional to the mixing coefficient. Since nearly all oceanic inorganic carbon and nutrients reside in the deep box, that is, Vd Cd cVs Cs , one can assume, to first order, that Cd is largely invariant, so that the transformed balance for the surface ocean concentration Cs ¼ Cd

F n

½6

reveals that the surface concentration depends primarily on the relative magnitude of the organic matter export to the magnitude of vertical mixing. Hence, [6] provides us with a first answer to our challenge: it states that, for a given amount of ocean mixing, the surface ocean concentration of inorganic carbon, and hence atmospheric CO2, will decrease with increasing marine export production. Relationship [6] also states that for a given magnitude of marine export production, atmospheric CO2 will increase with increased mixing. While very powerful, the two-box model has severe limitations. The most important one is that this model does not consider the fact that the deep ocean exchanges readily only with the high latitudes, which represent only a very small part of the surface ocean. The exchange of the deep ocean with the low latitudes is severly limited because diapcyncal mixing in the ocean is small. Another limitation is that the onebox model does not take into account that the nutrient concentrations in the high latitudes tend to be much higher than in the low latitudes (Figure 2(c)).

(a)

(b) High-lat

Surface Vs

Low-lat

Vh

Cs

Cl Vl

Ch T

ν

Cd

Vd

Deep

Three-box model

Two-box model

Φ

Φl

hd Φh Cd Vd

Deep

Figure 5 Schematic representation of the great biogeochemical loop in box models: (a) two-box model and (b) three-box model. C, concentrations; V, volume; n, exchange (mixing) coefficient; F, organic matter export fluxes; T, advective transport.



= +145

(a)

97

(c) = −120

(b)

Atmospheric pCO2 (ppm)

280

160

425

0

33

0

26

33

0

0

33

33

Nitrate balance (mmol−3) Window open

Preindustrial

Window closed

Figure 6 Schematic illustration of the impact of the high-latitude (Southern Ocean) window on atmospheric CO2. In (a) the highlatitude window is open, high-latitude nitrate is high, and atmospheric CO2 attains 425 ppm. Panel (b) represents the preindustrial ocean, where the window is half open, nitrate in the high latitudes is at intermediate levels, and atmospheric CO2 is 280 ppm. In (c), the high-latitude window is closed, high-latitude nitrate is low, and atmospheric CO2 decreases to about 160 ppm.

An elegant solution is the separation of the surface box into a high-latitude box, h, and a low-latitude box, l (Figure 5(b)). Intense mixing is assumed to occur only between the deep and the high-latitude boxes, nhd . A large-scale transport, T, is added to mimic the ocean’s global-scale overturning circulation. As before, the complex SMS terms are summarized by Fl and Fh. In steady state, the surface concentrations are given by Fl T Fh þ Fl Ch ¼ Cd T þ nhd Cl ¼ Cd

½7

which are structurally analogous to [6], but emphasize the different dynamics setting balances in the low and high latitudes. With nhd cT and FlEFh, [7] explains immediately why nutrient concentrations tend to be higher in high latitudes than in low latitudes (Figure 2(b)). In fact, writing [7] out for both nitrate ðNO 3 Þ and dissolved inorganic carbon (DIC), and using the observation that surface NO 3 in the low latitudes is essentially zero (Figure 2(c)), the DIC equations are DICl ¼ DICd rC:N ½NO 3 d DICh ¼ DICd

Fh þ rC:N T ½NO 3 d T þ nhd

½8

where rC:N is the carbon-to-nitrogen ratio of organic matter. Assuming that the deep-ocean nitrate concentration is time invariant, the analysis of [8] reveals that the low-latitude DIC concentration, DICl,

is more or less fixed, while the high-latitude DIC concentration, DICh, can be readily altered by changes in either the high-latitude export flux, Fh, or high-latitude mixing, nhd . Furthermore, if one considers the fact that the rapid communication between the high-latitude ocean and the deep ocean makes the high latitudes the primary window into the oceanic reservoir of inorganic carbon, it becomes clear that it must be the high latitudes that control atmospheric CO2 on the millenial timescales where the steadystate approximation is justified. This high-latitude dominance in controlling atmospheric CO2 is illustrated in Figure 6, which demonstrates that atmospheric CO2 can vary between about 160 and 425 ppm by simply opening or closing this highlatitude window. The exact magnitude of the atmospheric CO2 change depends, to a substantial degree, on the details of the model, but the dominance of high-latitude processes in controlling atmospheric CO2 is also found in much more complex and spatially explicit models. As a result, a change in the carbon cycle in the high latitudes continues to be the leading explanation for the substantially lower atmospheric CO2 concentrations during the ice ages. NP and NPZ Models, or What Simple Ecosystem Models Can Say about Oceanic Productivity

So far, we have represented the ecological, chemical, and physical processes that control the production of organic matter and its subsequent export with a single parameter, F. Clearly, in order to assess how marine biology responds to climate change and other perturbations, it is necessary to resolve these


98


processes in much more detail. Two fundamentally different approaches have been developed to model such ecological processes: concentration-based models and individually based models (IBMs). In the latter, the model’s equations represent the growth and losses of individual organisms, and the model then simulates the evolution of a large number of these individuals through space and time. This approach is most commonly used for the modeling of organisms at higher trophic levels, where life cycles play an important role (e.g., models of zooplankton, fish, and marine mammals). In contrast, concentration-based models are almost exclusively used for the modeling of the lower trophic levels, in particular to represent the interaction of light, nutrient, and grazing in controlling the growth of phytoplankton, that is, marine primary production. The simplest such model considers just one limiting nutrient, N, and one single phytoplankton group, P (see also Figure 7(a)): N P lP P KN þ N N P þ mP l P P SMSðNÞ ¼ Vmax KN þ N SMSðPÞ ¼ Vmax

½9

where the first term on the right-hand side of the phytoplankton eqn [9] is net phytoplankton growth (equal to net primary production) modeled here as a function of a nutrient-saturated growth rate Vmax, and a hyperbolic dependence on the in situ nutrient concentration (Monod-type), with a single parameter, the half-saturation constant, KN. The second term is the net loss due to senescence, viral infection, and grazing, modeled here as a linear process with a loss rate lP . A fraction mP of the phytoplankton loss is assumed to be regenerated inside the euphotic zone, while a fraction (1 mP) is exported to depth (Figure 7(a)). These two parametrizations for phytoplankton growth and losses reflect a typical situation for marine ecosystem models in that the growth terms are substantially more elaborate, reflecting the interacting influence of various controlling parameters, whereas the loss processes are highly simplified. This situation also reflects the fact that the processes controlling the growth of marine organisms are often more amenable to experimental studies, while the loss processes are much harder to investigate with careful experiments. When the SMS terms of [10] are inserted into the full tracer conservation eqn [1], the resulting equations form a set of coupled partial differential equations with a number of interesting consequences in steady state as shown in Figure 8(a). Below a certain nutrient concentration threshold, phytoplankton

growth is smaller than its loss term, so that the resulting steady-state solution is P ¼ 0, that is, no phytoplankton. Above this threshold, the abundance of phytoplankton (and hence primary production) increases with increasing nutrient supply in a linear manner. Phytoplankton is successful in reducing the dissolved inorganic nutrient concentration to low levels, so that the steady state is characterized by most nutrients residing in organic form in the phytoplankton pool. This NP model reflects a situation where marine productivity is limited by bottom-up processes, that is, the supply of the essential nutrients. Comparison with the nutrient and phytoplankton distribution shown in (Figure 2(b)) shows that this model could explain the observations in the subtropical open ocean, where nutrients are indeed drawn down to very low levels. However, the NP model would also predict relatively high P levels to go along with the low nutrient levels, which is clearly not observed (Figure 2(a)). In addition, this NP model also fails clearly to explain the high-nutrient regions of the high latitudes. However, we have so far neglected the limitation by light, as well as the impact of zooplankton grazing. The addition of a zooplankton compartment to the NP model (Figure 7(b)) dramatically shifts the nutrient allocation behavior (Figure 8(b)). In this case, as the nutrient loading increases, the phytoplankton abundance gets capped at a certain level by zooplankton grazing. Due to the reduced levels of biomass, the phytoplankton is then no longer able to consume all nutrients at high-nutrient loads, so that an increasing fraction of the supplied nutrients remains unused. This top-down limitation situation could therefore, in part, explain why macronutrients in certain regions remain untapped. Most recent research suggests that micronutrient (iron) limitation, in conjunction with zooplankton grazing, plays a more important role in causing these highnutrient/low chlorophyll regions. Another key limitation is light, which has important consequences for the seasonal and depth evolution of phytoplankton. Global 3-D Modeling of Ocean Biogeochemistry/ Ecology

The coupling of relatively simple NPZ-type models to global coarse-resolution 3-D circulation models has proven to be a challenging task. Perhaps the most important limitation of such NPZ models is the fact that all phytoplankton in the ocean are represented by a single phytoplankton group, which is grazed upon by a single zooplankton group. This means that the tiny phytoplankton that dominate the relatively nutrient-poor central gyres of the ocean, and the



99

(a) N−P model

Mortality and respiration

Phytoplankton (P )

P · P · P

Export

Remineralization

Phytoplankton growth

Nutrient supply

Γ(N )

N KN + N

P · Vmax ·

DIN + DON (N )

P · P

(1 – P ) · P · P

(b) N−P−Z model

Phytoplankton growth

Phytoplankton (P)

P · P · P P · P Export

Remineralization

Γ (N )

Nutrient supply

Remineralization

DIN + DON (N)

N KN + N

Mortality and respiration

Grazing

P · Vmax ·

g·Z·P KP

(1 – P ) · P · P

Zooplankton (Z )

(1 – Z ) ·

g·Z·P + Z · Z KP

Export

Z ·

(1 – Z ) · Mortality and respiration and egestion (1 – Z ) ·

(1 – Z ) ·

g·Z·P + Z ·Z KP

g·Z·P + Z · Z KP

Figure 7 A schematic illustration of (a) a nitrate–phytoplankton model and (b) a nitrate–phytoplankton–zooplankton model. The mathematical expressions associated with each arrow represent typical representations for how the individual processes are parametrized in such models. Vmax, maximum phytoplankton growth rate; KN, nutrient half saturation concentration; g , maximum zooplankton growth rate; KP, half saturation concentration for phytoplankton grazing; gZ, zooplankton assimilation efficiency; lP, phytoplankton mortality rate; lZ, zooplankton mortality rate; mP, fraction of dead phytoplankton nitrogen that is remineralized in the euphotic zone; mZ, as mP, but for zooplankton. Adapted from Sarmiento JL and Gruber N (2006) Ocean Biogeochemical Dynamics, 526pp. Princeton, NJ: Princeton University Press.


100


(a) 0.03

(b) 10

NP-model

NPZ-model

0.02 NT

=

Nn

P, Nn + P (mmol m–3)

P, Nn + P (mmol m–3)

8 +

P

0.01 Nn

P

6

Nn

4

2 Z

0

0

0.01

0.02 NT (mmol m–3)

0.03

0

P 0

2

4 6 NT (mmol m–3)

8

10

Figure 8 Nitrogen in phytoplankton (P), zooplankton (Z), and nitrate (Nn) as a function of total nitrogen. (a) Results from a twocomponent model that just includes N and P. (b) Results from a three-component model that is akin to that shown in Figure 7(b). The vertical axis is concentration in each of the components plotted as the cumulative amount. Adapted from Sarmiento JL and Gruber N (2006) Ocean Biogeochemical Dynamics, 526pp. Princeton, NJ: Princeton University Press.

large phytoplankton that dominate the highly productive upwellling regions are represented with a single group, whose growth characteristics cannot simultaneously represent both. An additional challenge of the large-scale coupling problem is the interaction of oceanic circulation/mixing with the marine ecological and biogeochemical processes. Often, small deficiencies in the physical model get amplified by the ecological/biogeochemical model, so that the resulting fields of phytoplankton abundance may diverge strongly from observations. The latter problem is addressed by improving critical aspects of the physical model, such as upper ocean mixing, atmospheric forcing, resolution, and numerical tracer transport algorithms, while the first problem is addressed by the consideration of distinct plankton functional groups. These functional groups distinguish themselves by their different size, nutrient requirement, biogeochemical role, and several other characteristics, but not necessarily by their taxa. Currently existing models typically consider the following phytoplankton functional groups: small phytoplankton (nano- and picoplankton), silicifying phytoplankton (diatoms), calcifying phytoplankton (coccolithophorids), N2fixing phytoplankton, large (nonsilicifying) phytoplankton (e.g., dinoflagellates and phaeocystis), with the choice of which group to include being driven by the particular question at hand. The partial differential equations for the different phytoplankton functional groups are essentially the same and follow a structure similar to [9], but are differentiated by varying growth parameters, nutrient/light requirements, and susceptibility to grazing. At present, most applications include between five and 10

phytoplankton functional groups. Some recent studies have been exploring the use of several dozen functional groups, whose parameters are generated stochastically with some simple set of rules that are then winnowed via competition for resources among the functional groups. Since the different phytoplankton functional groups are also grazed differentially, different types of zooplankton generally need to be considered as well, though fully developed models with diverse zooplankton functional groups are just emerging. An alternative is to use a single zooplankton group, but with changing grazing/ growth characteristics depending on its main food source. A main advantage of models that build on the concept of plankton functional groups is their ability to switch between different phytoplankton community structures and to simulate their differential impact on marine biogeochemical processes. For example, the nutrient poor subtropical gyres tend to be dominated by nano- and picoplankton, whose biomass tends to be tightly capped by grazing by very small zooplankton. This prevents the group forming blooms, even under nutrient-rich conditions. By contrast, larger phytoplankton, such as diatoms, can often escape grazing control (at least temporarily) and form extensive blooms. Such phytoplankton community shifts are essential for controlling the export of organic matter toward the abyss, since the export ratio, that is, the fraction of net primary production that is exported, tends to increase substantially in blooms. Results from the coupling of such a multiple functional phytoplankton ecosystem model and of a relatively simple biogeochemical model to a global 3-D circulation model show substantial success in


102


(a)

(b)

34.6° N

Nitrate (mmol−3)

Chlorophyll (mg m−3)

34.2° N 33.8° N 33.4° N 33.0° N 32.6° N 121° W

0

120° W

0.4 0.8 1.2 1.6

119° W

2

118° W

2.4 2.8 3.2 3.6

121° W

4

15

0.1

120° W

0.17

0.29

119° W

0.5

0.87

118° W

1.5

2.5

4.4

Figure 10 Snapshots of (a) surface nitrate and (b) surface chlorophyll as simulated by a 1-km, that is, eddy-resolving, regional model for the Southern California Bight. The snapshots represent typical conditions in response to a spring-time upwelling event in this region. The model consists of a multiple phytoplankton functional group model that has been coupled to the Regional Ocean Modeling System (ROMS). Results provided by H. Frenzel, UCLA.

simulations). This is currently feasible for only very short periods, so that limited domain models are often used instead. Eddy-resolving Regional Models

Limited domain models, as shown in Figure 10 for the Southern California Bight, that is, the southernmost region of the West Coast of the US, can be run at much higher resolutions over extended periods, permitting the investigation of the processes occurring at the kilometer scale. The most important such processes are meso- and submesoscale eddies and other manifestations of turbulence, which are ubiquitous features of the ocean. The coupling of the same ecosystem/biogeochemical model used for the global model shown in Figure 9 to a 1-km resolution model of the Southern California Bight reveals the spatial richness and sharp gradients that are commonly observed (Figure 10) and that emerge from the intense interactions of the physical and biogeochemical/ecological processes at the eddy scale. Since such mesoscale processes are unlikely to be resolved at the global scale for a while, a current challenge is to find parametrizations that incorporate the net impact of these processes without actually resolving them explicitly.

Future Directions The modeling of marine biogeochemical/ecological processes is a new and rapidly evolving field, so that

predictions of future developments are uncertain. Nevertheless, it is reasonable to expect that models will continue to evolve along the major axes of complexity, including, for example, the consideration of higher trophic levels (Figure 4). One also envisions that marine biogeochemical/ecological models will be increasingly coupled to other systems. A good example are Earth System Models that attempt to represent the entire climate system of the Earth including the global carbon cycle. In those models, the marine biogeochemical/ecological models are just a small part, but interact with all other components, possibly leading to complex behavior including feedbacks and limit cycles. Another major anticipated development is the increasing use of inverse modeling approaches to optimally make use of the increasing flow of observations.

Glossary advection The transport of a quantity in a vector field, such as ocean circulation. anthropogenic carbon The additional carbon that has been released to the environment over the last several centuries by human activities including fossil-fuel combustion, agriculture, forestry, and biomass burning. biogeochemical loop, great A set of key processes that control the distribution of bioreactive elements in the ocean. The loop starts with the photosynthetic production of organic matter in the light-illuminated upper ocean, a fraction of which



is exported to depth mostly by sinking of particles. During their sinking in the dark deeper layers of the ocean, this organic matter is consumed by bacteria and zooplankton that transform it back to its inorganic constituents while consuming oxidized components in the water (mostly oxygen). The great biogeochemical loop is closed by the physical transport and mixing of these inorganic bioreactive elements back to the near-surface ocean, where the process starts again. This great biogeochemical loop also impacts the distribution of many other elements, particularly those that are particle reactive, such as thorium. coccolithophorids A group of phytoplankton belonging to the haptophytes. They are distinguished by their formation of small mineral calcium carbonate plates called coccoliths, which contribute to the majority of marine calcium carbonate production. An example of a globally significant coccolithophore is Emiliania huxleyi. diatoms A group of phytoplankton belonging to the class Bacillariophyceae. They are one of the most common types of phytoplankton and distinguish themselves through their production of a cell wall made of amorphous silica, called opal. These walls show a wide diversity in form, but usually consist of two symmetrical sides with a split between them. diffusion The transport of a quantity by Brownian motion (molecular diffusion) or turbulence (eddy diffusion). The latter is actually some form of advection, but since its net effect is akin to diffusion, it is often represented as a diffusive process. export production The part of the organic matter formed in the surface layer by photosynthesis that is transported out of the surface layer and into the interior of the ocean by particle sinking, mixing, and circulation, or active transport by organisms. models, concentration based Models, in which the biological state variables are given as numbers per unit volume, that is, concentration. This approach is often used when physical dispersion plays an important role in determining the biological and biogeochemical impact of these state variables. Concentration-based models are most often used to represent lower trophic levels in the ocean. models, forward Models, in which time is integrated forward to compute the temporal evolution of the distribution of state variables in response to a set of initial and boundary conditions. models, individually based Models, in which the biological state variables represent individual organisms or a small group of individual

103

organisms. This approach is often used when life cycles play a major role in the development of these organisms, such as is the case for most marine organisms at higher trophic levels. models, inverse Models that aim to optimally combine observations with knowledge about the workings of a system as embodied in the model. Solutions to such inverse models can be improved estimates of the current state of the system (state estimation), improved estimates of the initial or boundary conditions, or an optimal set of parameters. A typical example of an inverse model is the optimal determination of ecological parameters, such as growth and grazing rates, given, for example, the observed distribution of phytoplankton, zooplankton, and nutrients. net primary production Rate of net fixation of inorganic carbon into organic carbon by autotrophic phytoplankton. This net rate is the difference between the gross uptake of inorganic carbon during photosynthesis and autotrophic respiration. organisms, autotrophic An autotroph (from the Greek autos ¼ self and trophe ¼ nutrition) is an organism that produces organic compounds from carbon dioxide as a carbon source, using either light or reactions of inorganic chemical compounds, as a source of energy. An autotroph is known as a producer in a food chain. organisms, heterotrophic A heterotroph (Greek heterone ¼ (an)other and trophe ¼ nutrition) is an organism that requires organic substrates to get its carbon for growth and development. A heterotroph is known as a consumer in the food chain. plankton Organisms whose swimming speed is smaller than the typical speed of ocean currents, so that they cannot resist currents and are hence unable to determine their horizontal position. This is in contrast to nekton organisms that can swim against the ambient flow of the water environment and control their position (e.g., squid, fish, krill, and marine mammals). plankton, phytoplankton Phytoplankton are the (photo)autotrophic components of the plankton. Phytoplankton are pro- or eukaryotic algae that live near the water surface where there is sufficient light to support photosynthesis. Among the more important groups are the diatoms, cyanobacteria, dinoflagellates, and coccolithophorids. plankton, zooplankton Zooplankton are heterotrophic plankton that feed on other plankton. Dominant groups include small protozoans or metazoans (e.g., crustaceans and other animals).


104


See also Biogeochemical Data Assimilation. Carbon Cycle. Carbon Dioxide (CO2) Cycle. Forward Problem in Numerical Models. Inverse Modeling of Tracers and Nutrients. Mesoscale Eddies. Nitrogen Cycle. Ocean Carbon System, Modeling of. Ocean Circulation. Phosphorus Cycle.

Further Reading DeYoung B, Heath M, Werner F, Chai F, Megrey B, and Monfray P (2004) Challenges of modeling ocean basin ecosystems. Science 304: 1463--1466. Doney SC (1999) Major challenges confronting marine biogeochemical modeling. Global Biogeochemical Cycles 13(3): 705--714. Fasham M (ed.) (2003) Ocean Biogeochemistry. New York: Springer.

Fasham MJR, Ducklow HW, and McKelvie SM (1990) A nitrogen-based model of plankton dynamics in the oceanic mixed layer. Journal of Marine Systems 48: 591--639. Glover DM, Jenkins WJ, and Doney SC (2006) Course No. 12.747: Modeling, Data Analysis and Numerical Techniques for Geochemistry. http://w3eos.whoi.edu/ 12.747 (accessed in March 2008). Rothstein L, Abbott M, Chassignet E, et al. (2006) Modeling ocean ecosystems: The PARADIGM Program. Oceanography 19: 16--45. Sarmiento JL and Gruber N (2006) Ocean Biogeochemical Dynamics, 526pp. Princeton, NJ: Princeton University Press.

Relevant Websites http://www.uta.edu – Interactive models of ocean ecology/biogeochemistry.


OCEAN CARBON SYSTEM, MODELING OF S. C. Doney and D. M. Glover, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Chemical species such as radiocarbon, chlorofluorocarbons, and tritium–3He are important tools for ocean carbon cycle research because they can be used to trace circulation pathways, estimate timescales, and determine absolute rates. Such species, often termed chemical tracers, typically have rather simple water-column geochemistry (e.g., conservative or exponential radioactive decay), and reasonably well-known time histories in the atmosphere or surface ocean. Large-scale ocean gradients of nutrients, oxygen, and dissolved inorganic carbon reflect a combination of circulation, mixing, and the production, transport, and oxidation (or remineralization) of organic matter. Tracers provide additional, often independent, information useful in separating these biogeochemical and physical processes. Biogeochemical and tracer observations are often framed in terms of ocean circulation models, ranging from simple, idealized models to full three-dimensional (3-D) simulations. Model advection and diffusion rates are typically calibrated or evaluated against transient tracer data. Idealized models are straightforward to construct and computationally inexpensive and are thus conducive to hypothesis testing and extensive exploration of parameter space. More complete and sophisticated dynamics can be incorporated into three-dimensional models, which are also more amenable for direct comparisons with field data. Models of both classes are used commonly to examine specific biogeochemical process, quantify the uptake of anthropogenic carbon, and study the carbon cycle responses to climate change. All models have potential drawbacks, however, and part of the art of numerical modeling is deciding on the appropriate model(s) for the particular question at hand. Carbon plays a unique role in the Earth’s environment, bridging the physical and biogeochemical systems. Carbon dioxide (CO2), a minor constituent in the atmosphere, is a so-called greenhouse gas that helps to modulate the planet’s climate and temperature. Given sunlight and nutrients, plants and some microorganisms convert CO2 via photosynthesis into organic carbon, serving as the building blocks and

energy source for most of the world’s biota. The concentration of CO2 in the atmosphere is affected by the net balance of photosynthesis and the reverse reaction respiration on land and in the ocean. Changes in ocean circulation and temperature can also change CO2 levels because carbon dioxide is quite soluble in sea water. In fact, the total amount of dissolved inorganic carbon (DIC) in the ocean is about 50 times larger than the atmospheric inventory. The air–sea exchange of carbon is governed by the gas transfer velocity and the surface water partial pressure of CO2 (pCO2), which increases with warmer temperatures, higher DIC, and lower alkalinity levels. The natural carbon cycle has undergone large fluctuations in the past, the most striking during glacial periods when atmospheric CO2 levels were about 30% lower than preindustrial values. The ocean must have been involved in such a large redistribution of carbon, but the exact mechanism is still not agreed upon. Human activities, including fossil-fuel burning and land-use practices such as deforestation and biomass burning, are altering the natural carbon cycle. Currently about 7.5 Pg C yr 1 (1 Pg ¼ 1015g) are emitted into the atmosphere, and direct measurements show that the atmospheric CO2 concentration is indeed growing rapidly with time. Elevated atmospheric CO2 levels are projected to heat the Earth’s surface, and the evidence for climate warming is mounting. Only about 40% of the released anthropogenic carbon remains in the atmosphere, the remainder is taken up in about equal portions (or 2 Pg C yr 1) by land and ocean sinks (Figure 1). The future magnitude of these sinks is not well known, however, and is one of the major uncertainties in climate simulations. Solving this problem is complicated because human impacts appear as relatively small perturbations on a large natural background. In the ocean, the reservoir of organic carbon locked up as living organisms, mostly plankton, is only about 3 Pg C. The marine biota in the sunlit surface ocean are quite productive though, producing roughly 50 Pg of new organic carbon per year. Most of this material is recycled near the ocean surface by zooplankton grazing or microbial consumption. A small fraction, something like 10–20% on average, is exported to the deep ocean as sinking particles or as dissolved organic matter moving with the ocean circulation. Bacteria and other organisms in the deep ocean feed on this source of organic matter from above, releasing DIC and associated nutrients back into the


105

106

OCEAN CARBON SYSTEM, MODELING OF

Atmosphere 750

50 ion

t

c du ro p n y ar atio rim spir p t e ne d r al an b o Gl 51.4

Vegetation 610 Soils and detritus 1580 2190

d

ing

lan

1.6 e us

Fossil fuels and cement production

g

an

Ch

0.5

92

n Erosio

Surface ocean 1020 40 0.8 Rivers

We

5.5

90.6

100

50

Marine biota 3

ring he at

8

91.6

2 DOC 40 S)

17

(b)

(a) 2.5

11 15

17 16 2.0

6 1 14

5 19 12 8 13 4 18 2 3

11 1.5

15

7

60 9 50

9

40 30

10

1.0

20 10

0.5

70

3.0 14

(c) Southern Ocean ( wb b @t Dt

½1

S is defined as positive when fluid is transferred into the thermocline (Figure 4); h is the thickness of the mixed layer, wb and ub are the vertical velocity and horizontal velocity vector, respectively, at the base of the mixed layer, and Db/Dt q/qt þ ubr is the Lagrangian rate of change following the horizontal


158

OCEAN SUBDUCTION

(A) September outcrop y x

March outcrop

(B) March position of isopycnal

z

September position of isopycnal

y

Seasonal boundary

Main thermocline

flow at the base of the mixed layer. The subduction rate is defined as the volume flux per unit horizontal area in [1] relative to the base of the instantaneous mixed layer. Hence, S includes a contribution from the change in thickness of the mixed layer, as well as from the vertical and horizontal movement of fluid particles. Equivalently, Dt in [1] can be evaluated from the vertical velocity and the Lagrangian change in mixed-layer thickness following the horizontal circulation. The subduction into the main thermocline is physically more important than that into the seasonal thermocline, since water-mass properties become shielded from the atmosphere for longer timescales within the main thermocline. The annual rate of subduction into the main thermocline, Dt, is evaluated from the volume flux per unit horizontal area passing through a control surface H that overlies the main thermocline (eqn [2]).

Interface

Figure 3 A schematic diagram illustrating the seasonal rectification of subduction for (A) a plan view at the sea surface and (B) a meridional section. Over the cooling season in the Northern Hemisphere, the outcrop of a density surface migrates equatorward from September to March (denoted by dashed and full lines). Consider a fluid parcel (circle) subducted from the vertically homogeneous mixed layer and advected equatorwards along this density surface. If the fluid parcel is advected past the March outcrop during the year, then it is subducted in to the main thermocline (the upper extent of which is shown by the dotted line), otherwise it is eventually entrained into the mixed layer. (Reproduced with permission from Williams et al. (1995).)

t

Sann ¼ wH uHd rH

½2

wH and uH are the vertical velocity and horizontal velocity vector at the depth H. The control surface is defined by the base of the seasonal thermocline given by the maximum thickness of the winter mixed layer (Figure 1). Alternatively, Sann may be evaluated directly from S in eqn [1] using a Lagrangian integration following the movement of a water column over an annual cycle. For example, for the mixed-layer cycle shown in Figure 2, Sann controls the vertical spacing between isopycnals subducted at the end of consecutive winters. Hence, a high rate of subduction leads to the formation of a mode water with a weak stratification (and a large vertical spacing between subducted isopycnals).

Gyre-scale Subduction h(t )

Dh Dt

Δt

S( t ) Δt bΔt

Figure 4 A schematic diagram showing a particle being subducted from the time-varying base of the mixed layer into the thermocline. The vertical distance between the particle and the mixed layer, S(t)Dt, is the sum of the vertical displacement of the particle, wbDt, and the shallowing of the mixed layer following the particle, (Dh/Dt)Dt; where h is the thickness of the mixed layer, wb is the vertical velocity and Dt is a time interval. (Reproduced with permission from Marshall et al. (1993).)

Subduction occurs over ocean basins predominantly through the gyre-scale circulation. The surface windstress drives anticyclonic and cyclonic recirculations, referred to as subtropical and subpolar gyres, respectively, within a basin. The wind forcing induces downwelling of surface fluid over the subtropical gyre and upwelling over the subpolar gyre. The gyrescale subduction rate defined in eqn [2] depends on both the vertical and horizontal circulations together with the thickness of the end of winter mixed layer. Subduction Rate over the North Atlantic

The North Atlantic is the most actively ventilated oceanic basin. The wind-induced downwelling


OCEAN SUBDUCTION

reaches magnitudes of 25–50 m y1 over the North Atlantic (Figure 5A), which is characteristic of most basins. However, the annual subduction rate is significantly enhanced over the vertical transfer through the lateral transfer across the sloping base of the mixed layer. The surface buoyancy loss to the atmosphere leads to the winter mixed layer thickening poleward, from typically 50 m in the subtropics (e.g., at 201N) to 500 m or more in the subpolar gyre (e.g., at 501N). Over the subtropical gyre, the annual subduction rate reaches between 50 m y1 and 100 m y1

_5 0 60

_ 50

_ 25

0 25

40

50

˚N

0 _ 50

25

_ 25

(A)

60

_ 200 _ 10 0 _ 50 0

050

_ 20

0

40

_ 100

˚N

100

50

20 50 0

80 (B)

50

_1 00

050 _

40

(Figure 5B). South of the Gulf Stream, there is a band of high subduction rates that are controlled by the lateral transfer. Elsewhere, the vertical and horizontal transfers are comparable over the subtropical gyre. Over the subpolar gyre, fluid is generally transferred from the main thermocline into the seasonal boundary layer. Fluid is eventually returned from the seasonal boundary layer and into the interior through deep convection events or through subduction along the western boundary of the subpolar gyre. The negative subduction rates reach several 100 m y1 over the North Atlantic (Figure 5B), which is controlled by the lateral transfer rather than the vertical transfer. This negative subduction leads to water-mass properties of the main thermocline being transferred or inducted into the downstream winter mixed layer. This induction process is particularly important for the biogeochemistry in transferring inorganic nutrients from the thermocline into the winter mixed layer. These nutrients are entrained into the euphotic zone and enable high values of biological production to occur. Evidence from Transient Tracers

50

20

159

0

˚W

Figure 5 Diagnostics for (A) wind-driven Ekman downwelling velocity (contours every 25 m y1) and (B) annual subduction rate into the main thermocline of the North Atlantic (contours every 50 m y1). The annual subduction rate is evaluated from SH ¼ wH uH :,H with the interface H defined from the mixed-layer thickness at the end of winter, uH evaluated from a density climatology using thermal-wind balance, and wH derived from the wind-stress climatology and linear vorticity balance. The subduction rate represents the volume flux per unit horizontal area which ventilates the stratified thermocline. (Reproduced with permission from Marshall et al. (1993).)

Characteristic signals of subduction and ventilation are provided by distributions of transient tracers, such as tritium and CFCs (chloroflurocarbons), over the ocean interior. Quantitative information on the ventilation process is provided when the transient tracers have a known source function and lifetime. For example, in the North Atlantic, the tritium–helium age distribution along potential density surfaces reveals the general ventilation of the subtropical gyre from the north east and gradual aging following the anticyclonic circulation (Figure 6). The invasion of transient tracers into the main thermocline provides an integrated measure of ventilation including the contribution of the time-mean circulation and the rectified contribution of eddies. For example, the rate of ventilation inferred from the tracer age distribution is two to three times greater than the rate of wind-induced (Ekman) downwelling, which is possibly consistent with the gyre-scale subduction rate diagnostics in Figure 5. However, the rate of ventilation inferred from influx of tracers also includes a diffusive contribution, which differs for different tracers, making a more exact comparison with the subduction rate based on volume fluxes difficult to achieve.

Subduction and the Main Thermocline Formation of the Main Thermocline

Throughout the ocean there is a persistent thermocline separating the mixed layer and the weakly


160

vertical convergence of downwelling of subtropical fluid and upwelling of denser fluid originating from outside the subtropical gyre, such as from the subpolar gyre or Southern Ocean.

20 ˚E

70˚N

100 ˚W 80 ˚W 60 ˚W 40 ˚W 20 ˚W 0˚

OCEAN SUBDUCTION

70˚N 60˚N

60˚N 50˚N

50˚N 3

30˚N

40˚N

1

3

40˚N

2

30˚N

4 6

5

20˚N

20˚N

7

10˚N

10˚N

70˚N

60˚ W

40˚ W

20˚ W

20˚ W

60˚N

0˚

50˚N

0˚ 20˚E

70˚N 50˚N

2

10

30˚N 20˚N

8

10˚N 0˚ 100˚ W 80˚ W

10˚N 0˚ 60˚ W

40˚ W

Q¼

40˚N

8

20˚N

(B)

The subduction process helps to determine the stratification within the main thermocline, and hence the distribution of a dynamic tracer, the potential vorticity. Fluid parcels conserve potential vorticity for adiabatic, inviscid flow. Potential vorticity depends on the absolute spin of the fluid and stratification. A large-scale estimate of potential vorticity is provided by eqn [3] where f ¼ 2O sin y is the Coriolis parameter or planetary vorticity, O is the Earth’s angular velocity, y is the latitude, s is a potential density and r is a reference density.

60˚N

0

40˚N 30˚N

0˚

˚E

0˚ 100˚ W 80˚ W

40˚ W

100 ˚W 80 ˚W 60 ˚W 40 ˚W 20 ˚W 0˚

(A)

60˚ W

20

80˚ W

Dynamical Tracer Distributions

20˚ W

0˚

20˚E

Figure 6 Observations of the tritium–helium age (years) on potential density surfaces sy ¼ 26:5 (A) and sy ¼ 26:75 (B) in the North Atlantic. These surfaces are ventilated from the north east and the tritium–helium age increases following the anticyclonic circulation of the subtropical gyre. (Reproduced with permission from Jenkins WJ (1988) The use of anthropogenic tritium and helium-3 to study subtropical gyre ventilation and circulation. Philosophical Transactions of the Royal Society of London A 325, 43–61.)

stratified deep ocean. In ideal thermocline theory, the subduction process forms an upper thermocline over the subtropical gyre. The wind-driven circulation drives downwelling over a subtropical gyre, which leads to fluid being subducted from the mixed layer into the adiabatic interior of the ocean — see the seminal work of Luyten et al. (1983) listed in the Further Reading section. The thermocline is an advective feature formed through the circulation tilting the horizontal contrast in surface density over the subtropical gyre into the vertical. The subducted thermocline extends over the mixed-layer density range within the subtropical gyre. However, observations reveal that the thermocline also extends over denser isopycnals that do not outcrop within the subtropical gyre. Thermocline models incorporating diffusion suggest that this denser thermocline may be formed through the

f @s r @z

½3

Climatological maps of large-scale potential vorticity over the main thermocline reveal different regimes consisting of open, closed, or blocked Q contours along potential density surfaces (Figure 7). The open Q contours thread from the stratified interior to the mixed-layer outcrop at the end of winter (Figure 7A). Conversely, closed Q contours do not intersect the mixed layer and usually contain regions of nearly uniform Q (Figure 7B). The blocked contours run zonally from coast to coast and are usually associated with weak meridional flow unless directly forced. The open Q contours dominate for lighter surfaces, whereas the closed or blocked Q contours dominate for denser surfaces. Competing Paradigms: Gyre-scale Subduction versus Eddy Stirring

Different hypotheses involving gyre-scale subduction or eddy stirring have been invoked to explain these contrasting tracer distributions seen in the data. Open tracer contours (such as in Figure 7A) are usually associated with gyre-scale subduction. Fluid is subducted from the end of winter mixed layer and, when the flow is adiabatic and inviscid, streamlines become coincident with Q contours in the thermocline. A thermocline model example is shown in Figure 8, where Q along potential density surfaces in the thermocline is determined by subduction from the overlying mixed layer. The isopleths of Q reveal the anticyclonic circulation of the subtropical gyre. The Q contrast in the thermocline can become small given realistic end of winter mixed-layer variations. In an ideal thermocline model, each potential


OCEAN SUBDUCTION

65˚

80˚

70˚

60˚

50˚

W 40˚

20˚ 10˚

30˚

0˚ 40

60˚

50˚ 200

5

30 0

0.2

5

0.2

100

20

40

30˚

0.30

20

10

30

40

30

_40

_20

_30

Longitude (degree)

40 0 1 cm s_1

(A) 0˚

40

60˚

50˚

100 30

50

N 40 ˚

Latitude (degree)

65˚

0.13

30

50

20

20 15

10

0.10

30˚

0.15

0.15 0.18

10

_60 20˚

_50

(B) 10

_40

_30

_20

Longitude (degree)

10 5

10˚ (B) 0˚

_50

(A)

20

20

0.50

0.30

_60

20˚ 10˚

0.1

0.13

200

200

N 40 ˚

Latitude (degree)

90˚

161

0 1 cm s_1

Figure 7 Diagnosed distribution of a dynamic tracer, potential vorticity Q (1011 m1 s1), over the North Atlantic. (A) Between the potential density sy ¼ 26:3 and 26.5 surfaces, the Q contours appear to be closed in the north-west Atlantic, or to be open and thread back to the winter outcrop (shaded area) in the north-east Atlantic. (B) Between the sy ¼ 26:5 and 27.0 surfaces, the Q field appears to be relatively homogeneous south of the winter outcrop over most of the subtropical gyre. (Reproduced with permission from McDowell et al. (1982).)

density surface can be separated into a ventilated zone, which divides unventilated regions of a western pool and an eastern shadow zone. The unventilated zones are not directly connected by the time-mean streamlines to the overlying mixed layer, although they may still be indirectly ventilated through mixing acting along the coastal boundaries. The ventilated zone extends over most of the subtropical gyre for

Figure 8 An ideal thermocline model solution for the dynamic tracer, potential vorticity Q (109 m1 s1), along (A) sy ¼ 26.4 surface and (B) the sy ¼ 26.75 surface. Fluid is subducted across the density outcrop in the mixed layer (thick dashed line) into the stratified thermocline within a ventilated zone (between the thin dashed lines). Streamlines are coincident with the potential vorticity contours in the adiabatic interior. Given appropriate mixed-layer variations, the subducted potential vorticity can become nearly uniform, as obtained in (B). These model solutions are for an imposed anticyclonic wind forcing and a mixed layer, which becomes thicker to the north and denser to the north and east. (Reproduced with permission from Williams (1991).)

light potential density surfaces, but contracts eastwards for denser surfaces. Nearly uniform tracer distributions (such as in Figure 7B) are instead usually associated with stirring by mesoscale eddies. The eddy stirring can homogenize conserved tracers within closed streamlines. An eddy-resolving model example is shown in Figure 9 for a pair of wind-driven gyres. Advection


OCEAN SUBDUCTION

0 12

100

120

120

100 100

100

80 80

80

80

80

(A)

100 100

100

10

0

0

100

0

100

10

by the gyre circulation leads to tracer contours and streamlines becoming nearly coincident and closed over much of the interior domain. Eddy stirring transfers tracer down-gradient within these closed contours, which forms extensive regions of nearly uniform tracer. This interpretation is supported through observations of nearly uniform distributions of tritium, as well as potential vorticity, along weakly ventilated potential density surfaces within the winddriven gyres of the North Pacific and North Atlantic. Gyre-scale subduction and eddy stirring are not mutually exclusive processes, and can occur simultaneously and modify each other (as revealed in general circulation model experiments). Ventilation helps control the input of tracer and the tracer contrast along the winter outcrop of the potential density surface, while eddies act to smear out subducted mode waters and tracer contrasts in the ocean interior. Diagnostics from transient tracers and float trajectories suggest that the implied Peclet number (a nondimensional measure of advection/diffusion) is typically less than 10. This value is much smaller than the high value expected from nondiffusive, ideal thermocline models. Hence, eddy diffusion appears to be significant along potential density surfaces throughout a basin and is particularly important in regions of weak background flow.

10

162

The Relationship between Subduction, Potential Vorticity, and Buoyancy Forcing

Local, kinematic connection The relationship between subduction rate, the mixed-layer evolution, and underlying potential vorticity is illustrated here. For a particle subducted from the mixed layer into the underlying thermocline (with a continuous match in potential density as depicted in Figure 2), the subduction rate [1] and potential vorticity [3] are kinematically related as in eqn [4]. Q¼

f Db sm rS Dt

½4

where Db rm =Dto0 is the Lagrangian change in mixed-layer potential density, sm , following the velocity ub at the base of the mixed layer, which is usually taken to be a geostrophic streamline. Hence, Q in the thermocline is diagnostically related to the evolution of the mixed-layer density following the horizontal flow and the subduction rate — see Figure 8 for an example of the resulting Q solution for an ideal thermocline model of a subtropical gyre. Alternatively, eqn [4] can be rearranged to highlight how subduction occurs only when the mixedlayer density becomes lighter following a geostrophic

(B)

Figure 9 An eddy-resolving, three-layer model solution for the dynamic tracer, potential vorticity, for a double wind-driven gyre over a rectangular basin: plan views of (A) climatological mean and (B) instantaneous potential vorticity for the intermediate layer. The wind forcing drives a pair of wind gyres antisymmetrically about the middle latitude, which are unstable, generating a vigorous mesoscale eddy circulation. The eddy stirring leads to the time-averaged potential vorticity (in (A)) becoming nearly uniform within the circulating gyres, while outside the gyres there is a poleward increase arising from the meridional gradient in planetary vorticity. The instantaneous potential vorticity (in (B)) shows nearly perfect homogenization within the pair of gyres. At the edges of the gyre, the eddy activity is apparent through the winding up of potential vorticity contours. An experiment by W. Holland. (Reproduced from Rhines PB and Young WR (1982) Homogenization of potential vorticity in planetary gyres. Journal of Fluid Mechanics 122: 347–367, with permission from Cambridge University Press).


OCEAN SUBDUCTION

streamline (eqn [5]).

Buoyancy fluxes

½5 Transformation

where S40 occurs when Db sm =Dto0. This relation reflects how subduction relies on fluid being capped by a lighter mixed layer and hence transferred into the thermocline (Figure 2). For gyre-scale subduction, the required buoyancy input can be provided by an annual average of a surface buoyancy flux or a convergent Ekman flux of buoyancy; the latter contribution dominates in climatological diagnostics over the subtropical gyre of the North Atlantic (as implied from buoyancy diagnostics associated with Figure 5). However, these kinematic relations [4] and [5] cease to be useful in a time-averaged limit in the presence of an active eddy circulation owing to the difficulty in defining an appropriate streamline. The Lagrangian change in mixed layer density, Db sm =Dto0, might even vanish following the timemean geostrophic streamline and provide a misleading result from [5]. Instead, the role of the time-varying circulation in subduction and an integrated view of the buoyancy forcing need to be considered. Integral connection An integral view of the ventilation process may be obtained by considering the conversion of water masses within a basin or restricted domain (Figure 10). Buoyancy fluxes at the sea surface and diffusive fluxes within the ocean lead to water masses being transformed from one density class to another. The convergence of this diapycnal volume flux, or water-mass transformation, defines the rate of water-mass formation within a particular density class. In the limit of no diffusive fluxes within the ocean, the rate of watermass formation rate is equivalent to the areaintegrated subduction rate over the particular density class into the main thermocline (Figure 10); this connection is exploited for the later example of a convective chimney in Figure 13. The contribution of surface buoyancy fluxes to the rate of water-mass formation has been estimated from climatological air–sea fluxes. While there are significant errors in the air–sea fluxes, the diagnostics reveal how the buoyancy forcing and subduction process leads to the preferential formation of distinct mode waters in the North Atlantic (Figure 11), rather than creating similar rates of formation for all densities.

Diffusive fluxes

Control surface

Subduction

Figure 10 A schematic diagram of the upper ocean showing the sea surface, two outcropping isopycnals, and a fixed control surface across which subduction is monitored. Buoyancy fluxes at the sea surface and diffusive fluxes in the interior transform water masses from one density class to another. If there are no diffusive fluxes in the interior, the convergence of the transformation flux controls the subduction flux across the control surface (which is chosen to be the interface between the base of the end of winter mixed layer and the main thermocline). (Reproduced from Marshall J, Jamous D and Nilsson J (1999) Reconciling thermodynamic and dynamic methods of computation of water-mass transformation rates. Deep-Sea Research I 46: 545–572. ^ 1999, with permission from Elsevier Science.)

_

f Db sm rQ Dt

Mass source (106 m3 s 1)

S¼

163

10

STMW

5

SPMW

M

0 _5 26.0

ME 26.4

26.8 27.2 _ Density (kg m 3 )

27.6

28.0

Figure 11 Diagnosed water-mass formation (106 m3 s1) versus density (kg m3) evaluated from climatological surface buoyancy fluxes (full line) and from Ekman pumping (dashed line) over the North Atlantic. The left-hand peak is centered near the density of subtropical mode water (STMW) and the right-hand peak spans subpolar mode water (SPMW) and Labrador Sea Water densities. (Reproduced with permission from Speer K and Tziperman E (1992). Rates of water mass formation in the North Atlantic. Journal of Physical Oceanography 22: 93–104.)

Role of Eddies and Fronts Mesoscale eddies can directly assist in the subduction process through modifying the subduction rate and diffusing the downstream properties of subducted fluid. Eddies lead to an effective stirring of watermass properties along isopycnals. This process can lead to subducted mode waters becoming rapidly smeared out, such that the downstream water-mass properties reflect an average of the upstream


164

OCEAN SUBDUCTION

characteristics determined in the mixed layer at the end of winter. Δz

v

Eddy-induced Transport and Subduction

Eddies provide a volume flux and transport of tracers, which can modify the subduction rate. The eddy-driven transport, u0 Dz0 , is given by the temporal correlation in velocity and thickness of an isopycnal layer, Dz; the overbar represents a time average over several eddy lifetimes and the primes represent a deviation from the time average. An example of an eddy-driven transport is shown in Figure 12A, where the oscillating velocity leads to an eddy-driven transport through the volume flux in the thick ‘blobs’ of fluid being greater than the return flux in the thin ‘blobs’ of fluid. An eddy-induced transport velocity or ‘bolus’ velocity within an isopycnic layer is defined by u*¼ u0 Dz0 /Dz. The eddy ‘bolus’ velocity is particularly large in regions of baroclinic instability, where the transport is generated through the flattening of isopycnals (Figure 12B). Evaluating the instantaneous subduction rate [1] in the presence of eddies is difficult, since timeaveraged correlations in velocity with mixed-layer thickness and velocity with mixed-layer outcrop area are required. However, the subduction rate into the main thermocline and across the control surface, H, can be evaluated, in principle, by including the eddyinduced ‘bolus’ velocity in eqn [2] to give eqn [6]. Sann ¼ ðw ¯ H þ w * Þ ðu¯ H þ u * Þ rH

½6

The vertical eddy-transport contribution, w*, is obtained from u* through continuity. Hence, the annual subduction rate includes contributions from the time-mean and time-varying circulation. The eddy contribution to subduction is expected to be large wherever the bolus velocity is significant, such as in baroclinic zones with strongly sloping isopycnals. Over a subtropical gyre, the subduction contribution from the time-mean circulation probably dominates over the eddy ‘bolus’ contribution, since the wind forcing drives an anticyclonic circulation across contours of mixed-layer thickness. However, the eddy contribution to subduction becomes crucial throughout the Southern Ocean, as well as across separated boundary currents and inter-gyre boundaries. Subduction Associated with Fronts and Convective Chimneys

Subduction and ventilation occurs over fine scales associated with narrow fronts and convective chimneys, as well as over the larger-scale circulation.

(A)

v∗

v∗

z

(B)

y

Figure 12 A schematic diagram illustrating the concept of the eddy rectified transport and ‘bolus’ velocity. (A) The timeaveraged meridional volume flux in an undulating layer, v Dz ¼ vDz ¯ þ v 0 Dz 0 , depends on advection by the time-mean meridional velocity v¯ and the time-averaged correlations in the temporal deviations in velocity v 0 and thickness of an isopycnal layer Dz 0 . The time-averaged volume flux is directed to the right as drawn here, since v > 0 correlates with large layer thickness, while v > 0 correlates with small layer thickness. (B) The slumping of isopycnals in baroclinic instability induces a secondary circulation in which the bolus velocity, v * ¼ v 0 Dz 0 =Dz, is poleward in the upper layer and equatorward in the lower layer. Eddy-driven subduction is likely to occur in baroclinic zones where there are strongly inclined isopycnals. (Reproduced with permission from Lee M-M, Marshall DP and Williams RG (1997) On the eddy transfer of tracers: advective or diffusive? Journal of Marine Research 55, 483–505.)

Secondary circulations develop across a front following the instability of the front and acceleration of the along stream flow. Frontal modeling studies demonstrate how this secondary circulation leads to frontal-scale subduction, which injects low stratification into the thermocline. Observational studies have identified upwelling and downwelling zones on either side of a narrow front (10 km wide) with vertical velocities reaching 40 m d1, which is much greater than the background, wind-driven, downwelling velocity. Indirect support for frontal-scale subduction is also provided by separate observations of bands of short-lived chlorophyll penetrating downward for several hundreds of meters into the stratified thermocline on the horizontal scale of


OCEAN SUBDUCTION

_2

100 W m

_B

in

_ (B + B in eddy)

0 _B

eddy

_ 1.5 Sv _ 0.1 Sv

0

27.8

2500m

0.1 Sv 0.2 Sv 5

27.8

u∗

0.3 Sv 0.4 Sv 0.4 Sv 0.2 Sv

27.90

200km Figure 13 A model solution for the eddy-driven subduction rate and volume fluxes into and out of a convective chimney (shaded) with isopycnals depicted by solid lines. Baroclinic instability of the chimney leads to an eddy-driven circulation. There is entrainment of warm, buoyant fluid over the upper kilometer with subduction of cold, dense fluid at greater depths – the net subduction of deep water is 1.6 Sv (1 Sv106 m3 s1). The subduction rate is determined by the buoyancy forcing, Bin þ Beddy , where Bin is the surface buoyancy loss to the atmosphere and Beddy is the eddy flux of buoyancy within the mixed layer. (Reproduced with permission from Marshall DP (1997).)

10 km; this frontal-scale process is also identified as being important in the atmosphere in transferring ozone-rich, stratospheric air into the troposphere An example of fine-scale subduction associated with a convective chimney is shown in Figure 13. Buoyancy loss to the atmosphere drives the conversion of light to dense water within the mixed layer and the concomitant thickening of the mixed layer. Baroclinic instability of the chimney leads to an eddy-driven transport, which fluxes light fluid into the chimney at the surface and, in turn, subducts denser fluid into the thermocline. Hence, the eddydriven transport enables recently ventilated dense waters to disperse away from a convective chimney. The eddy-driven influx of light fluid partly offsets the buoyancy loss to the atmosphere and can inhibit further mixed-layer deepening over the center of the chimney.

165

coupling between the mixed layer and ocean interior. Fluid is transferred into the permanent thermocline in late winter and early spring, rather than throughout the year. This transfer of fluid into the permanent thermocline helps to determine the relatively long memory of the ocean interior, compared with that of the surface mixed layer. Conversely, the reverse of the subduction process is also important. The induction of thermocline fluid into the seasonal boundary layer affects the downstream water-mass properties of the mixed layer, and hence alters the air–sea interaction and biogeochemistry. For example, biological production is enhanced wherever nutrients are vertically and laterally fluxed from the thermocline into the mixed layer and the euphotic zone. Subduction can occur over a range of scales extending over fronts, meoscale eddies, and gyres. A major challenge is to identify the relative importance of the frontal, eddy, and gyre contributions to subduction. Reliable estimates of subduction rate are difficult to achieve owing to the spatial and temporal variability in the circulation and thickness of the mixed layer and the difficulty in estimating the eddy transport contributions. An additional challenge is to determine the role that subduction plays in climate variability, particularly how subduction communicates atmospheric variability to the ocean interior.

Glossary Ventilation The transfer of fluid from the mixed layer into the ocean interior. Subduction The transfer of fluid from the mixed layer into the stratified thermocline. Subduction rate The volume flux per unit horizontal area passing from the mixed layer into the stratified thermocline. Thermocline A region of enhanced vertical temperature gradient over the upper 1 km of the ocean. Seasonal boundary layer A region over which the mixed layer and seasonal thermocline occur. The base of the seasonal boundary layer is defined by the maximum thickness of the winter mixed layer. Potential vorticity A dynamical tracer, conserved by fluid in adiabatic and inviscid flow, which is determined by the absolute spin of the fluid and the stratification. Sverdrup 1 Sv106 m3 s 1.

Conclusions The subduction process controls the rate at which the upper thermocline is ventilated, as well as determining the water-mass structure and stratification of the upper ocean. Subduction leads to an asymmetrical

See also Deep Convection. Ekman Transport and Pumping. Mesoscale Eddies. Open Ocean Convection.


166

OCEAN SUBDUCTION

Overflows and Cascades. Upper Ocean Mixing Processes. Upper Ocean Vertical Structure. Windand Buoyancy-Forced Upper Ocean. Wind Driven Circulation. Water Types and Water Masses.

Further Reading Cox MD (1985) An eddy-resolving model of the ventilated thermocline. Journal of Physical Oceanography 15: 1312--1324. Cushman-Roisin B (1987) Dynamics of the Oceanic Surface Mixed-Layer. Hawaii Institute of Geophysical Special Publications, Muller P Henderson D (eds), pp. 181–196. Jenkins WJ (1987) 3H and 3He in the Beta triangle: observations of gyre ventilation and oxygen utilization rates. Journal of Physical Oceanography 17: 763--783. Luyten JR, Pedlosky J, and Stommel H (1983) The ventilated thermocline. Journal of Physical Oceanography 13: 292--309. Marshall DP (1997) Subduction of water masses in an eddying ocean. Journal of Marine Research 55: 201--222. Marshall JC, Nurser AJG, and Williams RG (1993) Inferring the subduction rate and period over the North Atlantic. Journal of Physical Oceanography 23: 1315--1329. McDowell S, Rhines PB, and Keffer T (1982) North Atlantic potential vorticity and its relation to the general circulation. Journal of Physical Oceanography 12: 1417--1436.

Pedlosky J (1996) Ocean Circulation Theory. New York: Springer. Pollard RT and Regier LA (1992) Vorticity and vertical circulation at an ocean front. Journal of Physical Oceanography 22: 609--625. Price JF (2001) Subduction. In: Ocean Circulation and Climate: Observing and modelling the Global Ocean, G. Siedler, J. Church and J. Gould (eds), Academic Press, pp. 357–371 Rhines PB and Schopp R (1991) The wind-driven circulation: quasi-geostrophic simulations and theory for nonsymmetric winds. Journal of Physical Oceanography 21: 1438--1469. Samelson RM and Vallis GK (1997) Large-scale circulation with small diapycnal diffusion: the two-thermocline limit. Journal of Marine Research 55: 223--275. Stommel H (1979) Determination of watermass properties of water pumped down from the Ekman layer to the geostrophic flow below. Proceedings of the National Academy of Sciences of the USA 76: 3051--3055. Williams RG (1991) The role of the mixed layer in setting the potential vorticity of the main thermocline. Journal of Physical Oceanography 21: 1803--1814. Williams RG, Spall MA, and Marshall JC (1995) Does Stommel’s mixed-layer ‘Demon’ work? Journal of Physical Oceanography 25: 3089--3102. Woods JD and Barkmann W (1986) A Lagrangian mixed layer model of Atlantic 181C water formation. Nature 319: 574--576.


OCEAN THERMAL ENERGY CONVERSION (OTEC) S. M. Masutani and P. K. Takahashi, University of Hawaii at Manoa, Honolulu, HI, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1993–1999, & 2001, Elsevier Ltd.

Ocean thermal energy conversion (OTEC) generates electricity indirectly from solar energy by harnessing the temperature difference between the sun-warmed surface of tropical oceans and the colder deep waters. A significant fraction of solar radiation incident on the ocean is retained by seawater in tropical regions, resulting in average year-round surface temperatures of about 281C. Deep, cold water, meanwhile, forms at higher latitudes and descends to flow along the seafloor toward the equator. The warm surface layer, which extends to depths of about 100–200 m, is separated from the deep cold water by a thermocline. The temperature difference, DT, between the surface and thousand-meter depth ranges from 10 to 251C, with larger differences occurring in equatorial and tropical waters, as depicted in Figure 1. DT establishes the limits of the performance of OTEC power cycles; the rule-of-thumb is that a differential of about 201C is necessary to sustain viable operation of an OTEC facility. Since OTEC exploits renewable solar energy, recurring costs to generate electrical power are minimal. However, the fixed or capital costs of OTEC systems per kilowatt of generating capacity are very high because large pipelines and heat exchangers are

needed to produce relatively modest amounts of electricity. These high fixed costs dominate the economics of OTEC to the extent that it currently cannot compete with conventional power systems, except in limited niche markets. Considerable effort has been expended over the past two decades to develop OTEC by-products, such as fresh water, air conditioning, and mariculture, that could offset the cost penalty of electricity generation.

State of the Technology OTEC power systems operate as cyclic heat engines. They receive thermal energy through heat transfer from surface sea water warmed by the sun, and transform a portion of this energy to electrical power. The Second Law of Thermodynamics precludes the complete conversion of thermal energy in to electricity. A portion of the heat extracted from the warm sea water must be rejected to a colder thermal sink. The thermal sink employed by OTEC systems is sea water drawn from the ocean depths by means of a submerged pipeline. A steady-state control volume energy analysis yields the result that net electrical power produced by the engine must equal the difference between the rates of heat transfer from the warm surface water and to the cold deep water. The limiting (i.e., maximum) theoretical Carnot energy conversion efficiency of a cyclic heat engine scales with the difference between the temperatures at which these heat transfers occur. For OTEC, this difference is determined by DT and is very small; Longitude

40°E

80°E

120°E

160°E

160°W

120°W

80°W

40°W

0°W

Latitude

40°N 30°N 20°N 10°N Equator 10°S 20°S 30°S 40°S Less than 18°C 18°_ 20°C 20°_ 22°C

22°_ 24°C More than 24°C Depth less than 1000 m

Figure 1 Temperature difference between surface and deep sea water in regions of the world. The darkest areas have the greatest temperature difference and are the best locations for OTEC systems.


167

168

OCEAN THERMAL ENERGY CONVERSION (OTEC)

hence, OTEC efficiency is low. Although viable OTEC systems are characterized by Carnot efficiencies in the range of 6–8%, state-of-the-art combustion steam power cycles, which tap much higher temperature energy sources, are theoretically capable of converting more than 60% of the extracted thermal energy into electricity. The low energy conversion efficiency of OTEC means that more than 90% of the thermal energy extracted from the ocean’s surface is ‘wasted’ and must be rejected to the cold, deep sea water. This necessitates large heat exchangers and seawater flow rates to produce relatively small amounts of electricity. In spite of its inherent inefficiency, OTEC, unlike conventional fossil energy systems, utilizes a renewable resource and poses minimal threat to the environment. In fact, it has been suggested that widespread adoption of OTEC could yield tangible environmental benefits through avenues such as reduction of greenhouse gas CO2 emissions; enhanced uptake of atmospheric CO2 by marine organism populations sustained by the nutrient-rich, deep OTEC sea water; and preservation of corals and hurricane amelioration by limiting temperature rise in the surface ocean through energy extraction and artificial upwelling of deep water. Carnot efficiency applies only to an ideal heat engine. In real power generation systems, irreversibilities will further degrade performance. Given its low theoretical efficiency, successful implementation of OTEC power generation demands careful engineering to minimize irreversibilities. Although OTEC

consumes what is essentially a free resource, poor thermodynamic performance will reduce the quantity of electricity available for sale and, hence, negatively affect the economic feasibility of an OTEC facility. An OTEC heat engine may be configured following designs by J.A. D’Arsonval, the French engineer who first proposed the OTEC concept in 1881, or G. Claude, D’Arsonval’s former student. Their designs are known, respectively, as closed cycle and open cycle OTEC.

Closed Cycle Ocean Thermal Energy Conversion D’Arsonval’s original concept employed a pure working fluid that would evaporate at the temperature of warm sea water. The vapor would subsequently expand and do work before being condensed by the cold sea water. This series of steps would be repeated continuously with the same working fluid, whose flow path and thermodynamic process representation constituted closed loops – hence, the name ‘closed cycle.’ The specific process adopted for closed cycle OTEC is the Rankine, or vapor power, cycle. Figure 2 is a simplified schematic diagram of a closed cycle OTEC system. The principal components are the heat exchangers, turbogenerator, and seawater supply system, which, although not shown, accounts for most of the parasitic power consumption and a significant fraction of the capital expense. Also not included are ancillary devices such as separators to remove residual liquid downstream of the evaporator and subsystems to hold Cold seawater discharge

Warm sea water in

Evaporator

Warm seawater discharge

Working fluid vapor

Condenser

Turbogenerator

Working fluid pressurizer (boiler feed pump)

Cold sea water in

Working fluid condensate Figure 2 Schematic diagram of a closed-cycle OTEC system. The working fluid is vaporized by heat transfer from the warm sea water in the evaporator. The vapor expands through the turbogenerator and is condensed by heat transfer to cold sea water in the condenser. Closed-cycle OTEC power systems, which operate at elevated pressures, require smaller turbines than open-cycle systems.



and supply working fluid lost through leaks or contamination. In this system, heat transfer from warm surface sea water occurs in the evaporator, producing a saturated vapor from the working fluid. Electricity is generated when this gas expands to lower pressure through the turbine. Latent heat is transferred from the vapor to the cold sea water in the condenser and the resulting liquid is pressurized with a pump to repeat the cycle. The success of the Rankine cycle is a consequence of more energy being recovered when the vapor expands through the turbine than is consumed in repressurizing the liquid. In conventional (e.g., combustion) Rankine systems, this yields net electrical power. For OTEC, however, the remaining balance may be reduced substantially by an amount needed to pump large volumes of sea water through the heat exchangers. (One misconception about OTEC is that tremendous energy must be expended to bring cold sea water up from depths approaching 1000 meters. In reality, the natural hydrostatic pressure gradient provides for most of the increase in the gravitational potential energy of a fluid particle moving with the gradient from the ocean depths to the surface.) Irreversibilities in the turbomachinery and heat exchangers reduce cycle efficiency below the Carnot value. Irreversibilities in the heat exchangers occur when energy is transferred over a large temperature difference. It is important, therefore, to select a working fluid that will undergo the desired phase changes at temperatures established by the surface and deep sea water. Insofar as a large number of substances can meet this requirement (because pressures and the pressure ratio across the turbine and pump are design parameters), other factors must be considered in the selection of a working fluid including: cost and availability, compatibility with system materials, toxicity, and environmental hazard. Leading candidate working fluids for closed cycle OTEC applications are ammonia and various fluorocarbon refrigerants. Their primary disadvantage is the environmental hazard posed by leakage; ammonia is toxic in moderate concentrations and certain fluorocarbons have been banned by the Montreal Protocol because they deplete stratospheric ozone. The Kalina, or adjustable proportion fluid mixture (APFM), cycle is a variant of the OTEC closed cycle. Whereas simple closed cycle OTEC systems use a pure working fluid, the Kalina cycle proposes to employ a mixture of ammonia and water with varying proportions at different points in the system. The advantage of a binary mixture is that, at a given pressure, evaporation or condensation occurs over a

169

range of temperatures; a pure fluid, on the other hand, changes phase at constant temperature. This additional degree of freedom allows heat transferrelated irreversibilities in the evaporator and condenser to be reduced. Although it improves efficiency, the Kalina cycle needs additional capital equipment and may impose severe demands on the evaporator and condenser. The efficiency improvement will require some combination of higher heat transfer coefficients, more heat transfer surface area, and increased seawater flow rates. Each has an associated cost or power penalty. Additional analysis and testing are required to confirm whether the Kalina cycle and assorted variations are viable alternatives.

Open Cycle Ocean Thermal Energy Conversion Claude’s concern about the cost and potential biofouling of closed cycle heat exchangers led him to propose using steam generated directly from the warm sea water as the OTEC working fluid. The steps of the Claude, or open, cycle are: (1) flash evaporation of warm sea water in a partial vacuum; (2) expansion of the steam through a turbine to generate power; (3) condensation of the vapor by direct contact heat transfer to cold sea water; and (4) compression and discharge of the condensate and any residual noncondensable gases. Unless fresh water is a desired by-product, open cycle OTEC eliminates the need for surface heat exchangers. The name ‘open cycle’ comes from the fact that the working fluid (steam) is discharged after a single pass and has different initial and final thermodynamic states; hence, the flow path and process are ‘open.’ The essential features of an open cycle OTEC system are presented in Figure 3. The entire system, from evaporator to condenser, operates at partial vacuum, typically at pressures of 1–3% of atmospheric. Initial evacuation of the system and removal of noncondensable gases during operation are performed by the vacuum compressor, which, along with the sea water and discharge pumps, accounts for the bulk of the open cycle OTEC parasitic power consumption. The low system pressures of open cycle OTEC are necessary to induce boiling of the warm sea water. Flash evaporation is accomplished by exposing the sea water to pressures below the saturation pressure corresponding to its temperature. This is usually accomplished by pumping it into an evacuated chamber through spouts designed to maximize heat and mass transfer surface area. Removal of gases


170


Cold seawater discharge

Warm sea water in

De-aeration (Optional)

Vacuum chamber flash evaporator

Noncondensable gases

Desalinated water vapor Turbogenerator

Noncondensable Warm seawater discharge gases

Condenser

Vent compressor Desalinated water (Optional)

Cold sea water in

Figure 3 Schematic diagram of an open-cycle OTEC system. In open-cycle OTEC, warm sea water is used directly as the working fluid. Warm sea water is flash evaporated in a partial vacuum in the evaporator. The vapor expands through the turbine and is condensed with cold sea water. The principal disadvantage of open-cycle OTEC is the low system operating pressures, which necessitate large components to accommodate the high volumetric flow rates of steam.

dissolved in the sea water, which will come out of solution in the low-pressure evaporator and compromise operation, may be performed at an intermediate pressure prior to evaporation. Vapor produced in the flash evaporator is relatively pure steam. The heat of vaporization is extracted from the liquid phase, lowering its temperature and preventing any further boiling. Flash evaporation may be perceived, then, as a transfer of thermal energy from the bulk of the warm sea water of the small fraction of mass that is vaporized. Less than 0.5% of the mass of warm sea water entering the evaporator is converted into steam. The pressure drop across the turbine is established by the cold seawater temperature. At 41C, steam condenses at 813 Pa. The turbine (or turbine diffuser) exit pressure cannot fall below this value. Hence, the maximum turbine pressure drop is only about 3000 Pa, corresponding to about a 3:1 pressure ratio. This will be further reduced to account for other pressure drops along the steam path and differences in the temperatures of the steam and seawater streams needed to facilitate heat transfer in the evaporator and condenser. Condensation of the low-pressure steam leaving the turbine may employ a direct contact condenser (DCC), in which cold sea water is sprayed over the vapor, or a conventional surface condenser that physically separates the coolant and the condensate. DCCs are inexpensive and have good heat transfer characteristics because they lack a solid thermal boundary between the warm and cool fluids. Surface condensers are expensive and more difficult to maintain than DCCs; however, they produce a marketable freshwater by-product. Effluent from the condenser must be discharged to the environment. Liquids are pressurized to ambient levels at the point of release by means of a pump, or,

if the elevation of the condenser is suitably high, can be compressed hydrostatically. As noted previously, noncondensable gases, which include any residual water vapor, dissolved gases that have come out of solution, and air that may have leaked into the system, are removed by the vacuum compressor. Open cycle OTEC eliminates expensive heat exchangers at the cost of low system pressures. Partial vacuum operation has the disadvantage of making the system vulnerable to air in-leakage and promotes the evolution of noncondensable gases dissolved in sea water. Power must ultimately be expended to pressurize and remove these gases. Furthermore, as a consequence of the low steam density, volumetric flow rates are very high per unit of electricity generated. Large components are needed to accommodate these flow rates. In particular, only the largest conventional steam turbine stages have the potential for integration into open cycle OTEC systems of a few megawatts gross generating capacity. It is generally acknowledged that higher capacity plants will require a major turbine development effort. The mist lift and foam lift OTEC systems are variants of the OTEC open cycle. Both employ the sea water directly to produce power. Unlike Claude’s open cycle, lift cycles generate electricity with a hydraulic turbine. The energy expended by the liquid to drive the turbine is recovered from the warm sea water. In the lift process, warm seawater is flash evaporated to produce a two-phase, liquid–vapor mixture – either a mist consisting of liquid droplets suspended in a vapor, or a foam, where vapor bubbles are contained in a continuous liquid phase. The mixture rises, doing work against gravity. Here, the thermal energy of the vapor is expended to increase the potential energy of the fluid. The vapor is then condensed with cold sea water and discharged back into the ocean. Flow of the liquid through the



hydraulic turbine may occur before or after the lift process. Advocates of the mist and foam lift cycles contend that they are cheaper to implement than closed cycle OTEC because they require no expensive heat exchangers, and are superior to the Claude cycle because they utilize a hydraulic turbine rather than a low pressure steam turbine. These claims await verification.

Hybrid Cycle Ocean Thermal Energy Conversion Some marketing studies have suggested that OTEC systems that can provide both electricity and water may be able to penetrate the marketplace more readily than plants dedicated solely to power generation. Hybrid cycle OTEC was conceived as a response to these studies. Hybrid cycles combine the potable water production capabilities of open cycle OTEC with the potential for large electricity generation capacities offered by the closed cycle. Several hybrid cycle variants have been proposed. Typically, as in the Claude cycle, warm surface seawater is flash evaporated in a partial vacuum. This low pressure steam flows into a heat exchanger where it is employed to vaporize a pressurized, lowboiling-point fluid such as ammonia. During this process, most of the steam condenses, yielding desalinated potable water. The ammonia vapor flows through a simple closed-cycle power loop and is condensed using cold sea water. The uncondensed steam and other gases exiting the ammonia evaporator may be further cooled by heat transfer to either the liquid ammonia leaving the ammonia condenser or cold sea water. The noncondensables are then compressed and discharged to the atmosphere. Steam is used as an intermediary heat transfer medium between the warm sea water and the ammonia; consequently, the potential for biofouling in the ammonia evaporator is reduced significantly. Another advantage of the hybrid cycle related to freshwater production is that condensation occurs at significantly higher pressures than in an open cycle OTEC condenser, due to the elimination of the turbine from the steam flow path. This may, in turn, yield some savings in the amount of power consumed to compress and discharge the noncondensable gases from the system. These savings (relative to a simple Claude cycle producing electricity and water), however, are offset by the additional back-work of the closed-cycle ammonia pump. One drawback of the hybrid cycle is that water production and power generation are closely coupled. Changes or problems in either the water or

171

power subsystem will compromise performance of the other. Furthermore, there is a risk that the potable water may be contaminated by an ammonia leak. In response to these concerns, an alternative hybrid cycle has been proposed, comprising decoupled power and water production components. The basis for this concept lies in the fact that warm sea water leaving a closed cycle evaporator is still sufficiently warm, and cold seawater exiting the condenser is sufficiently cold, to sustain an independent freshwater production process. The alternative hybrid cycle consists of a conventional closed-cycle OTEC system that produces electricity and a downstream flash-evaporationbased desalination system. Water production and electricity generation can be adjusted independently, and either can operate should a subsystem fail or require servicing. The primary drawbacks are that the ammonia evaporator uses warm seawater directly and is subject to biofouling; and additional equipment, such as the potable water surface condenser, is required, thus increasing capital expenses.

Environmental Considerations OTEC systems are, for the most part, environmentally benign. Although accidental leakage of closed cycle working fluids can pose a hazard, under normal conditions, the only effluents are the mixed seawater discharges and dissolved gases that come out of solution when sea water is depressurized. Although the quantities of outgassed species may be significant for large OTEC systems, with the exception of carbon dioxide, these species are benign. Carbon dioxide is a greenhouse gas and can impact global climate; however, OTEC systems release one or two orders of magnitude less carbon dioxide than comparable fossil fuel power plants and those emissions may be sequestered easily in the ocean or used to stimulate marine biomass production. OTEC mixed seawater discharges will be at lower temperatures than sea water at the ocean surface. The discharges will also contain high concentrations of nutrients brought up with the deep sea water and may have a different salinity. It is important, therefore, that release back into the ocean is conducted in a manner that minimizes unintended changes to the ocean mixed layer biota and avoids inducing longterm surface temperature anomalies. Analyses of OTEC effluent plumes suggest that discharge at depths of 50–100 m should be sufficient to ensure minimal impact on the ocean environment. Conversely, the nutrient-rich OTEC discharges could be exploited to sustain open-ocean mariculture.


172


Economics of Ocean Thermal Energy Conversion Studies conducted to date on the economic feasibility of OTEC systems suffer from the lack of reliable cost data. Commercialization of the technology is unlikely until a full-scale plant is constructed and operated continuously over an extended period to provide these data on capital and personnel and maintenance expenses. Uncertainties in financial analyses notwithstanding, projections suggest very high first costs for OTEC power system components. Small land-based or near-shore floating plants in the 1–10 MW range, which would probably be constructed in rural island communities, may require expenditures of $10 000– $20 000 (in 1995 US dollars) per kW of installed generating capacity. Although there appears to be favorable economies of scale, larger floating (closed cycle) plants in the 50–100 MW range are still anticipated to cost about $5000 kW 1. This is well in excess of the $1000–$2000 kW 1 of fossil fuel power stations. To enhance the economics of OTEC power stations, various initiatives have been proposed based on marketable OTEC by- or co-products. OTEC proponents believe that the first commercial OTEC plants will be shore-based systems designed for use in developing Pacific island nations, where potable water is in short supply. Many of these sites would be receptive to opportunities for economic growth provided by OTEC-related industries.

been performed for larger metropolitan and resort applications. These studies indicate that air conditioning new developments, such as resort complexes, with cold seawater may be economically attractive even if utility-grid electricity is available. Mariculture

The cold deep ocean waters are rich in nutrients and low in pathogens, and therefore provide an excellent medium for the cultivation of marine organisms. The 322-acre NELHA facility has been the base for successful mariculture research and development enterprises. The site has an array of cold water pipes, originally installed for the early OTEC research, but since used for mariculture. The cold water is applied to cultivate flounder, opihi (limpet; a shellfish delicacy), oysters, lobsters, sea urchins, abalone, kelp, nori (a popular edible seaweed used in sushi), and macro- and microalgae. Although many of these ongoing endeavors are profitable, high-value products such as biopharmaceuticals, biopigments, and pearls will need to be advanced to realize the full potential of the deep water. The cold sea water may have applications for open-ocean mariculture. Artificial upwelling of deep water has been suggested as a method of creating new fisheries and marine biomass plantations. Should development proceed, open-ocean cages can be eliminated and natural feeding would replace expensive feed, with temperature and nutrient differentials being used to keep the fish stock in the kept environment.

Fresh Water

The condensate of the open and hybrid cycle OTEC systems is desalinated water, suitable for human consumption and agricultural uses. Analyses have suggested that first-generation OTEC plants, in the 1–10 MW range, would serve the utility power needs of rural Pacific island communities, with the desalinated water by-product helping to offset the high cost of electricity produced by the system. Refrigeration and Air Conditioning

The cold, deep sea water can be used to maintain cold storage spaces, and to provide air conditioning. The Natural Energy Laboratory of Hawaii Authority (NELHA), which manages the site of Hawaii’s OTEC experiments, has air-conditioned its buildings by passing the cold sea water through heat exchangers. A new deep seawater utilization test facility in Okinawa also employs cold seawater air conditioning. Similar small-scale operations would be viable in other locales. Economic studies have

Agriculture

An idea initially proposed by University of Hawaii researchers involves the use of cold sea water for agriculture. This involves burying an array of cold water pipes in the ground near to the surface to create cool weather growing conditions not found in tropical environments. In addition to cooling the soil, the system also drip irrigates the crop via condensation of moisture in the air on the cold water pipes. Demonstrations have determined that strawberries and other spring crops and flowers can be grown throughout the year in the tropics using this method. Energy Carriers

Although the most common scenario is for OTEC energy to be converted into electricity and delivered directly to consumers, energy storage has been considered as an alternative, particularly in applications involving floating plants moored far offshore.



Storage would also allow the export of OTEC energy to industrialized regions outside of the tropics. Longterm proposals have included the production of hydrogen gas via electrolysis, ammonia synthesis, and the development of shore-based mariculture systems or floating OTEC plant-ships as ocean-going farms. Such farms would cultivate marine biomass, for example, in the form of fast-growing kelp which could be converted thermochemically into fuel and chemical co-products.

See also Carbon Dioxide (CO2) Cycle. Geophysical Heat Flow. Heat and Momentum Fluxes at the Sea Surface. Heat Transport and Climate.

Further Reading Avery WH and Wu C (1994) Renewable Energy from the Ocean: A Guide to OTEC. New York: Oxford University Press.

173

Nihous GC, Syed MA, and Vega LA (1989) Conceptual design of an open-cycle OTEC plant for the production of electricity and fresh water in a Pacific island. Proceedings International Conference on Ocean Energy Recovery. Penney TR and Bharathan D (1987) Power from the sea. Scientific American 256(1): 86--92. Sverdrup HV, Johnson MW, and Fleming PH (1942) The Oceans: Their Physics, Chemistry, and General Biology. New York: Prentice-Hall. Takahashi PK and Trenka A (1996) Ocean Thermal Energy Conversion; UNESCO Energy Engineering Series. Chichester: John Wiley. Takahashi PK, McKinley K, Phillips VD, Magaard L, and Koske P (1993) Marine macrobiotechnology systems. Journal of Marine Biotechnology 1(1): 9--15. Takahashi PK (1996) Project blue revolution. Journal of Energy Engineering 122(3): 114--124. Vega LA and Nihous GC (1994) Design of a 5 MW OTEC pre-commercial plant. Proceedings Oceanology 94: 5. Vega LA (1992) Economics of ocean thermal energy conversion. In: Seymour RJ (ed.) Ocean Energy Recovery: The State of the Art. New York: American Society of Civil Engineers.


OCEAN ZONING M. Macleod, World Wildlife Fund, Washington, DC, USA M. Lynch, University of California Santa Barbara, Santa Barbara, CA, USA P. Hoagland, Woods Hole Oceanographic Institution, Woods Hole, MA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction During the last half-century, as the extent and variety of marine resource uses have expanded with population growth, technological advances, and changing human preferences, ocean space in many areas has become limited and thus more valuable. A multitude of ocean uses compete for available space, including shipping, naval operations, commercial fishing, energy development, waste disposal, marine recreation, oceanography, and environmental conservation. This phenomenon of increased demand for a resource that is in limited supply is most pronounced in coastal marine environments, including bays, estuaries, and near-shore waters. Generally speaking, ocean space becomes more available as one moves further offshore, but the cost of accessing that space becomes prohibitively expensive for all but a few high-valued uses, such as shipping, offshore hydrocarbon production, fishing, and oceanography. The intense competition for ocean space among human users in near-shore waters leads inevitably to costly disputes, accidental or purposeful destruction of property, and, on occasion, armed conflict. In most cases, the private ownership of ocean space is unauthorized, rendering traditional property markets unavailable for mitigating disputes over its use. Ocean space is typically allocated through a variety of mechanisms of collective action, such as government programs and policies. Various agencies at the municipal, state, and national levels employ a wide range of management measures in attempts to regulate the use of ocean space and the exploitation of ocean resources. The academic field of marine policy is organized around analyzing the effectiveness and fairness of alternate management measures for resolving such disputes. The absence of restrictions on uses or the adoption of ineffective management measures can be not only wasteful but may lead to the imposition of social

174

costs, including those associated with the degradation of the marine environment. Ocean zoning is a policy instrument intended to help resolve disputes over the use of ocean space. It involves the designation of areas of the ocean for one or more specific purposes. According to current practice, ocean zoning is a type of command-andcontrol or nonmarket policy instrument. Except in special cases, such as the sale at auction of Outer Continental Shelf (OCS) oil and gas leases, the implementation of ocean zoning does not usually involve the pricing of resources or ocean areas, per se, in order to create incentives for users to undertake socially preferred activities. The essential characteristics of an ocean zone are a boundary and a set of rules that identify the types and regulate the extent of activities that can occur inside the boundary. Ocean zones range in size from expansive exclusive economic zones (EEZs) extending 200 nautical mile (nmi) out from the shorelines of all coastal nations to relatively much more diminutive ‘areas to be avoided’ around a deep-water port (Figure 1 and Table 1). The former may involve millions of square nautical miles whereas the latter may involve less than 10 nmi2. In special circumstances, ocean zones may convey property rights or other legal interests in the ocean or the seabed from the public to private industry or to nongovernmental organizations (NGOs). Such rights may entitle their holders to exclude other human users from the relevant zone. Ocean zones may vary in the degree to which uses are restricted. The Great Barrier Reef Marine Park in Australia, for instance, includes some ocean zones where essentially all human uses are prohibited, other than certain scientific studies. At the other extreme, zones designated as essential fish habitats (EFHs) in the United States are merely a means of classification, aiming to inform fisheries policy but not restricting human uses per se. Depending upon the nature of the activities involved, an ocean zone may be designated exclusively for a specific use, or it may permit multiple uses. In some circumstances, zones can be nested, such as the location of oil and gas lease tracts within a regional 5-year planning area, which is itself located within the EEZ or on the OCS. Further, ocean zones may be of limited duration, such as temporary fishery closures, or they may be designed to exist in perpetuity, such as marine sanctuaries.


OCEAN ZONING

175

Legend Industrial sites Fishing areas Shipping lanes Stellwagen NMS MA ocean sanctuaries Cape Cod Bay Ocean Sanctuary Cape Cod Ocean Sanctuary North Shore Ocean Sanctuary South Essex Ocean Sanctuary

Figure 1 The locations and juxtapositions of ocean zones of various types off the east coast of the New England state of Massachusetts in the western Gulf of Maine.

Most ocean zones are fixed geographically. Some, however, have been designed to move with a resource, such as the buffers surrounding large whales that are the object of whale-watching ecotours. The ocean has

been zoned for national defense purposes, vessel traffic separation, fishery conservation, offshore oil and gas development, marine protection, ocean dumping, scientific research, among many other purposes.


176

OCEAN ZONING

Table 1

Some examples of spatial management schemes

Human use

Ocean zone example

Description

Planning level

Example of spatial scale

General

EEZ and OCS

Zones in which a coastal nation has sovereign rights over the exploitation of living and nonliving natural resources

National, following customary international law and the codified Law of the Sea

General

Submerged lands

Individual coastal states in the USA have ownership of the submerged lands

Local (state level)

Mineral exploration and production

OCS oil and gas leases

Geographic units allocated for hydrocarbon exploration, development, and production

Regional

Commercial fisheries

Fishery reserves

Reservation or closure of areas for purposes of stock recovery; protection of spawning grounds, juvenile fish, or habitat

Regional, local

Protection of species

Critical habitat

National, regional, local

Renewable energy development

Wind park

Protection of endangered or threatened species habitat under the US Endangered Species Act Areas designated for the siting of wind turbines

Shipping

Vessel traffic separation schemes

International, national, regional, local

Marine conservation

Particularly sensitive sea areas

Ocean dumping

Dredged material disposal sites

Designated lanes for inbound/outbound traffic approaching harbors and ports, crossing areas, and vessel types; associated precautionary areas and areas to be avoided Ship operators advised to take extra care within marked boundaries. IMO member states may develop additional regulatory measures (e.g., vessel traffic schemes, noanchoring zones, and discharge regulations) Areas where the disposal of materials dredged from harbors and channels is permitted

Extending 1112 km from the territorial sea baseline (water column and seabed); the continental shelf may exceed this distance under certain geological circumstances Extending 17 km from the territorial sea baseline, with regional exceptions On the OCS of the USA, the typical lease tract is 31 km2, as modified by the characteristics of a hydrocarbon deposit or industry pooling agreements Variable (both spatially and temporally); the largest fishery reserve is a closure to groundfish trawling of 950 000 km2 of waters off the Aleutian Islands, Alaska, USA Variable; 95 182 km2 for North Pacific right whales off Alaska, USA 65 km2 proposed for the Cape Wind facility in Nantucket Sound, off the coast of Massachusetts, USA Varies significantly depending upon the nature of the port and harbor and the roadstead; may be many miles in length

Marine aquaculture

Shellfish leases

Nonpermanent lease of areas for purposes of shellfish (clam, oyster, mussel) cultivation

National, local

International

Ten worldwide, of which the Great Barrier Reef in Australia is the largest at 343 000 km2

International, national, regional, local

Variable: New York/New Jersey disposal site is 54 km2, located 3.5 m off NJ and 7.7 m off NY coast Fractions of km2 (variable)

Local


OCEAN ZONING

Comprehensive Ocean Zoning The term ‘ocean zoning’ also has been used recently to refer to the application of land-based zoning principles to the ocean and in this context has become part of the lexicon of ocean management. This vision of ocean zoning, which often is a key implementation tool for so-called ‘ocean planning’ or ‘comprehensive ocean management’ schemes, may involve the designation of areas for particular uses before specific projects are proposed. Under some ocean planning schemes, the entire body of water in question is zoned; only certain regions are zoned in others; and many other ocean areas remain loosely regulated or largely open access. Potentially competing uses may be separated while compatible uses may be clustered, based on economic, ecological, or infrastructural criteria. Proponents of ocean zoning see the expansion of the concept as a moving away from what they perceive to be an ad hoc, fragmented, and uncoordinated regulatory environment and toward a more effective method for setting down a sensible mosaic of ocean uses. Detractors argue that such schemes could be unfair to historic users of the ocean and would be unable to address uncertain and changing future use patterns. A key element of any debate about ocean zoning concerns the ease with which zoning decisions can be modified as environmental conditions change or as human preferences shift. Opponents also claim that ocean zoning could prove to be an inefficient and expensive form of regulation, disproportionately subject to the influence of special interest groups and plagued by enforcement difficulties. The examples in Table 1 reflect the wide range of scales of implementation, spatial and temporal extents, and regulatory details that characterize spatial ocean management. EEZs or continental shelves comprise broad jurisdictions, establishing the authority of nations to regulate activities within their limits. Other more parochial zones, such as fisheries closures and shipping lanes, are examples of historically unregulated transient ocean uses; zoning for these uses perpetuates the public character of the ocean. Still others, such as offshore oil and gas lease tracts on the US OCS, wind farms off the coast of Denmark, or aquaculture development areas in Hawaiian territorial waters, are examples of zones for nontransient ocean uses in which private industry can occupy ocean areas or exploit ocean resources. Fisheries closures are unique among the examples in Table 1 in that they may be temporary, instituted in response to a narrowly defined management need. When the northwest Atlantic cod fishery in the Gulf

177

of Maine collapsed in the early 1990s, a series of emergency closures were established in order to allow stocks to recover from severe levels of recruitment overfishing. Such emergency closures can last from several months to several years. Spatial fisheries closures also can be utilized as part of a long-lasting management plan, either in the form of permanent no-take reserves or seasonal closures of breeding, feeding, or hatching grounds. Indeed, one set of groundfish reserves along the western boundary of the Gulf of Maine involves a wave of sequential closing and opening of areas during the fishing season to protect spawning grounds and juvenile fish. Fisheries closures further illustrate the varying degrees and methods by which uses can be restricted in an area. Nautical charts of Georges Bank, a highly productive fishing area off the coast of New England, include both strictly enforced do-not-enter closures alongside areas that are closed only to certain types of boats or gear. Closures also might apply to only a portion of the water column (e.g., from 1000 to 2000 m) or, when selective gear can be deployed, to particular species.

The Challenges of Ocean Management The uses of many ocean resources provide quintessential examples of Professor Hardin’s tragedy of open access, by which individual resource users, acting in their own self-interest, undermine collective long-term sustainability. Unfortunately, most extant public policies render traditional solutions, such as privatization and mutual agreements, at best challenging and at worst illegal. A thoughtful literature on the theory and practice of commons management has recognized certain resource and user characteristics, the existence of which increase the likelihood that unregulated ocean areas may become a true commons, to be managed collectively for the public good. These characteristics comprise a small zone, within the bounds of which a moderate level of resource use is undertaken, involving insignificant external effects. Further, resource users should share common cultural beliefs and social norms, exhibit an instinctive understanding of resource dynamics, and face nominal costs of monitoring resource stocks and flows. Regrettably, in only a few circumstances do these characteristics describe the ocean and its modern users. The reality of ocean management is characterized by complexity, inchoate science, and steep costs of monitoring and enforcement. The ocean is a fluid


178

OCEAN ZONING

environment, and its many living resources exhibit wide-ranging and incompletely understood migratory patterns. Resource users may not fully understand the cause-and-effect relationships that relate their actions to the responses of either a targeted resource or the larger ecosystem. In many places, the science of marine resource dynamics is incomplete or even unstudied, rendering monitoring and assessment, even when technologically and economically possible, uncertain and inconclusive. Further, ocean habitats are dynamic, sometimes chaotically so, implying that the sampled environmental characteristics used to designate a zone at one point in time could become immaterial at a later date. These unique characteristics of the ocean may preclude the use of spatial management schemes in some contexts or may require more careful (and costly) ongoing attention to the dimensional aspects of the ocean, such as current flows, temperature and salinity fluxes, plankton distributions, migration routes, species interactions, among many other variables.

Appropriate Scales for Ocean Zoning The physical and ecological features of the marine environment, in combination with unique institutional regimes, make the choice and application of management scales more problematic than on land. Comprehensive ocean planning involving zoning has received increased attention of late from academics and policymakers because it offers the hope that disputes over ocean uses may be mitigated by physically separating potentially competing uses. This approach assumes that there are appropriate spatial scales for particular uses, and that these scales, which may differ drastically across uses or for different resources, can be implemented successfully in practice. One of the most prevalent uses of ocean zoning has been the designation of various types of marine protected areas (MPAs). According to a definition adopted by the International Union for the Conservation of Nature: ‘‘[an MPA is] an area of intertidal or subtidal terrain, together with its overlying water and associated flora, fauna, historical and cultural features, which has been reserved by law or other effective means to protect part or all of the enclosed environment.’’ Importantly, many MPAs are implemented with multiple objectives, so that human uses are allowed to continue at some level, while natural features and processes are conserved. The most restrictive of MPAs are the ‘no-take’ fishery reserves, within which no fishing or other extractive uses are permitted. Fishery reserves are

implemented commonly for one of two purposes: (1) fisheries enhancement, for which effectiveness is not always apparent; or (2) biodiversity conservation. While reserves have a mixed record with regards to fisheries enhancement, they have repeatedly been shown to be an effective means of preserving biodiversity within their bounds. Most marine ecologists agree that the success of spatial conservation depends strongly upon reserve size and location. There is a growing awareness that optimal design depends upon many factors, including regional oceanographic characteristics, biological variables, such as patterns of larval dispersal, and the management goals of a particular reserve. Yet there remain many unanswered questions about the design of marine reserves and whether reserves should be linked together into a regional network. The optimal scale for reserves can vary by several orders of magnitude, and the need for and design of reserve networks further factors into considerations of scale. Linear programming approaches have received much recent attention as methods for determining the optimal spatial distribution of a group of protected or regulated areas within a region. These approaches involve the use of methods from the field of operations research to minimize the total area or boundary length of an MPA subject to user-defined conservation constraints. The effectiveness of such spatial modeling approaches depends critically upon the quality of the underlying data and the relationship of the data to the conservation goals. For example, one such application involves the use of mapped geological characteristics of the seafloor as a means for identifying which groundfish spawning grounds should be protected in a fishery. The effectiveness of this application obviously depends on the correlation between the relevant seafloor characteristics and fish-spawning behavior. Programming approaches also have been used to try to minimize the economic impact of the displacement of fishermen from protected areas. Of note, these programs tend to model a static situation, characterized by unchanging environmental parameters and historical patterns of fishing. Most applications to date have been limited by the lack of complementary models that predict the dynamic behavioral responses of the displaced fishermen. Notwithstanding these problems, the design of spatial progams for optimizing ocean management schemes continues to be a very active area of interdisciplinary research. In the future, other models and programs, some using emergent geographic information system capabilities, may be drawn upon to help design comprehensive, multiple-use ocean-zoning schemes.


OCEAN ZONING

A scientific consensus apparently is coalescing around the idea that biological connectivity over distance may be best preserved through networks of MPAs. Thus, networks arguably provide greater conservation benefits than isolated individual reserves. Consequently, the implementation of ocean zoning within regional planning units may make it easier for natural resource managers to coordinate resource protection, enhance fishery yields, and conserve biodiversity on the scale of ecosystems. Because of the multiplicity of jurisdictions and the large numbers of stakeholders, however, attempts at ocean zoning at a regional level often face daunting political obstacles.

Economic and Distributional Implications of Ocean Zoning Zoning occurs as a consequence of the public perception of inefficient or unfair distributions of existing human uses of the ocean. For example, the overexploitation that occurs in an open-access fishery may lead to the dissipation of economic rents associated with targeted fish stocks. Changes in ecological states that accompany overexploitation also may adversely affect commercial fishermen targeting other fisheries, recreational fishermen, and ecotourists who like to watch whales. The establishment of MPAs, fishery closures, EFHs, and other forms of zoning is one reaction to the wasteful use of the ocean as an open-access fishery. The delimitation of a zone is inherently an economic decision that reveals societal preferences for particular forms of the production of goods and services from the ocean. When a particular use or combination of uses of ocean space is perceived to be more appropriate than a historical pattern of uses, a zone may be created to modify the ‘production technology’. Thus a wind farm might be sited in an area that fishermen have historically used to trawl for fish. The term ‘ocean use’ is defined broadly here to include the production of goods and services that trade in existing markets (fish, transportation services, offshore oil and natural gas recovery, and sand and gravel mining) as well as the enjoyment of those that have no markets (natural habitat, biological diversity, esthetic views, and disposal of dredge spoils). Among the social sciences, only neoclassical economics provides a normative framework for judging whether or not the production technology represented by the allowed uses and nonuses of an ocean zone is socially optimal (i.e., economically efficient). Following that framework, ocean zones ought to be established to allow those combinations of uses that

179

maximize net present value as measured in monetary terms. In order for a combination of ocean uses to be considered economically efficient, the benefits of that combination of uses must exceed the opportunity costs of all other potentially excluded uses, either alone or in feasible combinations. Thus, in principle, there is an economically optimal mix of uses in an ocean area. This simple characterization of a decision rule for deciding upon efficient zoning in the ocean belies many complexities. Foremost among these is the problem of estimating the value of patterns of alternative uses and nonuses, such as habitat preservation. The net economic surpluses from alternative uses or compatible combinations of uses must be estimated, forecast into the future, and discounted into present values. These then would be compared with similar estimates of potentially excluded uses. Where nonuses are involved, specialized methodologies must be applied to estimate nonmarket or ‘passive use’ values. All valuation approaches also may involve uncertainty and risk aversion, which can make the decision process more complex. To circumvent these issues, some jurisdictions invoke legal decision rules, such as the so-called ‘public trust doctrine’, that establish priorities for certain uses over others. For example, navigation and fishing, which tend to be transitory uses of the ocean, usually are regarded as priority public trust uses applied at the individual state level in the United States in decisions over the allocation of uses in coastal waters and tidelands. Although the public trust doctrine may simplify decision making, and it may appear to be fair to some user groups, it does not necessarily imply that ocean-use decisions are economically efficient. The economic criterion tends to ignore the distributional effects of zoning decisions. For example, if an area of the ocean is zoned as a deep-water port, commercial fishermen may be displaced from the area. It may be economically efficient to establish such a zone, particularly when it displaces an open-access fishery in which resource rents have been dissipated, but the displaced fishermen often will complain about the inequitable result, particularly if they have been fishing in the zoned location for many years or even for generations. In the absence of well-defined property rights in ocean space or legal interests in ocean uses, it can be difficult for historical users to demand compensation for the loss of access to a newly zoned area. Notably, where such rights exist, users have a stronger argument for compensation. For example, fishermen in New Zealand who own transferable quota rights for wild fisheries have argued for compensation for areas of the fishery that have been zoned for marine aquaculture.


180

OCEAN ZONING

Institutional Approaches to Zoning The allocation of ocean space for specific uses through zoning can be implemented by executive fiat, legislative decree, or mutual agreement arrived at among a set of stakeholders acting independent of government. In some cases, ocean zones may be established through a joint effort of government agencies and stakeholders, a method of collective decision-making referred to by some commentators as ‘co-management’. Good examples exist of many of these means of implementation, and analyzing their pros and cons is an area of active social science research. Recently, ocean zoning by centralized management institutions, which exercise authority by fiat or decree, has fallen into disfavor. Comanagement or stakeholder agreements have received increasing attention as socially preferred forms of collective decision making, leading to gains in fairness and possibly even efficiency in the form of lower administrative costs. Although designation by executive fiat is one of the oldest forms of ocean zoning, for example, the establishment in 1935 by US President Franklin D. Roosevelt of the Fort Jefferson National Monument in the Dry Tortugas, it has been used less frequently in developed nations in recent decades. An exception concerns the recent designation in June 2006 of the Papha¯naumokua¯kea Marine National Monument by US President George W. Bush. This 137 792 km2 MPA, the largest MPA in the world at the time of this writing, is located in a remote region of small islands and atolls in the northwestern Hawaiian Islands. The national monument was established to phase out commercial fishing, ban seabed resource extraction, prohibit ocean dumping, preserve access for native Hawaiian cultural activities, enhance visitation around Midway Island, and provide for education and scientific research. The designation by executive fiat was justified in part on the need to forestall a comanagement process that was perceived to be contentious and needlessly slow. The Canada Oceans Act of 1997 provides an example of a co-management approach to ocean zoning. The act defines ‘integrated management’ as a collaboration that brings together all interested parties to incorporate social, cultural, environmental, and economic values in developing and implementing spatially explicit ‘ocean-use plans’. Under the auspices of the act, a pilot planning project has been organized for the integrated management of 325 000 km2 of eastern Canada’s Scotia Shelf. Human uses of the Scotian Shelf comprise commercial fishing, oil and natural gas exploration and production, national defense, laying of submarine

cables, scientific research, and recreation and tourism. Triggered by an oil and gas development proposal in a sensitive underwater canyon habitat, the Eastern Scotian Shelf Integrated Management (ESSIM) ocean-use plan has been touted as an ecosystem-based approach that focuses on defining management goals that will attempt to balance commercial uses with environmental protection of the area. The Canadian Department of Fisheries and Oceans (DFO) has been developing the plan with industry, conservation organizations, academics, First Nations, communities, and all levels of government through a multitiered stakeholder process.

Conclusions As the variety and scale of human uses of the coastal ocean increase, ocean space becomes more scarce and therefore more valuable in an economic sense. Ocean zoning is a nonmarket policy instrument for allocating scarce ocean space in order to mitigate real and potential conflicts over its use. The imposition of restrictions on particular uses in certain areas of the ocean is inevitably controversial, leading to the need for the selection of political institutions for implementing ocean zoning. Recently, co-managementtype institutions, involving agreements leading to zoning between government and stakeholders, have been in vogue, displacing zoning allocations by executive fiat or legislative decree. If economic efficiency were to become an overriding policy goal, the development of a market in ocean space would be an appropriate institution. Such an institution might involve the identification of certain areas that would be set beyond the reach of the market for conservation purposes. Increasingly, we expect to see the growing use of ocean zoning to carry out the goals of comprehensive, regional ocean management. As zoning is implemented, as legal interests begin to attach to the uses of zoned ocean space, and as technologies reduce the costs of establishing and enforcing these interests, markets for allocating ocean space will almost certainly emerge.

See also Law of the Sea. Marine Policy Overview. Marine Protected Areas.

Further Reading Crowder LB, Osherenko G, Young OR, et al. (2006) Resolving mismatches in US ocean governance. Science 313: 617--618.


OCEAN ZONING

Day JC (2002) Zoning – lessons from the Great Barrier Reef Marine Park. Ocean and Coastal Management 45: 139--156. Dietz T, Ostrom E, and Stern PC (2003) The struggle to govern the commons. Science 302: 1907--1912. Eckert RD (1979) The Enclosure of Ocean Resources: Economics and the Law of the Sea. Stanford, CA: Hoover Institution Press, Stanford University. Halpern BS (2003) The impact of marine reserves: Do reserves work and does reserve size matter? Ecological Applications 13: S117--S137. Hanna SS, Folke C, and Ma¨ler K-G (eds.) (1996) Rights to Nature: Ecological, Economic, Cultural, and Political Principles of Institutions for the Environment. Washington, DC: Island Press. Hardin G (1968) The tragedy of the commons. Science 162: 1243--1248. Johnston DM (1993) Vulnerable coastal and marine areas: A framework for the planning of environmental security zones in the ocean. Ocean Development and International Law 24: 63--79. Libecap GD (1989) Contracting for Property Rights. New York: Cambridge University Press.

181

McGrath K (2003) The feasibility of using zoning to reduce conflicts in the exclusive economic zone. Buffalo Environmental Law Journal 11: 183--220. Russ GR and Zeller DC (2003) From mare liberum to mare reservum. Marine Policy 27: 75--78.

Relevant Websites http://www.gbrmpa.gov.au – Great Barrier Reef Marine Park Authority. http://www.imo.org – International Maritime Organization. http://mpa.gov – Marine Protected Areas of the United States. http://www.un.org – United Nations, Office of Legal Affairs, Division for Ocean Affairs and the Law of the Sea. http://www.oceancommission.gov – US Commission on Ocean Policy. http://www.mms.gov – US Department of the Interior, Minerals Management Service, Offshore Minerals Management.


OFFSHORE SAND AND GRAVEL MINING E. Garel, CIACOMAR, Algarve University, Faro, Portugal W. Bonne, Federal Public Service Health, Food Chain Safety and Environment, Brussels, Belgium M. B. Collins, National Oceanography Centre, Southampton, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction Offshore sand and gravel extraction involves the abstraction of sediments from a bed which is always covered with seawater. This activity started in the early twentieth century (in the mid-1920s, in the UK), but did not reach a significant scale until the 1960s and 1970s, when markets for marine sand and gravel expanded and dredging technology improved (Figure 1). In the mining industry, the term ‘aggregates’ describes a variety of diverse particulate materials, which are provided in bulk, that is, sand and gravel, together with crushed rocks. The distinction

between sand and gravel varies, on the basis of the classification adopted. Hence, the nomenclature may change between the end-users or countries (or even regions within a particular country). For example, in the UK industry, ‘sand’ refers to (noncohesive) minerals with a mean grain size lying between 0.63 and 5 mm, and ‘gravel’ is reserved for coarser material; in Belgium, the industry considers as gravel the sediment with a mean grain size larger than 4 mm. For comparison, within the scientific literature, the limit between sand and gravel is generally at 2 mm.

Resource Origin Marine aggregate deposits from the continental shelf can be either relict or modern. Relict deposits have been formed under hydrodynamic and sedimentological regimes that no longer exist. They consist of river or coastal bank deposits, formed during the lower seawater stands, induced by the glacial periods affecting the Earth over the past 2 My. At that time, the rivers extended widely across the continental

140

120

Material extracted (million tonnes)

Land-won + marine aggregates 100

80

60

40 Marine aggregates only 20

0 1920

1940

1960 Year

1980

2000

Figure 1 Production of aggregates in the UK between 1900 and 2004 (excluding fill contract and beach nourishment, for marine aggregates). Note the rapid growth of extraction during the late 1960s and the early 1970s, due to the boom in construction in southern England. Sources: Crown-Estate, British Geological Survey.

182


OFFSHORE SAND AND GRAVEL MINING

Volume of sand extracted (106 m3)

shelf, transporting and depositing large amounts of sand and gravel in their valleys and coastal waters. This material has been submerged and possibly reworked, especially the mobile sandy deposits, during subsequent warmer interglacial periods, due to the retreating and melting of the ice. Concentration of deposits has resulted from the numerous repetitions of this warm–cold cycle. These sedimentary bodies have remained undisturbed since their formation and are immobile within the context of prevailing hydrodynamic conditions. Relict sand and gravel deposits include mostly thin sheets, banks, drowned beaches and spits, and infilled channel systems, such as river courses and terraces. In contrast, modern deposits have been formed and are controlled by the present prevailing hydrodynamic and sedimentological regime. Therefore, they are related to presentday coastal sediment budgets and sediment dynamics on the seabed. These deposits include, exclusively, sandbanks and sand bedform fields. The composition, roughness, and grain size of marine aggregate particles vary considerably, depending upon their mode of formation, depositional processes, and history of reworking. Although dependent upon the intensity of post-erosional transport, the grains consist usually of fragments derived from rocks of the surrounding emerged regions. Due to their glacial origin, relict deposits have a composition which is similar to those quarried on land (which are their upstream equivalent), having a low content of shell fragments. In comparison, modern deposits may include high concentration of shelly sediments, which affect the commercial quality of the deposit. In addition, marine aggregates tend to have an higher chloride content. As such, some objections have been raised on their use, in relation to potential alkali–silica

183

reactions, which could affect concrete, or chloride attack on steel reinforcement. However, experience and some experiments have shown that marine aggregates are as suitable as on-land quarried aggregates, for construction purposes.

Production and Usage Marine aggregates are used mostly in the construction industry, but also for beach replenishment and shore protection, for land reclamation, and in other fill-related uses (such as drainage and capping material). The quality required is governed generally by the usage of the product. High-quality marine aggregates (i.e., well sorted and free from impurities such as clay) are used for mixing with a cementing agent, to produce hard building material (e.g., concrete, mortar, and plaster), or are coated with bitumen for road surfacing. For other usages, such as base material under foundations, roads, and railways, or as the construction of embankments, lowergrade materials can be used. The quality of the material used for beach nourishment may also show strong variability, with location and local/regional policies. Nonetheless, not only should the grain size of the nourished material be similar to (or greater than) that eroded, but it should also not contain fresh biogenic material or contaminants. Production and usage of marine aggregates vary considerably between countries, with the largest global producers being Japan, the Netherlands, Hong Kong, Korea, and the UK (Figure 2). Clusters of extraction areas lie in the vicinity of these countries (e.g., in the North Sea, the Channel, and the Baltic Sea), although aggregate extraction takes place

50 40

39 32.3

32

30 20

15

12.8

10 0

7.2

Jap.∗

Neth.

HK †

S. Kor.‡

UK

US

5.8 Ger.§

5.6

Den.

2.4

1.6

Fra.

Bel.

Figure 2 Overview of sand extracted from coastal waters around the world (2002). Annual volumes in million cubic meters per year, for different countries (Japan, the Netherlands, Hong Kong, Korea, United Kingdom, United States, Germany, Denmark, France, Belgium). Key: data from 1998; w Hong Kong, averaged over 1990–98; z averaged over 1993–95; and y from 2000. Source: Roos PC (2004) Seabed Pattern Dynamics and Offshore Sand Extraction, 166pp. PhD Thesis, University of Twente, The Netherlands (ISBN 90-365-2067-3).


184


in shallow continental seas all over the world. The largest producers present several – if not all – of the following characteristics: limited, or depleted, onland aggregate resources, increasing environmental pressure on quarries, and large consumption of material. Besides, the yearly national productions may rise drastically as a result of major public works involving, temporarily, a huge volume of extracted material (e.g., Belgium, in 1997, for the construction of pipelines; and the Netherlands, in 2001, for major land reclamation schemes). Countries producing large quantities of marine aggregates extract different types of material, and for distinct purposes (e.g., concrete, filling, or beach recharge). In addition, they may import and export material to and from other countries. For example, in 1998, almost 90% of the marine aggregates production in the UK was for concrete (gravel and coarse sand), but a third of this (c. 7 Mt) was exported to other northwestern European countries. In Spain, by contrast, the dredging of marine aggregates is authorized only for beach replenishment.

Prospecting The identification of marine aggregate resources is based upon both desk studies and offshore prospecting surveys. The purpose of the desk study is to

locate suitable deposits, based upon a scientific literature review and other considerations. Prospecting surveys involve geophysical data acquisition and sediment sampling. In addition to an accurate vessel positioning system (e.g., GPS), the geophysical instrumentation includes usually: an echo sounder, to investigate the seafloor bathymetry; a high-resolution seismic system (e.g., Boomer, Pinger, and Sparker), to produce reflection profiles of the first (c. 10) few meters of the seabed layers; and, a side scan sonar, to characterize the seabed hardness and roughness (Figure 3). The collected data set provides information about the thickness, horizontal distribution, and surficial character of the deposit. Vertical and horizontal ground-truth information is provided by core and grab samples, respectively. In addition, grab sampling is useful to examine the surficial sediment grain-size distribution (and also to identify the benthic biological communities present within the area). The selection of the sampling equipment depends upon the difficulty of penetration into the various sediment layers. Large hydraulic grabs can be used to obtain surficial sediment samples, while vibrocorers (up to 10-m long) are capable of coring and recovering coarse-grained material. The level of accuracy of deposit recognition, dependent upon the resolution of both the instruments and the sampling spatial coverage, is governed by the cost of the surveys.

Seismic profile

Navigation antenna

Time Surface-towed seismic source Distance Seismic receivers (hydrophone streamer)

Research vessel

0 2−6 kt (in general)

Echo sounder antenna (multibeam)

Seismic wave ray paths

10 Schematic water depth (m) 20

Weight

Side scan sonar

Bathymetric map

30 0

20

40

60 80 100 Relative horizontal distance (m)

Seabed acoustic image

120

140

Figure 3 Schematic representation of a survey vessel undertaking geophysical prospecting operations: (multibeam) echo sounder mapping; side scan sonar imaging; and seismic reflection survey.



Extraction Process Offshore sand and gravel extraction is performed by anchor (or static) suction dredging, and trailer suction dredging. Both techniques utilize powerful centrifugal pumps to draw up the seabed material into the hopper dredger, through pipes of up to 1 m diameter. The dredged material displaces seawater within the hold of the ship, loaded previously as ballast. Anchor suction dredging involves a vessel anchoring or remaining stationary over a deposit, together with forward suction of the bed material through the pipe. As such, the technique is effective for small spatially restricted or, locally, thick deposits. This abstraction procedure creates crater-like pits, or a series of pits, up to 25 m deep and 200 m

185

in diameter, with a slope of about 51 (Figure 4). In contrast, trailer suction dredging requires the dredger to drag the lower end of its rear-facing pipe(s) slowly along the seabed, while the ship is underway. This technique permits relatively large areas with thin and more evenly distributed deposits to be worked. It is used also for thicker deposits, such as sandbanks, to limit the harmful environmental impacts to the superficial layers. At the bottom, the head of the pipe creates linear furrows, of 1–3 m in width and up to 50 cm in depth (Figure 5). However, repeated trailer suction hopper dredging over an area can lead to large dredged depressions (Figure 6). The total dredged cargo contains generally a mix of different grain sizes. In some cases, the dredgers

(a)

Multibeam Water depth (m)

0

20

40

60

80

100 m

(b)

20 m

Side scan sonar

Figure 4 Examples of gravel pits created by anchor hopper dredging (Tromper Wiek Area, Baltic Sea): (a) bathymetry draped with backscatter (illuminated from the west), obtained from a multibeam system; (b) side scan sonar. Source: Manso F, Diesing M, Schwarzer K, and Garel E (2004) Monitoring of dredged areas by combination of multibeam echosounder and sidescan sonar data. Conference Littoral 2004, Aberdeen, UK, 20–22 September 2004.


186


(a)

Furrows

100 m

Survey line

(b)

Survey line

Furrows

100 m

Figure 5 Side scan sonar mosaics showing sandy furrows, in two adjacent areas, at Graal Mu¨ritz, Baltic Sea, in December 2000 (with the survey lines in a NE–SW direction): (a) furrows are observed in the center of the image, with a subequatorial direction (note the reverse in direction of the operating dredger at the western tip of the tracks); (b) the furrows are oriented more randomly, although predominantly NE–SW in the lower part of the image. From Diesing M (2003) Die Regeneration von Materialentnahmestellen in der su¨dwestlichen Ostsee unter besonderer Beru¨cksichtigung der rezenten Sedimentdynamik, 158pp. PhD Thesis, Christian-AlbrechtsUniversita¨t, Kiel.

retain all the dredged material on board, for discharge and processing ashore. However, dredging operations often target only a specific type of aggregate, defined by market demand. Since it is more cost-effective to load and transport only a cargo of the required type, dredgers often have the facility to screen onboard the dredged material, spilling back to the sea the unwanted fines. This screening process spills an important amount of material, up to 3–4 times the retained cargo load, generating sediment

plumes that settle down and spread over the extraction zone (making subsequent extractions sometimes more difficult). Nowadays, some dredgers with cargo capacities of up to 8500 t operate with pumps capable of drawing aggregates at 2600 t of material per hour, in water depths up to 50 m (even 90 m, for Japanese vessels). Discharging on the wharf is performed by a range of machinery, such as bucket wheels, scrapers, or wirehoisted grabs, which place the aggregates on a



(a)

187

(b) Water level below chart datum (m)

15

km

N

–5

NW

SE

Dredged depression

–10 –15 –20 –25 500

1000 1500 Distance (m)

2000

Figure 6 Example of depressions created by repeated trailer suction hopper dredging on a sandbank: the Kwintebank, offshore Belgium. (a) Bathymetric map of the Kwintebank, where the depression is located with the red dotted ellipse; the red straight line indicates the location of the cross section (b); the bank is 15 km in length, 1–2 km in width, and 10–20 m in height. (b) Cross section perpendicular to the bank long-axis, including the dredged depression, 300 m in width, 1 km in length, and up to 5 m deep. (a) Source: Belgian Fund for Sand Extraction.

conveyor belt. Hydraulic discharge is also used, mainly for beach-nourishment schemes.

Environmental Impacts The environmental effects of sand and gravel extraction include physical effects, such as the modification of the sea bottom topography, the creation of turbidity plumes, substrate alteration, changes in the local wave and current patterns, and impacts on the coast. Biological effects include changes in the density, diversity, biomass, and community structure of the benthos or fish populations, as a consequence of the physical effects. One dredging activity may not have a significant direct or indirect environmental impact, but the cumulative effect of several (adjacent) dredging sites may induce significant changes. Physical Impacts

The estimation of the regeneration time of dredged features is still very difficult and depends upon the extracted material, the geometry of the excavation, the water depth, and the hydrodynamic regime of the system. Typical timescales for the regeneration of dredged furrows in sandy dynamic substrates lie within the range of months. In very energetic shallow sandy areas, such as those found in estuaries, they may recover after just one (or a few) tidal cycle. In contrast,

in deep and low-energy areas of fine sand (e.g., the German Baltic Sea), recovery could take decades, even for small features such as furrows. In waters of the Netherlands, southern North Sea, dredged tracks in a gravelly substrate exposed to long-period waves were found to disappear completely within 8 months, at 38-m water depth. In contrast, in waters of SE England, dredged tracks in gravelly sediments can take from 3 to more than 7 years to recover. Dredged depressions or pits created by static dredging have also been reported to remain, as recognizable seabed features, for several years and up to decades. In some cases, pits have been observed to migrate slowly in the direction of the dominant current. Surface turbid plumes are generated by the screening process and by the overflow of material from the hopper, during dredging. A further source of turbidity, with a far lesser quantity of suspended material, results from the mechanical disturbance of the bed sediment, by the head of the pipe. Large increases in suspended solid concentrations tend to be short-lived and localized, close to the operating dredger. However, turbid plumes with low suspended sediment concentrations can affect much larger areas of the seabed, over extended time periods (several days instead of several hours), especially when dredging is occurring simultaneously in adjacent extraction areas. Changes in sediment composition, as a direct result of dredging, range from minor alterations of


188


superficial granulometry to a large increase in the proportion of fines (sand or silt), or to an increase of gravel through the exposure of coarser sediments. In addition, dredged depressions or furrows, in gravel and coarse sand, can trap gravelly sand remobilized by storms, or finer sediment mobilized by tidal currents. This process may reduce significantly the volume of sediment bypassing in an area. In the case of relatively steep and deep dredged trough, a significant increase in silt and clay material, at the bottom, is sometimes observed following dredging, with long-term consequences (i.e., alteration) upon the in situ infaunal assemblage. Local hydrodynamic changes (i.e., in wave heights and current patterns) may result from the modification of the bathymetry by dredging. In turn, the local erosional and depositional patterns may be affected (near-field effects). In general, the dimensions of dredged features are, in relative terms, so small, that there is little influence of the deepened area on the macroscale hydrodynamics. However, local hydrodynamic changes induced by dredging may result in wider environmental effects (far-field effects). The main issue concerns possible erosion at the coastline, according to various scenarios. Beach erosion may occur, as a result of offshore dredging, due to a reduction of the sediment supply to the coast. This effect is induced either by the direct extraction of material which would normally supply the coast, or by trapping these sediments within the dredged depressions. Any ‘beach drawdown’ is a particular example of this case: if the area of extraction is not located far enough from the shore, that is, on the subtidal part of a beach, beach sediment may slump down into the excavated area (e.g., during winter storms) and not be able to return subsequently to the beach (in the summer). Coastal erosion may take place also if the protection of the coast is reduced. In particular, dredging may lower significantly the crest heights of (shallow) banks that normally reduce wave breaking and hence, when removed, would lead to larger wave heights at the coast. In addition, if the alterations in the sea bottom topography affect the wave refraction patterns, erosion may result from modification of the wave directions at the coastline. A substantial list of parameters has to be considered to evaluate whether dredging leads to erosion on the shore: water depth; the distance from the shore; the distribution of the banks; the degree of exposure; the direction of the prevailing waves and tidal currents; the frequency, direction, and severity of storms; the wave reflection and refraction pattern; and seabed type and sediment mobility. Predictions on the changes are performed generally using

mathematical models in which input data are provided through a number of sources, including surveys, charts, and offshore buoys. The modeling results have often a wide error range due to the relative uncertainties in the estimation of processes, especially for sediment transport calculations. Biological Impacts

The most obvious harmful direct effect of marine aggregate extraction results from the direct removal of substrata and associated benthic epifauna and inflora. Available evidence indicates that dredging causes an initial reduction in the abundance, species diversity, and biomass of a benthic community. Other impacts on benthic communities (and referred to as indirect effects) arise from the modification of abiotic factors, which determine the broad-scale benthic community patterns. These modifications concern the nature and stability of the sediment, through the exposition of underlying strata, and the tidal current strength, through the alteration of the local topography. The creation of relatively deep depressions with anoxic conditions is another source of concern. Sediment plumes arising from the dredging activities may also have adverse biological effects, by affecting water quality, increasing the sediment deposition, and modifying the sediment type. In the case of clean mobile sands, increased sedimentation and resuspension, as a consequence of dredging, are considered generally to be of less concern, as the fauna inhabiting such areas tend to be adapted to naturally high levels of suspended sediments, caused by wave and tidal current action. The effects of sediment deposition and resuspension may be more significant in gravelly habitats dominated by encrusting epifaunal taxa, due to the abrasive impacts of suspended sediments. In addition, turbidity plumes may have indirect impacts on fisheries, due to avoidance behavior by some species, the interruption of migratory pathways (e.g., lobsters), and the loss of access to traditional fishing grounds. The importance of the impacts that affect a benthic community is investigated commonly on the basis of the number of species, abundance, biomass, and the age of the population, compared to an undisturbed similar baseline. The existence of negligible or significant impacts depends largely upon the spatial and temporal intensity of dredging and, likewise, upon the nature of the benthos. While the recolonization of a dredged site by benthic species is generally rapid, the complete recovery (also ‘restoration’ or ‘readjustment’) of a benthic community can take many years. The recovery rate, after dredging, depends primarily upon several factors: the community



diversity and richness of the area prior to dredging; the life cycle and growth rate of species; the nature, magnitude, and duration of the dredging operations; the nature of the new sediment which is exposed, or subsequently accumulated at the extraction site; the larval and adult pool of potential new colonizers; the hydrodynamics and associated bed load transport processes; and the nature and intensity of the stresses (caused, for example, by storms) which the community normally withstands. Thus, among a number of studies addressing the consequences of long-term dredging operations on the recolonization of biota and the composition of sediment following cessation, some disparity appears in the findings. These results range from minimal effects of disturbance after dredging to significant changes in community structure, persisting over many years.

Supply and Demand: The Future of Offshore Sand and Gravel Mining Offshore sand and gravel constitute a limited and nonrenewable resource. The notion of supply is linked strongly to the usage of the material. As such, a distinction has to be made between the (reserves of) primary aggregates, suitable for use in construction, and the (resources of) secondary aggregates, which may include contaminants, and used typically as infilling material. For example, in-filling sand is available in relatively large quantities, whereas sand for construction, whose demand for quality is comparatively higher, is not so abundant. Gravel, which is also considered generally as a high-quality material, is in short supply. In the UK waters, the industry estimates that marine aggregate availability will last for 50 years, at present levels of extraction. However, such estimations are subject to important fluctuations, due to limited knowledge of the seafloor surface and subsurface. Furthermore, the level of accuracy of the identification of the resource varies greatly, from country to country (in general, data are far more complete in countries that have a greater reliance on marine sand and gravel). Another frequent difficulty toward a reliable assessment of the supply comes from the dispersal of the data relevant to the resource among governmental organizations, hydrographic departments, the dredging industry, and other commercial companies. In order to tackle this issue, a number of countries are undertaking seabed mapping programs, dedicated to the recognition of marine aggregates deposits. These programs range from detailed resource assessment, to reconnaissance level of the resource.

189

The precise estimation of the future demand for aggregates in general, including marine sand and gravel, is not an easy task. Difficulties arise from the highly variable and changing character of a number of parameters, such as the construction market and other economic factors, political strategies, environmental considerations, etc. In addition, the estimates can be distorted by extensive public work projects, involving large amounts of material. Moreover, the predictions should consider the exports and imports of material, and, therefore, take into account the demand at a much broader scale than the national one alone. Few countries have published estimates for the future demand of aggregate (only the UK and the Netherlands, in Europe). However, in a number of countries, on-land aggregate mining takes place within the context of increasing environmental pressure and the depletion of the resource. At the same time, large parts of the coasts, on a worldwide scale, experience increasing coastal erosion. Therefore, it is expected that the future marine sand and gravel extraction is bound to increase, in order to supply high-quality material and to provide the large quantities of aggregates required for the realization of large-scale infrastructure projects designed to manage coastal retreat and accommodate the development of the coastal zones (i.e., beach replenishment, dune restoration, foreshore nourishment, land reclamation, and other coastal defense schemes). Hence, in order to find these new resources, it is expected that the activity will move farther offshore, within the next decade.

See also Beaches, Physical Processes Affecting. MidOcean Ridge Seismicity. Sandy Beaches, Biology of. Sediment Chronologies. Sedimentary Record, Reconstruction of Productivity from the. Seismic Structure. Sonar Systems.

Further Reading Bellamy AG (1998) The UK marine sand and gravel dredging industry: An application of quaternary geology. In: Latham J-P (ed.) Engineering Geology Special Publications 13: Advances in Aggregates and Armourstone Evaluation, pp. 33--45. London: Geological Society. Boyd SE, Limpenny DS, Rees HL, and Cooper KM (2005) The effects of marine sand and gravel extraction on the macrobenthos at a commercial dredging site (results 6 years post-dredging). ICES Journal of Marine Science 62: 145--162.


190


Byrnes MR, Hammer RM, Thibaut TD, and Snyder DB (2004) Physical and biological effects of sand mining offshore Alabama, USA. Journal of Coastal Research 20(1): 6--24. Desprez M (2000) Physical and biological impact of marine aggregate extraction along the French coast of the Eastern English Channel. Short and long-term postdredging restoration. ICES Journal of Marine Science 57: 1428--1438. Diesing M (2003) Die Regeneration von Materialentnahmestellen in der Su¨dwestlichen Ostsee unter Besonderer Beru¨cksichtigung der Rezenten Sedimentdynamik, 158pp. PhD Thesis, Christian-AlbrechtsUniversita¨t, Kiel. Diesing M, Schwarzer K, Zeiler M, and Kelon H (2006) Special Issue: Comparison of Marine Sediment Extraction Sites by Mean of Shoreface Zonation. Journal of Coastal Research 39: 783–788. ICES (1992) Effects of extraction of marine sediments on fisheries. Cooperative Research Report, No. 182. Copenhagen: International Council for the Exploration of the Sea. ICES (2001) Report of the ICES Working Group on the effects of extraction of marine sediments on the marine ecosystem. Cooperative Research Report, No. 247. Copenhagen: International Council for the Exploration of the Sea. Kenny AJ and Rees HL (1996) The effects of marine gravel extraction on the macrobenthos: Results 2 years postdredging. Marine Pollution Bulletin 32(8/9): 615--622. Manso F, Diesing M, Schwarzer K, and Garel E (2004) Monitoring of dredged areas by combination of multibeam echosounder and sidescan sonar data. Conference Littoral 2004, Aberdeen, UK, 20–22 September 2004. Newell RC, Seiderer LJ, and Hitchcock DR (1998) The impact of dredging works in coastal waters: A review of the sensitivity to disturbance and subsequent recovery of biological resources on the sea bed. Oceanography and Marine Biology 36: 127--178. Roos PC (2004) Seabed Pattern Dynamics and Offshore Sand Extraction, 166pp. PhD Thesis, University of Twente, The Netherlands (ISBN 90-365-2067-3). van Dalfsen JA, Essink K, Toxvig madsen H, et al. (2000) Differential response of macrozoobenthos to marine sand extraction in the North Sea and the Western Mediterranean. ICES Journal of Marine Science 57(5): 1439--1445.

van der Veer HW, Bergman MJN, and Beukema JJ (1985) Dredging activities in the Dutch Waddensea: Effects on macrobenthic infauna. Netherlands Journal of Sea Research 19: 183--190. van Moorsel GWNM (1994) The Klaverbank (North Sea), geomorphology, macrobenthic ecology and the effect of gravel extraction. Report No. 94.24. Culemborg: Bureau Waardenburg BV. Van Rijn LC, Soulsby RL, Hoekstra P, and Davies AG (eds.) (2005) SANDPIT: Sand Transport and Morphology of Offshore Sand Mining Pits Process Knowledge and Guidelines for Coastal Management. Amsterdam: Aqua Publications.

Relevant Websites http://www.bmapa.org – British Marine Aggregate Producers Association (BMAPA). http://www.dredging.org – Central Dredging Association (CEDA): Serving Africa, Europe, and the Middle East. http://www.ciria.org – Construction Industry Research and Information Association (CIRIA) http://www.ciria.org – European Marine Sand and Gravel Group (EMSAGG). http://www.dvz.be – ICES Working Group on the Effect of Extraction of Marine Sediments on the Marine Ecosystem (WGEXT). http://www.iadc-dredging.com – International Association of Dredging Companies (IADC). http://www.ices.dk – International Council for the Exploration of the Sea (ICES). http://www.sandandgravel.com – Marine Sand and Gravel Information Service (MAGIS). http://www.westerndredging.org – Western Dredging Association (WEDA): Serving the Americas. http://www.woda.org – World Organisation of Dredging Associations (WODA).


OIL POLLUTION J. M. Baker, Clock Cottage, Shrewsbury, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 1999–2007, & 2001, Elsevier Ltd.

evaporation rates, and this is related to the molecular weights of the compounds they contain. Crude oils are also variable in their chemical composition and physical properties, depending on the field of origin. Chemical and physical properties are factors which partly determine environmental effects of spilt oil.

Introduction This article describes the sources of oil pollution, composition of oil, fate when spilt, and environmental effects. The initial impact of a spill can vary from minimal to the death of nearly everything in a particular biological community, and recovery times can vary from less than one year to more than 30 years. Information is provided on the range of effects together with the factors which help to determine the course of events. These include oil type and volume, local geography, climate and season, species and biological communities, local economic and amenity considerations, and clean-up methods. With respect to clean-up, decisions sometimes have to be made between different, conflicting environmental concerns. Oil spill response is facilitated by pre-spill contingency planning and assessment including the production of resource sensitivity maps.

Tanker Accidents Compared With Chronic Inputs Tanker accidents represent a small, but highly visible, proportion of the total oil inputs to the world’s oceans each year (Figure 1). The incidence of large tanker spills has, however, declined since the 1970s. Irrespective of accidental spills, background hydrocarbons are ubiquitous, though at low concentrations. Water in the open ocean typically contains 1–10 parts per billion but higher concentrations are found in nearshore waters. Sediments or organisms may accumulate hydrocarbons in higher concentrations than does water, and it is common for sediments in industrialized bays to contain several hundred parts per million. Sources of background hydrocarbons include:

Oil: A High Profile Pollutant

•

Consider some of the worst and most distressing effects of an oil spill: dying wildlife covered with oil; smothered shellfish beds on the shore; unusable amenity beaches. It is not surprising that ever since the Torrey Canyon in 1967 (the first major tanker accident) oil spills have been media events. Questions about environmental effects and adequacy of response arise time and time again, but to help answer these it is now possible to draw upon decades of experience from three main types of activity: post-spill case studies and long-term monitoring; field experiments to test clean-up methods; and laboratory tests to investigate toxicities of oils and dispersants. Oil is a complex substance containing hundreds of different compounds, mainly hydrocarbons (compounds consisting of carbon and hydrogen only). The three main classes of hydrocarbons in oil are alkanes (also known as paraffins), cycloalkanes (cycloparaffins or naphthenes) and arenes (aromatics). Compounds within each class have a wide range of molecular weights. Different refined products ranging from petrol to heavy fuel oil obviously differ in physical properties such as boiling range and

• •

Operational discharges (e.g. bilge water) from ships and boats; Land-based discharges (e.g. industrial effluents, rainwater run-off from roads, sewage discharges); Natural seeps of petroleum hydrocarbons, such as occur, for example, along the coasts of Baffin Island and California;

Natural sources 7% Atmosphere 9%

Exploration production 2% Industrial discharge and urban run-off 37%

Tanker accidents 12%

Vessel operations 33% Figure 1 Major inputs of oil to the marine environment.


191

192

• •

OIL POLLUTION

Airborne combustion products, either natural (e.g. from forest fires) or artificial (e.g. from the burning of fossil fuels); Organisms, e.g. leaf waxes and hydrocarbons synthesized by algae.

Notwithstanding the relatively great total amounts of oil from operational and land-based sources these discharges usually comprise diluted, dissolved and dispersed oil. The sections below focus on accidental spills because it is these which produce visible slicks which may coat wildlife and shores, and which typically require a clean-up response.

Natural Fate of Oil Slicks When oil is spilt on water, a series of complex interactions of physical, chemical, and biological processes takes place. Collectively these are called ‘weathering’; they tend to reduce the toxicity of the oil and in time they lead to natural cleaning. The main physical processes are spreading, evaporation, dispersion as small drops, solution, adsorption onto sediment particles, and sinking of such particles. Degradation occurs through chemical oxidation (especially under the influence of light) and biological action – a large number of different species of bacteria and fungi are hydrocarbon degraders. Case history evidence gives a reasonable indication of

natural cleaning timescales in different conditions. For open water sites, half-lives (the time taken for natural removal of 50% of the oil from the water surface) typically range from about half a day for the lightest oils (e.g. kerosene) to seven days or more for heavy oils (e.g. heavy fuel oil). However, for large spills near coastlines, some oil typically is stranded on the shore within a few days; once oil is stranded, the natural cleaning timescale may be prolonged. On the shore observed timescales range from a few days (some very exposed rocky shores) to more than 30 years (some very sheltered shores notably salt marshes). It is estimated that natural shore cleaning may take several decades in extreme cases. Shore cleaning timescales are affected by factors including:

•

• •

exposure of the shore to wave energy (Figure 2) from very exposed rocky headlands to sheltered tidal flats, salt marshes and mangroves. This in turn depends on a number of variables that include fetch; speed, direction, duration and frequency of winds; and open angle of the shore; localized exposure/shelter – even on an exposed shore, cracks, crevices, and spaces under boulders can provide sheltered conditions where oil may persist; steepness/shore profile – extensive, gently sloping shores dissipate wave energy;

12

Retention time (years)

10

8

6

4

2

0

Very sheltered Very sheltered porous boulder shores rocky shores

Moderately exposed rocky shores

Exposed rocky shores

no oil left

thick patches

oil stains or thin patches

liquid oil

Figure 2 Oil residence times on a variety of rocky shores where no clean up has been attempted.


OIL POLLUTION

0

Loose surface material including leaf fragments Live roots

Loose surface material including leaf fragments

5 Depth (cm)

193

Holes and burrows

Decaying root

10

15

Loose organic material oil mangrove peat

20 Figure 3 Two core samples taken from a mangrove swamp, showing penetration of oil down biological pathways.

•

• •

•

substratum – oil does not easily penetrate fine sediments (especially if they are waterlogged) unless they have biological pores such as crab and worm burrows or root holes (Figure 3). Penetration of shingle, gravel and some sand beaches can take place relatively easily, sometimes to depths of more than one metre; clay-oil flocculation – this process (first noticed on some Alaskan shores after the Exxon Valdez spill) reduces adherence of oil to shore substrata and facilitates natural cleaning; height of the stranded oil on the shore – oil spots taken into the supratidal zone by spray can persist for many years, conversely, oil on the middle and lower shore is more likely to be removed by water action. It is common to have stranded oil concentrated in the high tide area; oil type, e.g. viscosity affects movement into and out of sediment shores.

Oil Spill Response Aims

The aims of oil spill response are to minimize damage and reduce the time for environmental recovery by:

• •

guiding or re-distributing the oil into less sensitive environmental components (e.g. deflecting oil away from mangroves onto a sandy beach) and/or removing an appropriate amount of oil from the area of concern and disposing of it responsibly.

Initiation of a response, or decision to stop cleaning or leave an area for natural clean-up, needs to be focused on agreed definitions of ‘how clean is clean’, otherwise there is no yardstick for determining whether the response has achieved the desired result. The main response options are described below.

Booms and Skimmers

Booms and skimmers can be successful if the diversion and containment of the oil starts before it has had too much time to spread over the water surface. Booms tend to work well under calm sea conditions, but they are ineffective in rough seas. When current speeds are greater than 0.7–1.0 knots (1.3–1.85 km h 1), with the boom at right angles to the current, the oil is entrained into the water, passes under the boom and is lost. In some cases it may be possible to angle the boom to prevent this. Booms can also be used for shoreline protection, for example to stop oil from entering sheltered inlets with marshes or mangroves. The time available for protective boom deployment depends on the position and movement of the slick, and can vary from hours to many days. The efficiency of skimmers depends on the oil thickness and viscosity, the sea state and the storage capacity of the skimmer. Skimmers normally work best in sheltered waters. Because of their limitations, recovering 10% of the oil at a large spill in open seas is considered good for these mechanisms. Dispersants

Dispersants, which contain surfactants (surface active agents), reduce interfacial tension between oil and water. They promote the formation of numerous tiny oil droplets and retard the recoalescence of droplets into slicks. They do not clean oil out of the water, but can improve biodegradation by increasing the oil surface area, thereby increasing exposure to bacteria and oxygen. Information about dispersant effectiveness from accidental spills is limited for various reasons, such as inadequate monitoring and the difficulty of distinguishing between the contributions of different response methods to remove oil from the water surface. However, in at least some situations dispersants appear to remove a greater proportion of oil from the water surface than


194

OIL POLLUTION

mechanical methods. Moreover, they can be used relatively quickly and under sea conditions where mechanical collection is impossible. Dispersants, however, do not work well in all circumstances (e.g. on heavy fuel oil) and even for initially dispersable oils, there is only a short ‘window of opportunity’, typically 1–2 days, after which the oil becomes too weathered for dispersants to be effective. The main environmental concern is that the dispersed oil droplets in the water column may in some cases affect organisms such as corals or fish larvae.

In situ Burning

Burning requires the use of special fireproof containment booms. It is best achieved on relatively fresh oil and is most effective when the sea is fairly calm. It generates a lot of smoke and as with dispersants, the window of opportunity is short, typically a few hours to a day or two, depending on the oil type and the environmental conditions at the time of the spill. When it is safe and logistically feasible, in situ burning is highly efficient in removing oil from the water surface.

Nonaggressive Shore Cleaning

Nonaggressive methods of shore cleaning, methods with minimal impact on shore structure and shore organisms, include:

• • • • •

physical removal of surface oil from sandy beaches using machinery such as front-end loaders (avoiding removal of underlying sediment); manual removal of oil, asphalt patches, tar balls etc., by small, trained crews using equipment such as spades and buckets; collection of oil using sorbent materials (followed by safe disposal); low-pressure flushing with seawater at ambient temperature; bioremediation using fertilizers to stimulate indigenous hydrocarbon-degrading bacteria.

In appropriate circumstances these methods can be effective, but they may also be labor-intensive and clean-up crews must be careful to minimize trampling damage. Nonaggressive methods do not work well in all circumstances. Low-pressure flushing, for example, is ineffective on weathered firmly adhering oil on rocks; and bioremediation is ineffective for subsurface oil in poorly aerated sediments.

Aggressive Shore Cleaning

Aggressive methods of shore cleaning, those that are likely to damage shore structure and/or shore organisms, include:

• • •

removal of shore material such as sand, stones, or oily vegetation together with underlying roots and mud (in some cases the material may be washed and returned to the shore); water flushing at high pressure and/or high temperature; sand blasting.

In some cases these methods are effective at cleaning oil from the shore, e.g. hot water was more effective than cold water for removing weathered, viscous Exxon Valdez oil from rocks. However, heavy machinery, trampling and high-pressure water all can force oil into sediments and make matters worse. Net Environmental Benefit Analysis for Oil Spill Response

In many cases a possible response to an oil spill is potentially damaging to the environment. The public perception of disaster has sometimes been heightened by headlines such as ‘Clean-up makes things worse’. The advantages and disadvantages of different responses need to be weighed up and compared both with each other and with the advantages and disadvantages of entirely natural cleaning. This evaluation process is sometimes known as net environmental benefit analysis. The approach accepts that some response actions cause damage but may be justifiable because they reduce the overall problems resulting from the spill and response. Example 1 Consider sticky viscous fuel oil adhering to rocks which are an important site for seals. If effective removal of oil could only be achieved by high-pressure hot-water washing or sand blasting, prolonged recovery times of shore organisms such as algae, barnacles and mussels might be accepted because the seals were given a higher priority. A consideration here and in similar cases is that populations of wildlife species are likely to be smaller, more localized, and slower to recover if affected by oil than populations of abundant and widespread shore algae and invertebrates. Example 2 Consider a slick moving over shallow nearshore water in which there are coral reefs of particular conservation interest. The slick is moving towards sandy beaches important for tourism but of low biological productivity. Dispersant spraying will


OIL POLLUTION

Effects and Recovery Range of Effects

The range of oil effects after a spill can encompass:

• • • •

• • • • •

physical and chemical alteration of habitats, e.g. resulting from oil incorporation into sediments; physical smothering effects on flora and fauna; lethal or sublethal toxic effects on flora and fauna; short- and longer-term changes in biological communities resulting from oil effects on key organisms, e.g. increased abundance of intertidal algae following death of limpets which normally graze the algae; tainting of edible species, notably fish and shelfish, such that they are inedible and unmarketable (even though they are alive and capable of selfcleansing in the long term); loss of use of amenity areas such as sandy beaches; loss of market for fisheries products and tourism because of bad publicity (irrespective of the actual extent of the tainting or beach pollution); fouling of boats, fishing gear, slipways and jetties; temporary interruption of industrial processes requiring a supply of clean water from sea intakes (e.g. desalination).

The extent of biological damage can vary from minimal (e.g. following some open ocean spills) to the death of nearly everything in a particular biological community. Examples of extreme cases of damage following individual spills include deaths of thousands of sea birds, the death of more than 100 acres of mangrove forest, and damage to fisheries and/or aquaculture with settlements in excess of a million pounds. Extent of damage, and recovery times, are influenced by the nature of the clean-up operations, and by natural cleaning processes. Other interacting factors include oil type, oil loading (thickness), local geography, climate and season, species and biological communities, and local economic and amenity considerations. With respect to oil type, crude oils and products differ widely in toxicity. Severe toxic effects are associated with hydrocarbons with low boiling points, particularly aromatics, because these hydrocarbons are most likely to penetrate and disrupt cell membranes. The greatest toxic damage has been caused by spills of lighter oil particularly when

60

Medium height of shoots (cm)

minimize pollution of the beaches, but some coral reef organisms may be damaged by dispersed oil. From an ecological point of view, it is best not to use dispersants but to allow the oil to strand on the beaches, from where it may be quickly and easily cleaned.

195

50 40 30 20 10 0

Controls

LWFC HWFC Treatments (duplicate plots)

Figure 4 The effects of experimental oil treatments (duplicate plots) on shoot heights of the common salt marsh grass Spartina anglica. The measurements shown were taken four months after treatment. Lightly weathered Forties crude (LWFC) killed most of the grass shoots. Heavily weathered Flotta crude (HWFC) stimulated growth.

confined to a small area. Spills of heavy oils, such as some crudes and heavy fuel oil, may blanket areas of shore and kill organisms primarily through smothering (a physical effect) rather than through acute toxic effects. Oil toxicity is reduced as the oil weathers. Thus a crude oil that quickly reaches a shore is more toxic to shore life than oil that weathered at sea for several days before stranding. There have been cases of small quantities of heavy or weathered oils stimulating the growth of salt marsh plants (Figure 4). Geographical factors which have a bearing on the course of events include the characteristics of the water body (e.g. calm shallow sea or deep rough sea), wave energy levels along the shoreline (because these affect natural cleaning) and shoreline sediment characteristics. Temperature and wind speeds influence oil weathering, and according to season, vulnerable groups of birds or mammals may be congregated at breeding colonies, and fish may be spawning in shallow nearshore waters. Vulnerable Natural Resources

Salt marshes Salt marshes are sheltered ‘oil traps’ where oil may persist for many years. In cases where perennial plants are coated with relatively thin oil films, recovery can take place through new growth from underground stems and rootstocks. In extreme cases of thick smothering deposits, recovery times may be decades. Mangroves Mangrove forests are one of the most sensitive habitats to oil pollution. The trees are easily


196

OIL POLLUTION

killed by crude oil, and with their death comes loss of habitat for the fish, shellfish and wildlife which depend on them. Mangrove estuaries are sheltered ‘oil trap’ areas into which oil tends to move with the tide and then remain among the prop roots and breathing roots, and in the sediments (Figure 5). Coral reefs Coral reef species are sensitive to oil if actually coated with it. There is case-history evidence of long-term damage when oil was stranded on a reef flat at low tide. However, the risk of this type of scenario is quite low – oil slicks will float over coral reefs at most stages of the tide, causing little damage. Deep water corals will escape direct oiling at any stage of the tide.

which suggests that oil pollution has significant effects on fish populations in the open sea. This is partly because fish can take avoiding action and partly because oil-induced mortalities of young life stages are often of little significance compared with huge natural losses each year (e.g. through predation). There is an increased risk to some species and life stages of fish if oil enters shallow near-shore waters which are fish breeding and feeding grounds. If oil slicks enter into fish cage areas there may be some fish mortalities, but even if this is not the case there is likely to be tainting. Fishing nets, fish traps and aquaculture cages are all sensitive because adhering oil is difficult to clean and may taint the fish.

Fish Eggs, larvae and young fish are comparatively sensitive but there is no definitive evidence

Turtles Turtles are likely to suffer most from oil pollution during the breeding season, when oil at

North

0

0

6

50 meters

50 120

live Sonneratia mangrove live Rhizophora mangrove dead Rhizophora a mangrove oil concentration in surface 50 sediments (ppm dry weight)

dead Rhizophora

e

20

0

ive Sonneratia

meters

Figure 5 Plan and profile showing mangrove patches killed by small oil slicks.


OIL POLLUTION

197

Crabhall farm

IH

IH

IH

Dale flats

Windmill farm MS

MS

MS

Dale Slip

Section boundary Coastline

Rocky shore Boulder shore

Low water

Grid size = 1 km

Track Pathway

Detail is only provided for the section of coast referred to by the title and marked by the section boundaries

Land Sea

Access point

Pleasure craft

Diving birds

Recreational beach

Fisheries

Walkers

Sandy shore

Monitoring site

Houses

Muddy shore

Important habitat

Town

Shingle shore

Road

Waders

Saltmarsh Seagrass bed

Figure 6 An example of a sensitivity map. This large-scale map is part of an atlas covering the oil port of Milford Haven, Wales. It facilitates response when oil has come near shore or onshore, and the protection or clean-up of specific locations is being planned.

egg-laying sites could have serious effects on eggs or hatchlings. If oiling occurs, the effects from the turtle conservation point of view could be serious, because the various turtle species are endangered. Birds Seabirds are extremely sensitive to oiling, with high mortality rates of oiled birds. Moreover, there is experimental evidence that small amounts of oil transferred to eggs by sublethally oiled adults

can significantly reduce hatching success. Shore birds, notably waders, are also at risk. For them, a worst-case scenario would be oil impacting shore feeding grounds at a time when large numbers of migratory birds were coming into the area. Mammals Marine mammals with restricted coastal distributions are more likely to encounter oil than wide-ranging species moving quickly through


198

OIL POLLUTION

an area. Species at particular risk are those which rely on fur for conservation of body heat (e.g. sea otters). If the fur becomes matted with oil, they rapidly lose body heat and die from hypothermia. At sea, whales, dolphins and seals are at less risk because they have a layer of insulating blubber under the skin. However, there have been oil-related mortalities of young seals at breeding colonies.

areas, fish traps and aquaculture facilities. Other features include boat facilities such as harbors and slipways, industrial water intakes, recreational resources such as amenity beaches, and sites of cultural or historical significance. Sensitivities are influenced by many factors including ease of protection and clean-up, recovery times, importance for subsistence, economic value and seasonal changes in use.

Recovery

It is unrealistic to define recovery as a return to prespill conditions. This is partly because quantitative information on prespill conditions is only rarely available and, more importantly, because marine ecosystems are in a constant state of flux due to natural causes. These fluctuations can be as great as those caused by the impact of an oil spill. The following definition takes these problems into account. Recovery is marked by the re-establishment of a healthy biological community in which the plants and animals characteristic of that community are present and functioning normally. It may not have the same composition or age structure as that which was present before the damage, and will continue to show further change and development. It is impossible to say whether an ecosystem that has recovered from an oil spill is the same as, or different from, that which would have persisted in the absence of the spill.

Before a Spill

Before a spill it is important to identify what the particular sensitivities are for the area covered by any particular oil spill contingency plan, and to put the information on a sensitivity map which will be available to response teams. An example is shown in Figure 6. Maps should include information on the following.

• •

The response options need to be reviewed and finetuned throughout the response period, in the light of information being received about distribution and degree of oiling and resources affected. In extreme cases this process can be lengthy, for example over three years for the shoreline response to the Exxon Valdez spill in Prince William Sound, Alaska. In this case information was provided by shoreline clean-up assessment teams who carried out postspill surveys with the following objectives: assessment of the presence, distribution, and amount of surface and subsurface oil, and collection of information needed to make environmentally sound decisions on cleanup techniques. The standardized methods developed have subsequently been used as a model for other spills.

Acknowledgments

Assessment

•

After a Spill

Shoreline sensitivity. Shorelines may be ranked using the basic principles that sensitivity to oil increases with increasing shelter of the shore from wave action, penetration of oil into the substratum, natural oil retention times on the shore, and biological productivity of shore organisms. Typically, the least sensitive shorelines are exposed rocky headlands, and the most sensitive are marshes and mangroves forests. Other ecological resources such as coral reefs, seagrass and kelp beds, and wildlife such as turtles, birds and mammals. Socioeconomic resources, for example fishing areas, shellfish beds, fish and crustacean nursery

The International Petroleum Industry Environmental Conservation Association is gratefully acknowledged for permission to use material from its Report series.

See also Coral Reefs. Mangroves. Seabirds as Indicators of Ocean Pollution. Seabirds: An Overview. Seamounts and Off-Ridge Volcanism.

Further Reading American Petroleum Institute, Washington, DC. Oil Spill Conference Proceedings, published biennially from 1969 onwards. The primary source detailed papers on all aspects of oil pollution. IPIECA Report Series, Vol. 1 (1991) Guidelines on Biological Impacts of Oil Pollution; vol. 2 (1991) A Guide to Contingency Planning for Oil Spills on Water; vol. 3 (1992) Biological Impacts of Oil Pollution: Coral Reefs; vol. 4 (1993) Biological Impacts of Oil Pollution: Mangroves; vol. 5 (1993) Dispersants and their Role in Oil Spill Response; vol. 6 (1994) Biological Impacts of


OIL POLLUTION

Oil Pollution: Saltmarshes; vol. 7 (1995) Biological Impacts of Oil Pollution: Rocky Shores; vol. 8 (1997) Biological Impacts of Oil Pollution: Fisheries; vol. 9 (1999) Biological Impacts of Oil Pollution: Sedimentary Shores; vol. 10 (2000) Choosing Spill Response Options to Minimize Damage. London:

199

International Petroleum Industry Environmental Conservation Association. ITOPF (1987) Response to Marine Oil Spills. International Tanker Owners Pollution Federation Ltd, London. London: Witherby.


OKHOTSK SEA CIRCULATION oxygen, and other properties through ice processes, convection, and vigorous mixing before returning to the North Pacific. Relatively saline water from the Japan Sea assists in making Okhotsk Sea waters denser than those of the Bering Sea, which otherwise has similar processes but which does not produce intermediate water. Tides are exceptionally large within the Okhotsk which has broad, shallow continental shelves, providing a significant location for dissipation of tidal energy.

L. D. Talley, Scripps Institution of Oceanography, La Jolla, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2007–2015, & 2001, Elsevier Ltd.

Introduction The Okhotsk Sea (Figure 1) is one of the marginal seas of the north-western North Pacific. The circulation in the Okhotsk Sea is mainly counterclockwise. The Okhotsk Sea is the formation region for the intermediate water layer of the North Pacific. Water entering the Okhotsk Sea from the North Pacific is transformed in temperature, salinity, 135˚E

The Okhotsk Sea is enclosed by the Russian coastline to the north, Sakhalin and Hokkaido Islands to the west, the Kamchatka peninsula to the east and the

150˚E

145˚E

140˚E

Geography

160˚E

155˚E

RUSSIA

165˚E

y

Shelikov Bay

a ay

sk

in

60˚N

MAGADAN

OKHOTSK CITY

AYAN

55˚N

Pe

60˚N

Tinro Basin

nd

la

Is

h nz

Ba

KAMCHATKA

ky rs ta n a Sh Shantarsky Bay

Kashevarov Bank

55˚N

R.

ait

AM

Tatar S tr

50˚N

Terpenia Bay

ter

n

Bu

ss

ol

Kuril Basin

St

ra

IL

Soya Strait

45˚N

UR

HOKKAIDO

135˚E

140˚E

145˚E

ril

Ku

it

a

K

Japan Sea

50˚N

Kr St uze ra ns it h

IS LA ND S Tr en ch

UR

SAKHALIN

Derugina Basin

atk

45˚N

h mc

a -K

150˚E

155˚E

Figure 1 Okhotsk Sea geography.

200


160˚E

165˚E

OKHOTSK SEA CIRCULATION

northward at the eastern side of each strait and southward at the western side, leading to a net clockwise circulation of water around each of the Kuril Islands. Within the Okhotsk Sea, the maximum tidal currents and sea surface height displacements are in the northern bays and on Kashevarov Bank (Figure 2). Sea surface displacements can reach several meters in Penzhinskaya Bay in the north east. Maximum energy dissipation occurs in Shelikov Bay, with a somewhat less important site on Kashevarov Bank.

60

Latitude

55

50

45

145

140 (A)

155

160

165

155 150 Longitude

160

165

150 Longitude

167 cm/s 60

55 Latitude

Kuril Islands to the south east. Two other marginal seas are nearby: the Bering Sea east of Kamchatka, and the Japan (East) Sea west of Sakhalin and Hokkaido. These three marginal seas are characterized by limited connections to the North Pacific, relatively deep basins, and the presence of sea ice in winter. The Okhotsk Sea has a highly productive fishery, and is the site of explorative oil lease sites. The Okhotsk also occupies a unique role as the highest density surface source of waters for the North Pacific, feeding into the intermediate depth layer down to about 2000 m. All deeper water (and indeed, much of the water even in this intermediate layer) comes from the Southern Hemisphere. The Okhotsk Sea is connected to the North Pacific through 13 straits between the numerous Kuril Islands. Because of the political history of this region, most straits and islands bear both Russian and Japanese names. The two deepest straits are Bussol’ and Kruzenshtern, with sill depths of 2318 m and 1920 m, respectively, through which there is significant exchange of water with the North Pacific. Other straits of importance for exchange are Friza and Chetvertyy Straits, both with sill depths of about 600 m. The Okhotsk Sea is connected to the Japan Sea through two straits on either end of Sakhalin: Soya (La Perouse) Strait to the south and Tatar Strait to the north. Soya Strait, while shallow (55 m), is an important source of warm, saline water for the Okhotsk Sea. Tatar Strait is extremely shallow (5 m) and there is little exchange through it. The Okhotsk Sea is not connected directly to the Bering Sea, but much of the water flowing into the Okhotsk Sea originates in a current from the western Bering Sea. Within the Okhotsk Sea there are three deep basins, with the greatest depth being 3390 m in the Kuril Basin. An important characteristic of the Okhotsk Sea is its very broad continental shelves in the north; these impact tidal energy dissipation and formation of dense waters in winter. The fairly shallow, isolated Kashevarov Bank is found in the north west. A major river, the Amur, drains into the Okhotsk Sea north-west of Sakhalin.

201

50

Tides The Okhotsk Sea is one of the major tidal dissipation areas of the world ocean as a result of its very broad continental shelves. Tides in the North Pacific rotate counterclockwise. The tidal energy passing by the Kuril Islands enters the Okhotsk Sea. Tides around the Kuril Islands can have associated currents of 4–8 knots (20–40 cm s 1). The currents through each of the Kuril Straits associated with the tides are

45

140 (B)

145

Figure 2 (A) The amplitude (cm) and phase and (B) current ellipses of the dominant semi-diurnal tide (M2) in the Okhotsk Sea, from Kowalik and Polyakov (1998).


202


Circulation and Eddy Field The mean circulation of the Okhotsk Sea is counterclockwise (Figure 3). Because ocean currents are in geostrophic balance, this means that there is low pressure in the center of the Okhotsk Sea. The flow in the Okhotsk Sea is driven by the same wind field that drives the cyclonic circulation of the adjacent subpolar North Pacific. The prevailing winds are the western side of the Aleutian Low. These winds create upwelling in the subpolar region and Okhotsk Sea. Upwelling over a broad region causes counterclockwise (cyclonic) flow in the Northern Hemisphere. North Pacific water enters the Okhotsk Sea through the northernmost passages through the Kurils, primarily Kruzenshtern and Chetvertyy Straits. The North Pacific water comes from the East Kamchatka Current, which is a narrow southward boundary current along the eastern coast of Kamchatka, from

135˚E

140˚E

145˚E

the Bering Sea. The water that leaves the Okhotsk Sea through the southern Kuril Islands turns southward along the Kurils and joins the narrow, strong Oyashio. The Oyashio flows to the southern coast of Hokkaido, where it turns eastward and enters the North Pacific gyres. The net exchange between the Okhotsk Sea and the North Pacific is superimposed on clockwise flow around each of the Kuril Islands, driven by the strong tidal currents. Within the Okhotsk Sea, the inflow from the Pacific feeds a broad northward flow in the east called the West Kamchatka Current. The West Kamchatka Current is fragmented and broad and often contains an inshore countercurrent. The warmth of the West Kamchatka Current region keeps this region ice-free throughout the year. Along the northern shelves there is westward flow. The shelf flow rounds the northern tip of Sakhalin (Cape Elizabeth) and collects into a swift, narrow boundary current flowing southward along the east

150˚E

155˚E

160˚E

165˚E

60˚N

60˚N

55˚N

55˚N

50˚N

50˚N

45˚N

45˚N

135˚E

140˚E

145˚E

150˚E

155˚E

160˚E

165˚E

Figure 3 Mean circulation of the Okhotsk Sea, after numerous sources. (See Talley and Nagata, 1995, for collection of the many cartoons of the flow.)



side of Sakhalin, called the East Sakhalin Current. Where the East Sakhalin Current ends at Cape Terpeniya at the southern end of Sakhalin, it feeds into an eddy field with net transport toward Bussol’ Strait, in the center of the Kuril Islands. Southward motion of the winter ice pack from Cape Terpeniya to Hokkaido suggests that some of the East Sakhalin Current also continues southward. Water also enters the Okhotsk Sea from the Japan Sea, through Soya Strait. The narrow Soya Current carries this water along the northern coast of Hokkaido, moving towards the Kuril Islands. The typical speed of the Soya Current is 25–50 cm s 1, with greater speeds in Soya Strait. The Soya Current peaks in summer and is submerged or very weak in winter, which allows ice to form along Hokkaido in the absence of this relatively warm water. The Soya Current water joins the other waters exiting the Okhotsk Sea through the southern Kuril Islands, including through Bussol’ Strait. The Okhotsk Sea has a vigorous eddy field, meaning that often it is difficult to discern the mean flow because of the presence of moving, transient eddies of about 100–150 km scale. The eddy field is especially pronounced in the Kuril Basin as tracked by satellite imagery of the sea surface temperature. The two to four Kuril Basin eddies that are formed each year are clockwise (anticyclonic). A mean clockwise flow, perhaps divided into smaller subgyres, has been discerned in this eddy-rich basin. This anticyclonic permanent flow or eddy field conveys the East Sakhalin Current waters towards the central Kuril Islands and hence to the exit through Bussol’ Strait. The Soya Current often has dramatic eddies that form as the current becomes unstable and rolls up horizontally into backward-breaking waves. Eddies

Figure 4 Loose sea ice in the Soya Current, Showing a counterclockwise eddy off the Hokkaido coast from an aircraft at altitude 1000 ft. The eddy was about 20 km in diameter. From Wakatsuchi and Ohshima (1990).

203

of the Soya Current also form mushroom-shaped vortices. Soya Current eddies have been photographed and tracked using loose sea ice in winter (Figure 4).

Sea Ice in the Okhotsk Sea Ice forms every winter in the Okhotsk Sea and melts away completely every summer. Thus all ice in the Okhotsk is first-year ice. A small amount of ice is usually present by the end of October in the coastal areas of Shelikof and Penzhinskaya Bays, with formation by mid-November in Shantarsky Bay and off the west coast of Kamchatka. Ice expansion is rapid and by December ice is found throughout much of the northern Okhotsk Sea and around Sakhalin. Ice forms off Hokkaido by early January. Maximum ice extent in the Okhotsk occurs in late March when ice can cover almost the entire Okhotsk Sea in a heavy-ice winter. The south-eastern Okhotsk Sea, where relatively warm North Pacific waters enter, remains ice-free in even the most extreme winters. Fast ice (attached to land) is found in late winter in Shantarsky, Shelikov, and Penzhinskaya Bays and around Sakhalin. Maximum ice thickness in the Okhotsk is about 1.5 m in the north and 1 m in the central Okhotsk Sea. Circulation along the coast of Hokkaido is dominated by the warm, saline Soya Current entering from the Japan Sea. The Soya Current inhibits ice formation. It also exhibits large seasonality, being nearly completely submerged under fresh, cold water in winter and so ice can form along the Hokkaido coast. Pack ice from the Sakhalin area also reaches Hokkaido in early February. Ice melt begins in late March and usually finishes by the end of June or early July, with the last vestiges of ice usually found in Shantarsky Bay. In heavy-ice winters, this last ice melts in late July. In winter, pack ice often flows out of the Okhotsk Sea into the Pacific through the southern Kuril Islands. This and the fresh water generated by ice melt form the fresh coastal part of the Oyashio along the southern coast of Hokkaido. Within the ice-covered zones, there are usually several areas that are either free of ice or contain only thin frazil ice. Such openings are called polynyas. A polynya often forms over Kashevarov Bank in the north, where strong tidal currents continually upwell warm waters to the sea surface and keep the region icefree. The heat flux maintaining this polynya is provided by an upwelling of 0.3–0.6 m d 1 of water at 21C. Polynyas are also usually found in a narrow band along the north-western and northern coast between Shantarsky and Tauskaya Bays. The coastal polynyas are kept open by northerly or north-westerly winds


204


that force the coastal ice offshore. Therefore ice forms continuously within these polynyas. Similar wind-forced polynyas are found in several spots along Sakhalin and in Terpeniya Bay at the southern end of Sakhalin. Sea ice is always fresher than the sea water from which it forms since salt is rejected from the developing ice lattice. The rejected salt collects in pockets of brine within the ice and drips out at the bottom of the ice. This brine rejection process increases the salinity of the waters underneath the sea ice, which thus increases the density of these waters. This densification process is most effective in areas of active ice formation, such as the wind-created coastal polynyas, and where the water depth is not too great, so that the brine is less diluted as it mixes into the underlying water. In the Okhotsk Sea, the broad northern shelf with its coastal polynya is a site of active dense water formation.

Water Properties of the Okhotsk Sea The main water source for the Okhotsk Sea is the North Pacific just east of the Kuril Islands and upstream of the Okhotsk Sea in the East Kamchatka Current. These North Pacific waters are characterized by a shallow temperature minimum below the

sea surface, which is often called the ‘dichothermal’ layer (Figure 5A). Below this is a temperature maximum (the ‘mesothermal’ layer). The temperature minimum is supported by the existence of a low salinity surface layer (Figure 5B) since both temperature and salinity contribute to seawater density. The temperature minimum is a remnant of winter cooling, although it may reflect cooling farther upstream in the flow. The Okhotsk Sea waters also have this structure, although modified from the North Pacific’s, as described below. The second major source of water for the Okhotsk Sea is the Japan Sea, through Soya Strait. Japan Sea water is relatively saline, since it all originates as a branch of the Kuroshio at much lower latitude. Even though net precipitation in the Japan Sea reduces the salinity of the Kuroshio waters that enter it, the Soya Current waters feeding into the Okhotsk Sea are relatively saline. There is no similar source of relatively high salinity water for the Bering Sea. The resulting difference in overall salinity between the Okhotsk and Bering Seas is likely the main reason that the intermediate waters of the North Pacific are ventilated (originate at the sea surface) in the Okhotsk Sea rather than in the higher latitude Bering Sea. Salinity within the Okhotsk Sea is reduced by flow from the Amur River and local precipitation. Sea ice

Station Number

Station Number 1

5

10

15

20

25

30

Pressure (dbar)

Pressure (dbar)

100

200 0

Pressure (dbar)

Pressure (dbar)

5

0

200

10

15

20

25

30

100

200 0

1000

2000

1000

2000

3000

3000 0

(A)

1 0

0

200

400

600

800

1000

1200

Distance (km.)

1400

1600

1800

400

(B)

600

800

1000

1200

1400

1600

1800

Distance (km.)

Figure 5 Cross-sections of (A) potential temperature and (B) salinity extending from the North Pacific through Bussol’ Strait and to the northwest Okhotsk Sea, from Freeland et al. (1998).



production each winter creates higher density waters through brine rejection from the ice, and sea ice melt in spring and summer produces a freshened surface layer. Within the Okhotsk Sea the dichothermal layer is colder than outside in the Pacific. Ice formation throughout most of the Okhotsk Sea depresses the temperature of the minimum. In the north-west Okhotsk Sea, the temperature minimum is close to the freezing point due to ice production in this shallow region in winter, causing the waters to the shelf bottom to be near freezing. Fresh, cold, oxygenated water penetrates much deeper with the Okhotsk Sea than in the East Kamchatka Current

60˚N

135˚E

140˚E

145˚E

which is its primary source. The temperature maximum in the Okhotsk Sea is near 1000 m, considerably deeper than the mesothermal layer in the East Kamchatka Current which lies at about 300 m. In summer, the surface salinity is very low, o32.8 PSU, and considerably lower near the Amur River outflow. The generally low salinity is due to ice melt and river discharge. Salinity at the temperature minimum is about 33.0 PSU and then increases gradually to the bottom, consonant with the North Pacific source of the deep waters. Several important water transformation processes occur in the Okhotsk: densification resulting from ice formation, convection resulting from cooling,

150˚E

155˚E

60˚N

55˚N

55˚N

50˚N

50˚N

45˚N

45˚N

135˚E

140˚E

145˚E

205

150˚E

155˚E

Figure 6 Bottom density (in units of kg/m 3–1000) for depths less than 300 m is shown with the heavy contours. Bottom density is shaded where especially dense shelf water is found, with density greater than 1026.9 kg m 3. The light contours show bottom depths of 300 and 1000 m. The dots indicate where observations were made. Large diamonds show observation positions where the water depth is less than 300 m. Large stars indicate that the temperature at the bottom is colder than 1.01C. The data are from Kitani (1973) and from surveys of the Okhotsk Sea in 1994 and 1995, with the latter data provided by Rogachev (Pacific Oceanological Institute, Vladivostok, Russia) and Riser (University of Washington).


206


input of low salinity at the surface from rivers, precipitation and ice melt, and mixing which is greatly accentuated in the Okhotsk because of its large tidal amplitudes. Ice formation over the northern shelves, particularly in the coastal polynyas, creates a dense shelf water that moves cyclonically around to Shantarsky Bay and then out past the northern end of Sakhalin and down along the sloping side to feed the dichothermal layer. The density of the new shelf water (Figure 6) can be surmised from nonwinter observations. Based on data collected over many years, it appears that the maximum density of the shelf waters and their offshore mixture is 1027.2 kg m 3. Convection due to heat loss occurs in the Okhotsk Sea, notably in the Kuril Basin to a depth of about 500 m. The maximum density affected by convection is about 1026.85 kg m 3, and so ice formation on the shelves creates denser water than does convection. This limit on the density created by convection is set by the salinity of surface waters in the Kuril Basin and the maximum density they can reach when cooled to freezing. (Densification through ice formation has negligible effect in deep water since the rejected salt mixes into a thick water column.) However, the convective mixing is important as a signature of Okhotsk Sea water transformation as the waters enter the Oyashio and move southward into the North Pacific, since the thickness of the newly convected layer is retained to some extent. Mixing is a much more significant process in the Okhotsk than in the open North Pacific because of the large tidal amplitudes and topography within the sea. Two locations especially deserve mention – in the straits between the Kuril Islands and over Kashevarov Bank in the north west. Mixing over the latter moves relatively warm water to the sea surface, melting out the sea ice there. Mixing over the nearby shelves may also be an important factor in setting the maximum density of the winter shelf waters, since the mixing could bring higher salinity waters from offshore onto the shelves. In the Kuril Straits, tidal currents are large, and oppose each other on opposite sides of each strait. In particular, water properties in Bussol’ Strait have long been observed to be strongly mixed, to the bottom of the strait. This mixing brings the high oxygen of the upper waters down to the sill depth and is the major source for the North Pacific down to this depth of oxygen and other atmospheric gases such as chlorofluorocarbons.

‘ventilation’ processes of the Okhotsk directly affect densities higher than elsewhere in the North Pacific. Deep and bottom waters are not formed at the sea surface in the North Pacific – globally they are formed only in the North Atlantic and around Antarctica. However, the intermediate layer of the North Pacific, between about 500 and 2000 m, is ventilated through Okhotsk Sea processes, similar to the impact of intermediate water formation in the Labrador Sea of the North Atlantic and Southern Hemisphere intermediate water formation around southern South America. The intermediate layer of the North Pacific lies between the salinity minimum found in the subtropical region, at a density of 1026.8 kg m 3, and the deep waters found below 2000 m, or a density of about 1027.6 kg m 3. The most recently ventilated water in the North Pacific, as marked by laterally high oxygen (Figure 7) and chlorofluorocarbons and other gases of atmospheric origin, is found in the north west, in the neighborhood of the Okhotsk Sea. The intermediate layer is also freshest in this area. The evidence described above indicates that the Okhotsk Sea is the source of the ventilation. The bottom of the North Pacific’s ventilated intermediate water layer is set by the sill depth of Bussol’ Strait, that is, by the maximum depth and density of the vertical mixing that moves ventilated waters downward in the Okhotsk Sea, with outflow into the North Pacific. The water that leaves the Okhotsk Sea through Bussol’ Strait turns southward and becomes the Oyashio. Its properties are significantly different from those of the East Kamchatka Current, that feeds water into the Okhotsk Sea farther north. The Oyashio continues southward to the southern end of Hokkaido and then turns offshore and to the east.

60˚N

50˚

40˚

Connection of the Okhotsk Sea with North Pacific Processes

130˚

The Okhotsk Sea is an important factor in air–sea exchange and overturn in the North Pacific. The

Figure 7 Oxygen on a constant density surface in the North Pacific Intermediate Water layer, from Talley (1991). This surface lies at about 300–400 m depth within the Okhotsk Sea.

140˚


150˚

160˚

170˚

180˚


The Oyashio waters, including the Okhotsk Sea products, then enter the interior of the North Pacific.

See also Bottom Water Formation. Kuroshio and Oyashio Currents. Polynyas. Sea Ice. Sea Ice: Overview. Tidal Energy. Tides. Upper Ocean Mixing Processes. Water Types and Water Masses. Wind Driven Circulation.

Further Reading Alfultis MA and Martin S (1987) Satellite passive microwave studies of the Sea of Okhotsk ice cover and its relation to oceanic processes, 1978–1982. Journal of Geophysical Research 92: 13 013--13 028. Favorite F, Dodimead AJ, and Nasu K (1976) Oceanography of the Subarctic Pacific region, 1960–71. Vancouver: International North Pacific Fisheries Commission 33: 1–187. Freeland HJ, Bychkov AS, Whitney F, et al. (1998) WOCE section P1W in the Sea of Okhotsk – 1. Oceanographic data description. Journal of Geophysical Research 103: 15 613--15 623. Kitani K (1973) An oceanographic study of the Okhotsk Sea – particularly in regard to cold waters. Bulletin of Far Seas Fisheries Research Laboratory 9: 45--76.

207

Kowalik Z and Polyakov I (1998) Tides in the Sea of Okhotsk. Journal of Physical Oceanography 28: 1389--1409. Moroshkin KV (1966) Water Masses of the Okhotsk Sea. Moscow: Nauka. (Translated from the Russian by National Technical Information Services, 1968.) Preller RH and Hogan PJ (1998) Oceanography of the Sea of Okhotsk and the Japan/East Sea. In: Robinson AR and Brink KH (eds.) The Sea, vol. 11, New York: John Wiley and Sons. Reid JL (1973) Northwest Pacific ocean waters in winter. Johns Hopkins Oceanographic Studies, Baltimore, MD: The Johns Hopkins Press 5: 1–96. Talley LD (1991) An Okhotsk Sea water anomaly: implications for ventilation in the North Pacific. DeepSea Research 38 (Suppl): S171--S190. Talley LD and Nagata Y (1995) The Okhotsk Sea and Oyashio Region. PICES Scientific Report No. 2. Sidney BC. Canada: North Pacific Marine Science Organization (PICES), Institute of Ocean Sciences Wakatsuchi M and Martin S (1990) Satellite observations of the ice cover of the Kuril Basin Region of the Okhotsk Sea and its relation to the regional oceanography. Journal of Geophysical Research 95: 13 393--13 410. Wakatsuchi M and Ohshima K (1990) Observations of iceocean eddy streets in the Sea of Okhotsk off the Hokkaido coast using radar images. Journal of Physical Oceanography 20: 585--594. Warner MJ, Bullister JL, Wisegarver DP, et al. (1996) Basin-wide distributions of chlorofluorocarbons CFC11 and CFC-12 in the North Pacific – 1985–1989. Journal of Geophysical Research 101: 20 525--20 542.


ONE-DIMENSIONAL MODELS H. Yamazaki, Tokyo University of Marine Science and Technology, Tokyo, Japan H. Burchard, Baltic Sea Research Institute Warnemu¨nde, Warnemu¨nde, Germany K. Denman, University of Victoria, Victoria, BC, Canada T. Nagai, Tokyo University of Marine Science and Technology, Tokyo, Japan & 2009 Elsevier Ltd. All rights reserved.

Introduction A one-dimensional model is a function of one spatial variable and time. In oceanographic applications, the spatial variable is normally the vertical coordinate, z. Since realistic flow in the ocean depends on all three independent coordinates, x, y, z, and time t, a one-dimensional water column model works well only under limited conditions: when the dynamics are horizontally homogeneous and lateral processes are negligible. Since the computational cost is low, a one-dimensional model is useful for developing conceptual models, testing numerical schemes, and performing sensitivity analyses. Walter Munk in 1966 argued, based on a onedimensional model, that a canonical value for the eddy diffusivity, Kz, should be 10 4 m2 s 1. Namely, if the thermal structure in the main thermocline remains in steady state despite thermal diffusivity acting continuously over millions of years, the diffusivity must be counterbalanced by vertical advection. He argued that, globally, the vertical advection should be equal to the formation of deep water in the North Atlantic Ocean and the Southern Ocean. In order to maintain the thermal structure in the rest of the ocean, vertical advection must balance turbulent diffusion locally throughout the water column: w

@T @2T ¼ Kz 2 @z @z

½1

where temperature, T, is only a function of z and w is the average vertical advection estimated from the deep water formation. Once the temperature structure is specified, the eddy diffusivity is easily estimated to be 10–4 m2 s–1, which is sometimes referred to as 1 Munk unit. The argument is straightforward, and should hold in a global average sense. Since Kz is an average quantity, this argument cannot predict how local Kz is varies in space. In fact, numerous field

208

experiments over many different regimes in the ocean interior show that Kz is about 10–5 m2 s 1. To arrive at an average 10 times larger, intense mixing must occur in confined areas, clearly showing the limitation of one-dimensional models. However, this onedimensional model is still useful in illustrating the role of mixing in maintaining the vertical temperature gradient in the ocean. Mixing is particularly strong in the upper ocean due to wind stress and convective processes. When the forcing for the upper ocean mixing is horizontally homogenous, upper ocean dynamics may be expressed in terms of a one-dimensional model, often referred to as a water column model. First, for the dynamics to be physically correct, these models are based on the Navier–Stokes equations (NSEs). Next, to deal with biological processes in the upper ocean, a proper ecosystem model must be solved simultaneously with the dynamic physical model. These equations are usually expressed in an Eulerian framework, where the state variables are functions of space and time. However, the Lagrangian approach, which follows individual ‘particles’ or organisms in space, can capture behavior and variability in population characteristics more directly. The next section introduces several classes of the upper ocean water column dynamics. In the subsequent section we couple the physics and the primary producers of the marine ecosystem, the plankton. Then, we present key examples, and in the last section we discuss some general perspectives.

Numerical Model of the Upper Ocean Mixed Layer The NSEs, developed more than 150 years ago, provide dynamic equations for the turbulent velocity structure, and include all the physics necessary for describing ocean dynamics, from millimeter to global scales. However, even within the ocean mixed layer these equations cannot be solved numerically, since cubes with edge lengths of the order of the mixed layer depth have to be resolved by means of cubes of millimeter scale, which is about the smallest scale of relevant turbulent motions. Solutions over these scales will not be feasible during the next decades. Therefore, substantial simplifications of the NSEs are necessary for predictive upper ocean modeling. Since the ocean is turbulent and turbulence is a three-dimensional phenomenon, throughout the upper ocean there are strong gradients of all properties in all


ONE-DIMENSIONAL MODELS

directions (and in time). Thus, the basic assumption of one-dimensional modeling – neglect of horizontal gradients – is in contradiction to resolving the upper ocean physics. However, only the statistical properties of turbulence and not the instantaneous spatial and temporal patterns are relevant for measures like mixed layer depth or turbulent fluxes through the thermocline which are statistical properties themselves. Therefore, we can focus our upper ocean modeling on reproducing statistics of the turbulence. The foundations of this approach are the Reynolds averaged NSEs together with dynamic equations for potential temperature and salinity. The strategy is to decompose any physical quantity u into a mean and a fluctuating part: u ¼ u¯ þ u0

½2

where the over bar denotes mean quantities and the prime denoting fluctuations. The mean part is meant to represent an ensemble average, that is, the spaceand time-dependent fields that would result from averaging over a large number of flow realizations with statistically identical external forcing. It is thus required that the mean of the fluctuating part vanishes. Inserting this so-called Reynolds decomposition into the NSEs and the dynamic equation for potential temperature (salinity variations are neglected here for simplicity), and carrying out a number of algebraic operations, leads to dynamic equations for averaged quantities such as mean velocity components u; ¯ v; ¯ w, ¯ ¯ In the dynamic and the mean potential temperature y. equations for these mean quantities, we obtain, due to nonlinearities in the NSEs and the dynamic temperature equation, second-order statistical moments such as u0 w0 , v0 w0 (the vertical turbulent momentum fluxes or Reynolds stresses), and w0 y0 (the vertical turbulent heat flux):

209

parametrizations of the turbulent fluxes through the thermocline. Direct parametrizations of eddy-induced mixing coefficients are given by the K-profile parametrization. More powerful is the statistical turbulence modeling approach, which is based on deriving dynamic equations for the higher-order statistical moments of the turbulent flow. Analytical Near-surface Layer Models

With highly simplified assumptions, instructive analytical models can be constructed. If homogeneous density and constant viscosity and surface stress are assumed, the classical Ekman spiral results, with current speed decreasing exponentially with depth, with surface currents turned by 451 toward the right (in the Northern Hemisphere) of the wind direction due to Earth’s rotation, and with the vertically integrated transport being turned by 901 toward the right. Although the assumption of constant eddy viscosity is unrealistic, the estimated e-folding depth of the velocity is a good estimator for the mixed layer depth: if the latter is less than the former, significant mixed layer deepening can be expected. Bulk Models

@ u¯ @ þ ðw0 u0 Þ f v¯ ¼ 0 @t @z

½3

Bulk models are obtained by vertically integrating over the mixed layer depth. Mixed layer deepening is parametrized as function of surface fluxes, gradients across the thermocline, and mixed layer depth, a process that causes entrainment of ambient properties from below into the mixed layer. Decreasing mixed layer depth is parametrized as detrainment, typically caused by decreasing wind stress and increased surface buoyancy flux into the mixed layer. In the 1960s and 1970s, bulk models were popular because of their low computational cost. At present, they are mainly used for parametrizing the surface ocean as the lower boundary condition for atmospheric models.

@ v¯ @ þ ðw0 v0 Þ þ f u¯ ¼ 0 @t @z

½4

K-profile Parametrization

@ y¯ @ þ ðw0 y0 Þ ¼ 0 @t @z

½5

where f is the Coriolis parameter that accounts for planetary rotation. As they stand, these equations are not directly applicable for upper ocean studies, since they are mathematically not closed. Simple empirical models for the vertical mixed layer structure may be formulated based on steady-state approximations of these equations. Bulk models for the mixed layer result from vertically integrating these equations and

The K-profile parametrization (KPP) provides an algorithm for directly calculating the eddy viscosity and diffusivity for the mixed layer and ambient ocean below. The vertical shape of these eddy coefficient profiles is given as nondimensional cubic functions of depth, decreasing toward surface and thermocline below the mixed layer. Those are then scaled with the mixed layer depth (diagnosed from vertical density distribution) and a surface fluxrelated velocity scale. This velocity scale is composed of the wind stress-related friction velocity and the Deardorff velocity scale which is related to the


210


surface buoyancy flux. In the thermocline, eddy viscosity and diffusivity values depend on internal wave activity diagnosed from the gradient Richardson number. Various extensions of the KPP have been developed, such as the consideration of nonlocal momentum and density fluxes. The KPP is mainly applied for large-scale ocean models with fairly low vertical resolution, where these KPP models have provided stable and realistic performance. Turbulence Closure Models

Dynamic equations for second-order statistical moments can be obtained by further analytical treatment of the NSEs and the dynamic temperature equations, but these will include third-order moments such as w0 y0 y0 (vertical turbulent flux of temperature variance). In turn, equations for these third-order moments can be derived, which then include fourth-order moments and so forth, such that we obtain an infinite series of equations (the socalled Friedmann–Keller series), not leading to a mathematical closure. This classical turbulence closure problem can be solved by expressing statistical moments of a certain order by moments of a lower order. For upper ocean modeling, secondmoment closures have proved to be a powerful tool, expressing all third moments in terms of second and first moments. Furthermore, the dynamic equations for the second moments are often assumed to be in turbulent equilibrium, which transform them into algebraic relations. The resulting second moments are then given by means of a down-gradient parametrization as w0 u0 ¼ Km

@ u¯ ; @z

w0 v0 ¼ Km

w0 y0 ¼ Kh

@ v¯ @z

@ y¯ @z

½6

k2 ; e

K h ¼ Sh

k3=2 Lp e

½10

Another type of kmen-equation is the o-equation with o ¼ e/k, leading to the k o model. In many studies, an algebraic expression is constructed for the dissipation rate. In principle, it should be possible to combine all approaches for the calculation of the dissipation rate profiles and for the stability functions. However, some physically based adjustments have to be made for each combination. All these considerations have been implemented into a public domain water-column model – General Ocean Turbulence Model (GOTM).

½7

with the eddy viscosity Km and the eddy diffusivity Kh . The eddy mixing coefficients are functions of the turbulent kinetic energy k (TKE, kinetic energy of the velocity fluctuations) and its dissipation rate into heat, e: K m ¼ Sm

where the first term on the right-hand side is the shear production and the second term on the righthand side is the buoyancy production (negative for stable stratification). The last term on the right-hand side is the dissipation of turbulent kinetic energy into heat. The only term in the TKE equation that is not closed is the vertical transport term F(k) for the parametrization of which many different suggestions have been made. More difficult is the determination of the dissipation rate e. Models solving equations for k and e are called k e models. Often, instead of an e-equation, dynamic equations for a quantity of the form kmen are constructed in a similar way as done for the e-equation. Once k and kmen are calculated, e can directly be determined. A well-known example for kmen equations is the Mellor–Yamada model with m ¼ 5/2 and n ¼ 1, where the kmen-equation is formulated as a kL-equation with the turbulence integral length scale L:

k2 e

½8

Sm and Sh are nondimensional stability functions. Several methods for calculating k and e have been suggested. For the TKE, k, a dynamic equation can be driven by means of Reynolds decomposition of the NSE. This equation is of the following form: @k @ þ FðkÞ ¼ Km M2 Kh N 2 e @t @z

½9

Coupling to Planktonic Ecosystem Eulerian Approach

Initial ecosystem models were slab models (0-D) with no or specified exchange with a reservoir below. They were followed by 0 þ -D models where the thickness of the mixed layer was prescribed over an annual cycle and exchange of selected properties with a reservoir below was allowed, usually as a fraction of the volume of the layer per day. The Eulerian formulation for the planktonic ecosystem including its supporting nutrients and light requires that the planktonic groups are small enough, passive enough, and numerous enough to satisfy a ‘continuum hypothesis’, that is, that they are transported (advected, diffused, and mixed) by the fluid environment in the same manner as a dissolved



substance (salt) or as an ‘intensive’ property of the fluid (temperature, or more correctly heat). Each component of the planktonic ecosystem Ci can be represented by a general conservation/transportation equation @Ci @ @Ci þ wi Ci Kh @t @z @z ¼ SourcesðCi Þ sinksðCi Þ

½11

where wi represents the vertical swimming speed or sinking/rising speed due to buoyancy difference with the fluid, Kh represents the vertical turbulent diffusion, and terms on the right-hand side represent the time rate of sources minus sinks due to ecological processes. The ‘ecological’ scientist on the other hand is concerned mostly with the sources and sinks for each component of the planktonic ecosystem Ci which can be represented in each vertical layer by a general mass conservation equation dCi ¼ Gki ðCi Þ Lik ðCi Þ dt

½12

where the Gki(Ci) terms represent the flux of mass from component Ck to component Ci and Lij(Ci)

represents the flux of mass lost from component Ci to component Cj due to ‘ecological’ processes. Ecological processes include growth of phytoplankton involving availability of light and uptake of various nutrients, loss of phytoplankton due to mortality or sinking, loss of phytoplankton due to grazing/feeding by zooplankton, etc. A gain by one component Ci usually represents a loss by one or more other components Cj, and vice versa. Each of the terms represents an intrinsic transfer rate and dependences on the components gaining and losing mass, representing 1 or more parameters to which values must be assigned. Components may be nutrients, planktonic functional types, dissolved organic matter, stages of plants or animals, etc. Originally, simple models were nutrient–phytoplankton–zooplankton (NPZ), but now multiple functional groups are standard as shown schematically in Figure 1. Zooplankton can also be represented by separate weight and number (weight number ¼ biomass) equations. Multiple stages or a continuous stage variable can also be employed. Within a mixed layer model, the ecosystem equations can be treated in the same way as equations for other passive scalars temperature and salinity. During each time step, the right-hand side of eqn [12] is evaluated and then vertical mixing is applied as for

PL

PS PS

Small phytoplankton

PL

Large phytoplankton i.e., diatoms

Ni

Nitrate

Na

Ammonium

D

Sinking detritus

Z1

Microzooplankton

Z2

Mesozooplankton

211

Na

Z2

Z1

Ni

D

Entrainment + mixing

Detritus Aggregates

Sinking

Fecal pellets

Figure 1 Schematic diagram of a contemporary planktonic ecosystem model embedded in the GOTM mixed layer model and in a Mellor–Yamada 2.5 mixed layer model. Downward arrows represent sinking fluxes and the ‘entrainment þ mixing’ upward arrow represents input of the nutrient nitrate via restoring to a constant or prescribed concentration over the bottom two layers. The model is also coupled to inorganic carbon, oxygen, and silica models via fixed ‘Redfield ratios’. Reproduced from Denman K, Voelker C, Penã A, and Rivkin R (2006) Modelling the ecosystem response to iron fertilization in the subarctic NE Pacific: The influence of grazing, and Si and N cycling on CO2 drawdown. Deep-Sea Research II 53 (doi:10.1016/j.dsr2.2006.05.026), with permission from Elsevier.


212


temperature and salinity. In physical mixed layer models, there is an efficient implicit algorithm used to mix temperature and salinity that can be called each time step to mix the ecosystem variables Ci. Animals and some plants with significant directed motility and/or development cannot usually be adequately represented by Eulerian representation and instead a Lagrangian or individual based model (IBM) is employed where individual organisms or groups are followed in the flow field. Hybrid models have also been used where nonmotile organisms are treated in an Eulerian model while fish or some zooplankton are treated in an IBM or Lagrangian model and are able to graze on their food sources. Lagrangian Approach

Many planktonic organisms are motile and undergo a daily vertical migration. For example, zooplankton can migrate toward the sea surface to feed during nighttime, and dinoflagellates (phytoplankton) can migrate downward to obtain nutrients during nighttime. When the time history of each organism is important, an Eulerian approach that treats the state of organisms only as a function of space is not an appropriate representation for realistic ecosystem dynamics. Then we need to follow each individual organism as a ‘particle’, in space and time, which is the Lagrangian approach. Each ‘particle’ can ‘remember’ the time history of its biological state and exhibit specific behavior patterns that are prescribed by a certain behavior model. The location of the ith particle at time t, Zi(t), is advanced each time step by both physical, WP(Zi(t), t), and biological, WB(Zi(t), t), processes. Over the time span t, its position becomes Zi ðt þ tÞ ¼ Zi ðtÞ þ WP ðZi ðtÞ; tÞ þ WB ðZi ðtÞ; tÞ

½13

Unless vertical advection is important, WP in a vertical one-dimensional model is due mostly to turbulent diffusion. WB is a behavior term that controls the swimming or sinking vector of each particle. When turbulent diffusion is spatially homogenous, WP can be expressed in terms of a constant eddy diffusivity, Kh. Each step is drawn from a Gaussian distribution: WP ðZi ðtÞ; tÞ ¼ Nð0; 2Kh tÞ

½14

where N is a Gaussian distribution with mean zero and variance 2Kht. Thus, the random step size is drawn from a unique Gaussian distribution. When turbulent diffusion varies in space, special attention is required to avoid unrealistic particle aggregation

in regions of low diffusivity: WP ðZi ðtÞ; tÞ ¼

dKh ðZi ðtÞÞ t dz þ N 0; 2Kh Zi ðtÞ 1 dKh ðZi ðtÞÞ þ t t 2 dz

½15

The first term on the right-hand side is a correction to avoid artificial aggregation; the spatial variability also affects where the diffusivity should be evaluated. These corrections are for numerical reasons, so no corresponding physical processes exist. Since the corrections require that the first derivative of Kh be a continuous function, special attention is needed where the derivative changes abruptly at a grid point for closure models. Another artificial accumulation may occur where the model diffusivity approaches zero. To avoid this problem, the following effective diffusivity, Keff(z), should be used: Keff ðzÞ ¼ Kh ðzÞ þ KBG

½16

where KBG is a constant background diffusivity. For phytoplankton diffusion, 10–6 m2 s–1 is a good value to use. For the biological step, WB includes effects from simple sinking to behavior-driven complex motion, such as migration and grouping behavior. When the behavior is preconditioned, WB is straightforward to implement. But organisms exhibit complex adaptive behavior; the model should take this adaptive aspect into consideration.

Case Studies Ocean Station P

Ocean Station Papa (OSP, 501 N, 1451 W) in the NE subarctic Pacific was occupied by a weather ship more or less continuously from the early 1940s until June 1981. Because of this long time series, about half a dozen intensive international studies have taken place at OSP, from the mixed layer experiment (MILE) in the early 1980s up to the recent 2002 Subarctic Ecosystem Response to Iron Fertilization Study (SERIES). Most mixed layer models have been applied to OSP, both because of the more than 50 years of observations and a belief that currents are sluggish in the open NE Pacific allowing the mixed layer to be essentially in a one-dimensional balance. Physics The suitability of station P for testing onedimensional models results from the clear annual



213

SST (°C)

14 10 6 2 0

1

2

3

4

5

0

1

2

3

4

5

0

1

2

3

4

5

MLD (m)

0 50 100 150 200

Q (109 Jm−2)

3 2 1 0 −1 −2

t (y) since 1 January 1961 Figure 2 Mixed layer physics at station P during a 5-year period. Top: sea surface temperature, middle: mixed layer depth, bottom: heat content of upper 250 m relative to 1 January 1961. From Burchard H (2002) Lecture Notes in Earth Sciences, Vol. 100: Applied Turbulence Modelling in Marine Waters, 229pp. New York: Springer.

cycle of surface fluxes, sea surface temperature (SST) and mixed layer depth, as well as from the assumptions of small horizontal advection and a closed heat budget. This annual cycle is shown in Figure 2 for five subsequent years. An SST minimum of typically 5 1C is reached in early spring, roughly coinciding with the maximum mixed layer depth of 100–150 m and a minimum heat content of the upper 250 m. Then, the surface mixed layer evolves in intermittent steps, switching between periods of stable thermal stratification and suppressed vertical mixing due to a net heat flux into the water column, and periods of increased mixing during episodic strong wind events. The SST maximum of around 14 1C is reached during a short period in late summer, when the mixed layer depth is reduced to a few meters and the heat content reaches a maximum. In autumn, net cooling causes convective mixing with erosion of the thermocline and subsequent deepening of the mixed layer. Simulations with a two-equation turbulence closure model employing an algebraic second-moment closure reveal that temperature structures at the base of the surface mixed layer is not sufficiently reproduced by such models, probably due to the neglect of mixing induced by internal wave activity (Figure 3). The SST is however well reproduced by such models during spring and summer. During autumn and winter, however, advection events, such as freshwater

fluxes and Ekman pumping, which are not parametrized here, prohibit complete agreement between observations and model simulations of SST (Figure 4). Planktonic ecosystem OSP is located in one of the three high-nitrogen-low-chlorophyll (HNLC) regions globally where phytoplankton production and biomass are limited by the micronutrient iron, rather than by nitrogen or phosphorous. HNLC regions are characterized by low-standing stocks of phytoplankton chlorophyll with no pronounced spring bloom and relatively high concentrations of nitrate, even throughout the summer productive season. A series of planktonic ecosystem models have been applied to OSP, first to understand factors reducing any spring bloom, and more recently to determine the major factors controlling the response of the ecosystem to changes in the micronutrient iron. Figure 5 shows an annual cycle from the ecosystem model with two classes of phytoplankton (shown schematically in Figure 1) embedded in a mixed layer model. In this simulation each phytoplankton class had a different constant degree of iron limitation, with the large phytoplankton (diatoms) having a greater limitation. One-dimensional ecosystem models are successful in reproducing the extent and timing of the ecosystem response to iron fertilization at OSP – provided that



Measurements OWS Papa 1961/62

T (°C)

Simulation OWS Papa 1961/62

18 17 16 15 14 13 12 11 10 9 8 7 6 5 4

Depth (m)

0

0

90 120 150 180 210 240 270 300 330 360 390 420

T (°C) 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4

Depth (m)

214

90 120 150 180 210 240 270 300 330 360 390 420

Julian Day 1961/62

Julian Day 1961/62

Figure 3 Comparison between observed (left) and simulated (right) temperature at station P for the period March 1961–March 1962. The simulations were carried out using a two-equation turbulence closure model with an algebraic second-moment closure. From Burchard H (2002) Lecture Notes in Earth Sciences, Vol. 100: Applied Turbulence Modelling in Marine Waters, 229pp. New York: Springer.

16 Observation Simulation

SST (°C)

14 12 10 8 6 4 100

150

200

250 300 Julian Day 1961/62

350

400

450

Figure 4 Comparison between observed (dotted line) and simulated (thin line) sea surface temperature (SST) at station P for the period March 1961–March 1962. The simulations have been carried out with a two-equation turbulence closure model with an algebraic second-moment closure.

the silica:nitrate uptake ratio by the diatoms varies inversely with in situ iron concentration. This dependence of Si:N uptake ratio provides a delay in the peak diatom biomass (and silica depletion) and results in better agreement with observations. Use of a cell quota model for iron uptake can delay the peak response even further. Convective Mixing and Internal Wave Mixing

One-dimensional models parametrize convective mixing driven by surface cooling based on laboratory results that the buoyancy flux through the base of the mixed layer as a result of penetrative convection is about 20% of the surface buoyancy flux, after removal of the influence of wind. These models are relatively successful in reproducing observations of penetrative convection (Figure 6). Internal wave mixing at the base of the mixed layer, on the other hand, is highly variable and in general cannot be modeled as a constant background vertical diffusion coefficient.

Perspectives The physical structure of the surface mixed layer is far more complicated than what one-dimensional models can reproduce. The major reason for this is the presence of surface and internal waves. Surface waves interact in a complex way with near-surface currents and turbulence. The interaction between surface-wave-induced Stokes drift and current shear drives Langmuir circulation, which is basically counter-rotating vortices aligned with the wind direction and penetrating the whole mixed layer depth. These coherent structures, strongly modified by large and energetic turbulent eddies, provide additional vertical homogenization of the mixed layer and cyclically transport phytoplankton through the euphotic zone, affecting photoadaptation. Furthermore, breaking surface waves inject turbulence into the near-surface zone, generating balances far away from a constant stress layer, with strong implications for the vertical transport of matter, solutes, and energy exchanged with the atmosphere. Internal wave



215

Small phytoplankton

(a)

0.4

Depth (m)

0.3

0.2

0.1

1 Apr.

1 Jul.

1 Oct.

1 Jan.

0

Large phytoplankton

(b)

0.4

Depth (m)

0.3

0.2

0.1

1 Apr. (c)

1 Jul.

1 Oct.

1 Jan.

0

1.0 Diatoms PL Small phyto Ps

0.8 0.6 0.4 0.2 0.0

100

200 Year day

300

Figure 5 Annual cycle at OSP of: (a) small ‘background’ phytoplankton, (b) larger ‘diatoms’, and (c) their average concentrations over the upper 50 m, from the model shown in Figure 1, embedded in a Mellor–Yamada 2.5 mixed layer model. The annual cycle of the base of the mixed layer can be seen in (a) from color contrast. Color bars are in mmol–N m–3. Reproduced from Denman K, Voelker C, Penã A, and Rivkin R (2006). Modelling the ecosystem response to iron fertilization in the subarctic NE Pacific: The influence of grazing, and Si and N cycling on CO2 drawdown. Deep-Sea Research. II 53 (doi:10.1016/j.dsr2.2006.05.026), with permission from Elsevier.

dynamics determine the turbulence level below the mixed layer in the thermocline, with substantial impact on diapycnal nutrient fluxes into the euphotic zone. Surface waves, Langmuir circulation, and internal waves may also influence each other. For all these processes, we need to develop improved parametrizations for use in one-dimensional models. Horizontal variability in physical and biological processes, which cannot be resolved by the onedimensional models, is a ubiquitous feature in the

ocean. One-dimensional mixed layer parametrizations are, however, relevant for such two- and threedimensional flows and related biology. For example, oceanic fronts and eddies involve horizontal structures, which can strongly enhance interaction between the adiabatic ocean interior and the diabatic surface boundary layer. Boundary layer mixing plays a crucial role in such interactions and in resulting water mass formation, subduction, upwelling, and biology near eddies and fronts. Therefore, successful


216


Time of day (h) 0

3

6

0

27.8

(a) −2 −4

27.6 −6 −8 27.4

−10 −12

27.2

0 −2

(b)

−4

27

−6 −8

26.8

−10 −12 Depth (m)

26.6 −4

(c)

Julian Date −6.5

−6 −8 −10

−7 −12

−4

(d) −7.5

−6 −8 −10

−8 −12 242.9

242.95

243

243.05

243.1 Julian date

243.15

243.2

243.25

243.3

243.35

Figure 6 Comparison between the Mellor–Yamada closure model (a, c) and observed data (b, d). (a) and (b) show temperature; (c) and (d) are TKE dissipation rate (W kg–1). Modified from Nagai T, Yamazaki H, Nagashima H, and Kantha LH (2005) Field and numerical study of entrainment laws for surface mixed layer. Deep Sea Research II 52: 1109–1132.



modeling of fully complex three-dimensional flow evolves naturally from one-dimensional models.

See also Okhotsk Sea Circulation. Open Ocean Convection.

Further Reading Baumert HZ, Simpson JH, and Dermann JS (eds.) (2005) Marine Turbulence. Theories Observations and Models. Cambridge: Cambridge University Press. Burchard H (2002) Lecture Notes in Earth Sciences, Vol. 100: Applied Turbulence Modelling in Marine Waters, 229pp. New York: Springer. Denman KL (2003) Modelling planktonic ecosystems: Parameterizing complexity. Progress in Oceanography 57: 429--452. Denman K, Voelker C, Penã A, and Rivkin R (2006) Modelling the ecosystem response to iron fertilization in the subarctic NE Pacific: The influence of grazing, and Si and N cycling on CO2 drawdown. Deep-Sea Research II 53 (doi:10.1016/j.dsr2.2006.05.026). Kantha LH and Clayson CA (2000) International Geophysics Series, Vol. 67: Small-Scale Processes in Geophysical Fluid Flows. New York: Academic Press. Monahan AH and Denman KL (2004) Impacts of atmospheric variability on a coupled upper-ocean/ ecosystem model of the subarctic Northeast Pacific.

217

Global Biogeochemical Cycles 18: GB2010 (doi: 10.1029/2003GB002100). Large WG, McWilliams JC, and Doney SC (1994) Oceanic vertical mixing: A review and a model with nonlocal boundary layer parameterisation. Reviews of Geophysics 32: 363--403. Nagai T, Yamazaki H, and Kamykowski D (2003) A Lagrangian photoresponse model coupled with 2nd order turbulence closure. Marine Ecology Progress Series 265: 17--30. Nagai T, Yamazaki H, Nagashima H, and Kantha LH (2005) Field and numerical study of entrainment laws for surface mixed layer. Deep Sea Research II 52: 1109--1132. Thorpe SA (2005) The Turbulent Ocean, 439pp. Cambridge, UK: Cambridge University Press. Umlauf L and Burchard H (2005) Second-order turbulence closure models for geophysical boundary layers. A review of recent work. Continental Shelf Research 25: 795--827. Yamazaki H, Mackas D, and Denman K (2002) Coupling small scale physical processes with biology. In: Robinson AR, McCarthy JJ, and Rothschild BJ (eds.) The Sea: Biological–Physical Interaction in the Ocean, ch. 3, pp. 51--112. New York: Blackwell.

Relevant Websites http://www.gotm.net – General Ocean Turbulence Model, GOTM.


OPEN OCEAN CONVECTION A. Soloviev, Nova Southeastern University, FL, USA B. Klinger, Center for Ocean-Land-Atmosphere Studies (COLA), Calverton, MD, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2015–2022, & 2001, Elsevier Ltd.

Introduction Free convection is fluid motion due to buoyancy forces. Free convection, also referred to as simply convection, is driven by the static instability that results when relatively dense fluid lies above relatively light fluid. In the ocean, greater density is associated with colder or saltier water, and it is possible to have thermal convection due to the vertical temperature gradient, haline convection due to the vertical salinity gradient, or thermohaline convection due to the combination. Since sea water is about 1000 times denser than air, the air–sea interface from the waterside can be considered a free surface. So-called thermocapillary convection can develop near this surface owing to the dependence of the surface tension coefficient on temperature. There are experimental indications that in the upper ocean layer more than 2 cm deep, buoyant convection dominates. Surfactants, however, may affect in the surface renewal process. This article will mainly consider convection without these capillary effects. Over most of the ocean, the near-surface region is considered to be a mixed layer in which turbulent mixing is stronger than at greater depth. The strong mixing causes the mixed layer to have very small vertical variations in density, temperature, and other properties compared to the pycnocline region below. Convection is one of the key processes driving mixed layer turbulence, though mechanical stirring driven by wind stress and other processes is also important. Therefore, understanding convection is crucial to understanding the mixed layer as well as property fluxes between the ocean and the atmosphere. Thermal convection is associated with the cooling of the ocean surface due to sensible (QT ), latent (QL ), and effective long-wave radiation (QE ) heat fluxes. QT may have either sign; its magnitude is, however, much less than that of QE or QL (except perhaps in some extreme situations). The top of the

218

water column becomes colder and denser than the water below, and convection begins. In this way, cooling is associated with the homogenization of the water column and the deepening of the mixed layer. Warming due to solar radiation occurs in the surface layer of the ocean and is associated with restratification and reductions in mixed layer depth. The most prominent examples of this mixing/restratification process are the diurnal cycle (nighttime cooling and daytime warming) and the seasonal cycle (winter cooling and summer warming). There are also important geographical variations in convection, with net cooling of relatively warm water occurring more at higher latitudes and a net warming of water occurring closer to the Equator. For this reason, mixed layer depth generally increases towards the poles, though at very high latitudes ice-melt can lower the surface salinity enough to inhibit convection. Over most of the ocean, annual average mixed layer depths are in the range of 30–100 m, though very dramatic convection in such places as the Labrador Sea, Greenland Sea, and western Mediterranean Sea can deepen the mixed layer to thousands of meters. This article discusses convection reaching no deeper than a few hundred meters. Dynamically, the convection discussed here differs from deep convection in being more strongly affected by surface wind stress and much less affected by the rotation of the Earth. Convection directly affects several aspects of the near-surface ocean. Most obviously, the velocity patterns of the turbulent flow are influenced by the presence of convection, as is the velocity scale. The convective velocity field then controls the vertical transport of heat (or more correctly, internal energy), salinity, momentum, dissolved gases, and other properties, and the vertical gradients of these properties within the mixed layer. Convection helps to determine property exchanges between the atmosphere and ocean and the upper ocean and the deep ocean. The importance of convection for heat and gas exchange has implications for climate studies, while convective influence on the biologically productive euphotic zone has biological implications as well.

Phenomenology The classical problem of free convection is to determine the motion in a layer of fluid in which the top surface is kept colder than the bottom surface. This is an idealization of such geophysical examples as an


OPEN OCEAN CONVECTION

ocean being cooled from above or the atmosphere being heated from below. The classical problem ignores such complications as wind stress on the surface, waves, topographic irregularities, and the presence of a stably stratified region below the convection region. The study of convection started in the early twentieth century with the experiments of Benard and the theoretical analysis of Rayleigh. One might expect that heavier fluid would necessarily exchange places with lighter fluid below as a result of buoyancy forces. This happens by means of convective cells or localized plumes of sinking dense fluid and rising light fluid. However, such cells or plumes are retarded by viscous forces and are also dissipated by thermal diffusion as they sink into an environment with a different density. When the buoyancy force is not strong enough to overcome the inhibitory effects, the heavy-over-light configuration is stable and no convection forms. The relative strengths of these conflicting forces is measured by the Rayleigh number, a nondimensional number given by eqn [1]. Ra ¼

gaDTh3 ðkT vÞ

½1

Here g is the acceleration of gravity, a is the thermal expansion coefficient of sea water (a ¼ 2:6 104 1C1 at T ¼ 201C and S ¼ 35 PSU), DT is the temperature difference between the top and bottom surfaces, h is the convective layer thickness, and m and kT are the molecular coefficients of viscosity and thermal diffusivity, respectively (n ¼ 1:1 106 m2 s1 and kT ¼ 1:3 107 m2 s1 at T ¼ 201C and S ¼ 35PSU). The term DT ¼ Dr=r represents the fractional density difference between top and bottom. Convection occurs only if Ra is greater than a critical value, Racr , which depends somewhat on geometrical and other details of the fluid. For the classical problem of water bounded above and below by solid plates, Racr ¼ 657. For sea water under typical conditions, even a temperature difference of 0.11C makes Ra > Racr as long as h is greater than a centimeter. For Ra > Racr, the Rayleigh number still serves a useful purpose as a guide to the nature of the convective activity (though the problem also depends on the Prandtl number, Pr ¼ n=kT ). For a fixed Pr and for Ra only slightly larger than Racr , motion occurs in regular, steady cells. As Ra is increased, the motion becomes time-dependent. Regular oscillations occur, and these increase in number and frequency for higher Ra. At sufficiently high Ra, the flow is turbulent and intermittent. The value of Ra in the ocean

219

is very large (typically greater than 1014 for a temperature difference of 0.11C over 10 m), so convection is usually turbulent. Turbulent convection is usually characterized by the formation of descending parcels of cold water. In laboratory experiments, it has been found that water from the cooled surface layer collects along lines, producing thickened regions that become unstable and plunge in vertical sheets (Figure 1). In analogy to atmospheric convection, we will here call these parcels thermals, although — in contrast to the atmosphere — in the ocean they are colder than the surrounding fluid. In 1966, Howard formulated a phenomenological theory that represented turbulent convection as the following cyclic process. The thermal boundary layer forms by diffusion, grows until it is thick enough to start convecting, and is destroyed by convection, which in turn dies down once the boundary layer is destroyed. Then the cycle begins again. This phenomenological theory has implications for the development of parametrizations for the air–sea heat and gas exchange under low wind speed conditions (see later). The descending parcels of water have a mushroomlike appearance. In the process of descending to deeper layers, the descending parcels developing as a result of the local convective instability of the thermal molecular sublayer join and form larger mushroomlike structures. The latter descend faster and eventually form bigger structures. This cascade

Figure 1 Orthogonal views of convective streamers in warm water that is cooling from the surface. The constantly changing patterns appear as intertwining streamers in the side view. (From Spangenberg WG and Rowland WR (1961)) Convective circulation in water induced by evaporative cooling. Physics of Fluids 4: 743–750. ^ 1961 American Institute of Physics.


220


process produces a hierarchy of convective scales, which is illustrated in Figure 2 on the example of the haline convection.

Penetrative Convection The unstable stratification of the mixed layer is usually bounded below by a stratified pycnocline. One can imagine the mixed layer growing in depth with thermals confined to the statically unstable depth range. Suppose the density at the top of the pycnocline is r1 (Figure 3A). As surface buoyancy loss and convection increase the average density of the mixed layer, the mixed layer density increases to r2 , which is slightly denser than r1 (Figure 3B). The static instability now allows convection to act on the pycnocline down to density r2 (Figure 3C), so that the mixed layer grows at the expense of the pycnocline. This is known as nonpenetrative convection. In reality, the largest thermals acquire enough kinetic energy, as they fall through the mixed layer, that they can overshoot the base of the mixed layer, working against gravity. This is penetrative

convection. The penetrative convection produces a countergradient flux that is not properly accounted for if we model convective mixing as merely a very strong vertical diffusion. Unlike the smooth density profile at the base of a mixed layer that is growing by nonpenetrative convection (Figure 3C), penetrative convection is characterized by a density jump at the base of the mixed layer (Figure 3D). The cooling of the ocean from its surface is compensated by the absorption of solar radiation. The latter is a volume source for the upper meters of the ocean. The thermals from the ocean surface, as they descend deeper into the mixed layer, produce heat flux that is compensated by the volume absorption of solar radiation. This is another type of the penetrative convection in the upper ocean, which will be considered in more detail in a later section.

Relative Contributions of Convection and Shear Stress to Turbulence For the limiting case in which the only motion in the mixed layer is due to convection, there are simple estimates of average speed and temperature fluctuations associated with the plumes. When the Rayleigh number is high enough that the flow is fully turbulent, the plume characteristics should be largely independent of the viscosity and diffusivity throughout most of the mixed layer. In that case, ignoring the Earth’s rotation and influences from the pycnocline, the governing parameters of the system are simply the mixed layer depth h and the surface buoyancy flux B0 . B0 is based on the surface heat fluxes according to eqn [2], where r is the water density, cp is the specific heat capacity of water (E4 103 Jkg1 K1 ), L is the latent heat released by evaporation (E2:5 106 Jkg1 ), S is the surface salinity, and b is the coefficient of salinity expansion (b ¼ 7:4 104 PSU1 at T ¼ 201C and S ¼ 35PSU). h i 1 B0 ¼ gr1 ac1 ð Q þ Q þ Q Þ þ bQ L S S E L L p

Figure 2 Shadowgraph picture of the development of secondary haline convection. From Foster TD (1974) The hierarchy of convection. Colloques Internationaux du CNRS N215. Processus de Formation Des Eaux Oceaniques Profondes, pp. 235–241. ^ 1974 Centre National de la Recherche Scientifique.

½2

The first term in the square bracket in the right side of [2] relates to the buoyancy flux due to surface cooling; the second term relates to the buoyancy flux due to the surface salinity increase because of evaporation. Given all the above restrictions, the velocity scale, o * , is then given by the Priestly formula ([3]). w * ¼ ðB0 hÞ1=3

½3

This is the only combination of B0 and h that will give the proper units for velocity. Similarly, if we



(A) Original density

(B) After cooling

Depth

Mixed layer

Depth

221

Pycnocline

1

Statically unstable

1

2

Density

Density (D) Penetrative

Depth

Depth

(C) Nonpenetrative

1

2

1

2

Density

Density Figure 3 Schematic diagram of nonpenetrative and penetrative convection.

define the buoyancy to be b ¼ gDr=r, the buoyancy scale, b * , is given in [4] 1=3 b * ¼ B2 =h

½4

Laboratory experiments have shown that these scales are in good agreement with actual fluctuations during convection. For typical oceanic parameters (for instance, heat flux of Q0 ¼ 100Wm2 and h ¼ 100 m), o * is a few centimeters per second and

b * is equivalent to temperature fluctuations of about 0.011C. Two major sources of turbulent kinetic energy in the upper ocean are the wind stress and buoyant convection. Upper ocean convection is usually accompanied by near-surface currents induced by wind and wind waves. The near-surface shear is then an additional source of near-surface turbulent mixing. In the 1950s, Oboukhov proposed the buoyancy length scale, LO ¼ ku3* =B0 , where k is the Von Karman constant (k ¼ 0:4), B0 is the surface


222


buoyancy flux (e.g., defined by [2]), and u * is the boundary layer velocity scale (friction velocity) defined as u * ¼ ðt=rÞ1=2 , where t represents the bottom stress in the atmospheric case and the windstress in the oceanic case (r is the density of air or water, respectively). Later, Monin and Oboukhov suggested the stability parameter, x ¼ z=LO (where z is the height in the atmosphere or the depth in the ocean), to characterize the relative importance of shear and buoyant convection in the planetary boundary layer. Experimental studies conducted in the atmospheric boundary layer show that at xo 0:1 the flow is primarily driven by buoyant convection. From the analogy between the atmospheric and oceanic turbulent boundary layers, the Monin–Oboukhov theory is often applicable to the analysis of the oceanic processes as well. In particular, it provides us a theoretical basis to separate the layers of free and forced convection in the upper ocean turbulent boundary layer. For a 5 m s1 wind speed and Q0 ¼ 100Wm2 , the Oboukhov scale is LO B15 m. This means that the shear-driven turbulent flow is confined within a 1.5 m thick near-surface layer of the ocean. In a 50 m deep mixed layer, 97% of its depth will be driven by the buoyant convection during nighttime, with the rate of dissipation of turbulent kinetic energy there about equal to the surface buoyancy flux, B0 , as shown by Shay and Gregg.

Convection and Molecular Sublayers Convection is driven by the horizontal–mean vertical density gradient. At high Ra, typical vertical velocities are much lower near the top and bottom boundaries than they are in the bulk of the water column. Since the vertical density gradient is reduced by the convective motion, the velocity distribution causes most of the vertical density gradient to occur near the boundaries. Indeed, under low-wind, lowwave conditions in which convection dominates, the mixed layer temperature gradient is largely confined to a region only about 1 mm deep. Because the vertical heat flux at the base of the convection region is typically much smaller than at the surface, the large temperature gradient only occurs at the surface, where this thermal sublayer is often referred to as the cool skin. The temperature jump across the cool skin can be related to the vertical flux of heat at the air–sea interface and constants of molecular viscosity and heat diffusion in water using convection laws. The vertical heat flux, Q0 , can be written in nondimensional form as the Nusselt number ([5]).

Nu ¼ Q0 =cr r =ðkT DT=hÞ

½5

In [5] the heat flux is normalized by the heat flux due to vertical diffusion in the absence of convection. This quantity must be a function of the given nondimensional parameters of the system, which, for thermal convection in the absence of other driving mechanisms, are just Ra and the Prandtl number Pr (here we ignore the Earth’s rotation and entrainment from the pycnocline). A further simplifying assumption is that for high Ra (greater than 107), typical of the mixed layer, the convection is fully turbulent and does not depend on the mixed layer thickness, h, which implies [6], where AðPrÞ is a dimensionless coefficient depending on Prandtl-number (according to laboratory measurements, AE0:16 0:25). Nu ¼ AðPrÞRa1=3

½6

Given the definitions of Ra and Nu, this relation can be rearranged to yield the temperature difference across the cool skin, DT, as a function of the surface heat flux, Q0 ¼ QL þ QE þ QT ([7]). 1=4 3=4 Q0 =cp r DT ¼ A3=4 agk2T =v

½7

DT is 0.2–0.41C under typical oceanic conditions but can be as much as 11C in regions of very high heat loss to the atmosphere (e.g., Gulf Stream at high latitudes). While the term ‘sea surface temperature’ (SST) is often used to represent the temperature of the mixed layer as a whole, the existence of a cool skin means that the temperature of the literal surface of the ocean can be somewhat lower than the rest of the mixed layer. Satellite measurements of SST are based on infrared emissions from a thin layer of several micrometers, so that these measurements can be somewhat different from ship-based ‘surface’ measurements, which are generally based on sampling within several upper meters of the ocean. Indeed, while the first experimental evidence of the cool skin was obtained in the 1920s, the phenomenon was not widely recognized by the oceanic community until sophisticated methods, including remote sensing by infrared techniques, had been gradually helping to incorporate the cool skin into modern oceanography. The accuracy of current satellite remote sensing techniques is, nevertheless, still below that level at which the cool skin becomes of crucial importance. The effect of the cool skin on the heat exchange between ocean and atmosphere is also basically below the resolution of widely used bulk flux algorithms. However, one interesting practical application of the cool skin phenomenon emerged in the



1990s. Similar laws govern the thermal sublayer of the ocean (the cool skin) and diffusive sublayers associated with air–sea gas exchange. Such gas exchange is a key biogeochemical variable, and for greenhouse gases such as CO2 is of climatological importance as well. The rate at which gases cross the air–sea interface is measured by the piston velocity, K. Boundary layer laws relate K to DT in the eqn [8]. A0 Q0 ðm=vÞ1=2 K¼ cp rDT

½8

A0 is a dimensionless constant (E1.85) and m is the molecular gas diffusion coefficient in water (m ¼ 1:6 109 m2 s1 for CO2 at T ¼ 201C and S ¼ 35PSU). The more readily available cool skin data can then be used for an adjustment of the gas transfer parametrization. The convective parametrizations for the cool skin and air–sea gas exchange are valid within the range of wind speed from 0 to 3–4 m s1. Under higher wind speed conditions, the cool skin and the interfacial air–sea gas exchange are controlled by the

wind stress and surface waves. The transition is observed when the surface Richardson number, Rf0i ¼ agQ0 =ðcp ru4 Þ, reaches a value of approxi* mately 1.5 105 (here r is the water density).

Diurnal and Seasonal Cycles of Convection For much of the year, much of the ocean experiences a cycle of daytime heating and nighttime cooling that leads to a strong diurnal cycle in convection and mixed layer depth. Such behavior is illustrated in Figure 4. At night, when there is cooling, the convective plumes reach the base of the mixed layer, which deepens as the mixed layer grows colder and denser. During the day, convection is inhibited within the bulk of the mixed layer but may still occur near the surface of the mixed layer, even if the mixed layer experiences a net heat gain. This is because the vertical distribution of cooling and heating are somewhat different. Heat loss is dominated by latent heat flux associated with evaporation and hence this forcing occurs at the top surface. Heat gain is

2

4

0

0

bold =

0 Jq

0 _

_4

_2

10 2 J q /W m

_2

_1 7

0

10 J b / W kg

223

_8

_4 0 0.1

p/MPa

0.2 0.3 0.4 0.5

0.6

_

D

_1

= 10 8 w kg

0.7 13

14

15

16

17

19 18 October 1986

20

21

22

23

Figure 4 Diurnal cycles in the outer reaches of the California Current (341N, 1271W). Each day the ocean lost heat and buoyancy starting several hours before sunset and continuing until a few hours after sunrise. These losses are shown by the shaded portions of the surface heat and buoyancy fluxes in the top panel. In response, the surface turbulent boundary layer slowly deepened (lower panel). The solid line marks D, the depth of the surface turbulent boundary layer, and the lightest shading shows 108 W kg1oso107 W kg1 where s is the dissipation rate of the turbulent kinetic energy. The shading increases by decades, so that the 0=mn ¼ B0 , and darkest shade is s >105 W kg1. Note that 1 MPa in pressure p corresponds to approximately 100 m in depth, Jb Jq0 ¼ ðQ0 þ QR Þ, where QR is the solar radiation flux penetrating ocean surface. (From Lombardo CP and Gregg MC (1989) Similarity scaling during nighttime convection. Journal of Geophysical Research 94: 6273–6274.) ^ 1989 American Geophysical Union.


224


dominated by solar radiation that is absorbed by the water over a range of depths that can extend tens of meters in many parts of ocean. For example, one can have surface heat loss of 100 W m2 occurring at the surface and net radiative heat gain of 500 W m2 distributed over the top 30 m of the ocean. Calculating the rate of change of heat due to the forcing between the surface and a depth z, we find that there is actually heat loss for small z down to a depth known as the thermal compensation depth. Below this depth, the mixed layer restratifies and convection occurs only through the mechanism of penetrative convection. For most of the World Ocean, the thermal compensation depth is less than 1 m between sunrise and sunset. Usually, the rate of turbulent kinetic energy production in the mixed layer is dominated by the convective term at night, but by the wind stress term during most of the day. Because the thermal compensation depth is generally quite small, turbulent kinetic energy generated by convection makes no contribution to turbulent entrainment of water through the bottom of the mixed layer, which lies much deeper. Under low wind speed conditions and strong solar insolation, the thickness of the surface convective layer of the ocean may reduce to only several centimeters. In that case, convection in the upper ocean may be of a laminar or transitional nature. Stable stratification inhibits turbulent mixing below the relatively thin near-surface convection layer. Vertical mixing of momentum is confined to the shallow daytime mixed layer so that, during the day, flow driven directly by the wind stress is confined to a similarly thin current known as the diurnal jet. In the evening, when convection is no longer confined by the solar radiation effect, convective plumes penetrate deeper into the stratified part of the mixed layer, increasing the turbulent mixing of momentum at the bottom of the diurnal jet. The diurnal jet then releases its kinetic energy during a relatively short time. This process is so intensive that the releasing kinetic energy cannot be dissipated locally. As a result, a Kelvin–Helmholtz type instability is formed, which generates billows — a kind of organized structure. The billows intensify the deepening of the diurnal mixed layer. Although the energy of convective elements is relatively small, it serves as a catalyst for the release of the kinetic energy by the mean flow. In the equatorial ocean, the shear in the upper ocean is intensified by the Equatorial Undercurrent; the evening deepening of the diurnal jet is therefore sometimes so intense that it resembles a shock, which radiates very intense high-frequency internal waves in the underlying thermocline.

The diurnal cycle is often omitted from numerical ocean models for reasons of computational cost. However, the mixed layer response to daily-averaged surface fluxes is not necessarily the same as the average response to the diurnal cycle. Neglecting the diurnal cycle replaces periodic nightly convective pulses with chronic mixing that does not reach as deep. Open ocean convection is a mechanism effectively controlling the seasonal cycle in the ocean as well. Resolution of diurnal changes is usually uneconomical when the seasonal cycle is considered. Because of nonlinear response of the upper ocean to the atmospheric forcing, simply averaged heat fluxes cannot be used to estimate the contribution of the convection on the seasonal scale. The sharp transition between the nocturnal period, when convection dominates mixing in the surface layer, and the daytime period, when the sun severely limits the depth of convection, leaving the wind stress to control mixing, may simplify the design of models for the seasonal cycle of the upper ocean. Incorporation of convection adjustment schemes into the oceanic component of the global circulation models leads to an appreciable change of troposphere temperature in high latitudes, which affects the global ocean and atmosphere circulation. Parametrization of the convection on the seasonal and global scales is therefore an important task for the prediction of climate and its changes.

Conclusions Observation of the open ocean convection is a difficult experimental task. Though convective processes have been observed in several oceanic turbulence studies, most of our knowledge of this phenomenon in the ocean is based on the analogy between atmospheric and oceanic boundary layers and on laboratory studies. Many intriguing questions regarding the convection in the open ocean remain, however. Some of them, like the role of penetrative convection in mixed layer dynamics, are of crucial importance for improvement of the global ocean circulation modeling. Others, like the role of surfactants in the surface renewal process, are of substantial interest for studying the air–sea exchange and global balance of greenhouse gases like CO2.

See also Air–Sea Gas Exchange. Breaking Waves and NearSurface Turbulence. Deep Convection. NonRotating Gravity Currents. Photochemical



Processes. Rotating Gravity Currents. Satellite Remote Sensing of Sea Surface Temperatures. Single Compound Radiocarbon Measurements. Small-scale Patchiness, Models of. ThreeDimensional (3D) Turbulence. Upper Ocean Mixing Processes. Wind- and Buoyancy-Forced Upper Ocean.

Further Reading Busse FH and Whitehead JA (1974) Oscillatory and collective instabilities in large Prandtl number convection. Journal of Fluid Mechanics 66: 67--79. Foster TD (1971) Intermittent convection. Geophysical Fluid Dynamics 2: 201--217. Fru NM (1997) The role of organic films in air–sea gas exchange. In: Liss PS and Duce RA (eds.) The Sea Surface and Global Change, pp. 121--172. Cambridge: Cambridge University Press. Gregg MC, Peters H, Wesson JC, Oakey NS, and Shay TJ (1984) Intense measurements of turbulence and shear in the equatorial undercurrent. Nature 318: 140--144. Holland WR (1977) The role of the upper ocean as a boundary layer in models of the oceanic general circulation. In: Kraus EB (ed.) Modelling and Prediction of the Upper Layers of the Ocean. Oxford: Pergamon Press.

225

Katsaros KB (1980) The aqueous thermal boundary layer. Boundary-Layer Meteorology 18: 107--127. Kraus EB and Rooth CGH (1961) Temperature and steady state vertical heat flux in the ocean surface layers. Tellus 13: 231--238. Caldwell DR, Lien R-C, Moum JN, and Gregg MC (1997) Turbulence decay and restratification in the equatorial ocean surface layer following nighttime convection. Journal of Physical Oceanography 27: 1120--1132. Shay TJ and Gregg MC (1986) Convectively driven turbulent mixing in the upper ocean. Journal of Physical Oceanography 16: 1777--1791. Soloviev AV and Schluessel P (1994) Parameterization of the cool skin of the ocean and of the air-ocean gas transfer on the basis of modeling surface renewal. Journal of Physical Oceanography 24: 1339--1346. Thorpe SA (1988) The dynamics of the boundary layers of the deep ocean. Science Progress (Oxford) 72: 189--206. Turner JS (1973) Buoyancy Effects in Fluids. Cambridge: Cambridge University Press. Woods JD (1980) Diurnal and seasonal variation of convection in the wind-mixed layer of the ocean. Quarterly Journal of the Royal Meteorological Society 106: 379--394.


OPEN OCEAN FISHERIES FOR DEEP-WATER SPECIES J. D. M. Gordon, Scottish Association for Marine Science, Oban, Argyll, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2023–2030, & 2001, Elsevier Ltd.

Introduction Deep-water fisheries are considered to be those that exploit fish or shellfish that habitually live at depths greater than 400 m. With the exception of some localized line fisheries around oceanic islands, such as for Aphanopus carbo (black scabbardfish) at Madeira in the Atlantic and for Ruvettus in Polynesia, the fisheries are mostly of recent origin. The deepwater fisheries of the continental slopes only developed in the 1960s when Soviet trawlers discovered and exploited concentrations of roundnose grenadier (Coryphaenoides rupestris) off Canada and Greenland. At the same time, the Soviet fleet was exploiting similar resources in the North Pacific. Since then, despite a decline in Russian landings, deep-water fisheries have continued to expand on a global scale. Some of these fisheries target species that had never previously been exploited, such as orange roughy (Hoplostethus atlanticus). Some species have a depth range that extends from the shallow continental shelf depths into deeper water, and there is an increasing trend for the fisheries on these species to extend into deeper water and exploit all stages of the life history. Examples of such species in the North Atlantic are Greenland halibut (Reinhardtius hippoglossoides), anglerfish (Lophius spp.) and deep-water redfish (Sebastes mentella). The United Nations Food and Agriculture Organization (FAO) has divided the world’s oceans into statistical areas (Figure 1) and for some species there are data on landings. However, in many countries deep-water species are frequently landed unsorted in grouped categories that sometimes makes it difficult, if not impossible, to specify a proportion of the catch as deep-water. For example, many sharks are grouped and the term ‘sharks various’ can include inshore, deep-water, and oceanic pelagic species. Deep-water fishes are generally perceived as being slow-growing and having a high age at first maturity and a low fecundity, all of which contribute to making them susceptible to overexploitation. In a

226

new fishery the initial catch rates will be high, but, as the accumulated biomass of older fish is removed, the catch rates will decrease. For some species the sustainable yield may be only 1–2% per year of the pre-exploitation biomass. Few deep-water fisheries are managed, but where they are managed an objective has often been to maintain the stock at above 30% of the virgin biomass. Some deep-water species, such as sharks and orange roughy, have a global distribution, while others, such as roundnose grenadier, Sebastes spp. and Patagonian toothfish (Dissostichus eleginoides), are widely distributed within different oceans. Virtually nothing is known about the stock structure of most deep-water species and this can cause problems when developing management strategies. The following examples have been chosen to illustrate some of the problems associated with the sustainable exploitation and management of deep-water fisheries.

Black Scabbardfish (Aphanopus carbo) (Figure 2) One of the longest-established deep-water fisheries is the fishery for black scabbardfish in the eastern North Atlantic. It began as a longline fishery around the oceanic island of Madeira and probably originates from the seventeenth century. Records of landings exist from about the 1930s, but because of the artisanal nature of the fishery these may be unreliable. At the beginning the fishery was prosecuted from open boats using vertical longlines set at about 1000 m depth and which were hauled by hand. However, in recent years the fishery has evolved and now uses larger vessels, mechanized line haulers and horizontal floating longlines. As a result the landings have been increasing, partly as a result of increasing efficiency and also because of the discovery of new fishing grounds farther from the islands. In 1983 a longline fishery for black scabbardfish developed off mainland Portugal and catches increased rapidly. When a mixed bottom-trawl fishery developed on the continental slope to the west of the British Isles in the 1990s black scabbardfish was an important bycatch and landings increased considerably. There is little doubt that the catch per unit of effort (CPUE) in the trawl fishery is decreasing and in 2000 the International Council for the Exploration of the Sea (ICES) recommended a 50% decrease in fishing effort. However, there is a lack


OPEN OCEAN FISHERIES FOR DEEP-WATER SPECIES

227

18 27 27 21 67 61

37 31

34

77 71 51 87

57

41

47 51

81 58

58

48 88

Figure 1 Map showing major fishing areas used by FAO for statistical purposes. The numbered areas are referred to by name in the text and Figures 2–7 and can be identified by reference to Table 1. (From Gordon, 1999.)

Reported landings (t)

14 000 12 000 10 000 8 000 6 000

4 000 2 000

0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 Year Madeira

Mainland Portugal

Rockall Trough

Figure 2 The reported landings (tonnes) of black scabbardfish (Aphanopus carbo) caught by longline around Madeira and off mainland Portugal and by bottom trawl in the Rockall Trough (all north-east Atlantic).

the southern areas and the subadults migrate northward on a feeding migration. Studies of the diet of the subadult fish in the northern areas have shown that they feed on other fish such as blue whiting (Micromesistius poutassou) and argentine (Argentina silus) that shoal along the upper continental slope. If this hypothesis proves to be correct, effective management will have to take into account the three established fisheries and another that is developing in the MidAtlantic. An important by-catch of the longline fisheries for black scabbardfish are deep-water squalid sharks, such as Centroscymnus coelolepis, Centrophorus squamosus, and Centrophorus granulosus. Deep-water shark fisheries have their own problems (see below).

Roundnose Grenadier (Coryphaenoides rupestris) (Figure 3) of basic biological knowledge on which to base any effective management. The eggs, larvae, and smallest juveniles of black scabbardfish are unknown. Estimated ages range from about 8 to over 20 years, but without information on the juvenile stages these cannot be validated. Mature female fish are found around Madeira, off mainland Portugal, and probably also from the Azores and northward along the Mid-Atlantic and Reykjanes Ridges. To the west of the British Isles the fish are all immature. One hypothesis to explain this distribution is that there is a single northeastern Atlantic stock and that the adult fish spawn in

Concentrations of roundnose grenadier were discovered along the continental slope of the western North Atlantic north of 501N in the late 1950s. This led to fishery investigations and exploratory voyages by Russia, the then German Democratic Republic, and Poland, and also coincided with the introduction of the large factory-freezer stern trawlers. The first directed fishery for roundnose grenadier was by Russia in the north-west Atlantic in the mid 1960s. This fishery began off the north-west of Newfoundland, and later extended northwards to Labrador, Baffin Island and


228


Figure 3 The reported landings (tonnes) of roundnose grenadier (Coryphaenoides rupestris) caught by bottom trawl in the northeast Atlantic and the north-west Atlantic (FAO statistical areas).

western Greenland. The fishery developed rapidly, reaching a peak of over 83 000 t in 1971. The subsequent decline in the fishery has often been ascribed to overexploitation and, while this might be an important factor, other contributory factors are undoubtedly the establishment of 200-mile national fishery zones, changes in the allowable by-catches in the Greenland halibut fishery, and probably some misreporting of initial catches. Scientific research on roundnose grenadier lagged far behind the fishery and it was not until 1975 that a precautionary total allowable catch (TAC) was imposed. However, the landings continued to decline and were soon well below the level of the TAC, which was also being steadily reduced. It was only in the 1990s that landings reached the level of the TAC as a result of the by-catch of the European trawlers fishing for Greenland halibut in international waters around the Flemish Cap. A similar fishery developed in the international waters of the eastern North Atlantic (including the Mid-Atlantic and Reykjanes Ridges) in about 1973. Historically, the largest catches were from Russian vessels fishing the southern part of the Reykjanes Ridge. The landings peaked in 1975 and have fluctuated ever since. ICES has expressed concern that some of the catches in international waters are not being reported. By-catches of other species in these mixed, bottom-trawl catches have also been inadequately reported. In the early 1970s German trawlers and later French trawlers began to exploit spawning aggregations of blue ling (Molva dypterygia) in the northern parts of the Rockall Trough. Some French trawlers, which traditionally fish along the shelf edge for species such as saithe (Pollachius virens), began to move onto the upper and midcontinental slope of the Rockall Trough to exploit blue ling and by 1989 were beginning to land a bycatch of species such as roundnose grenadier, black scabbardfish, deep-water sharks, and several other

less abundant species. As the fishery evolved it moved into deeper waters and soon roundnose grenadier was a target species in its own right. The landings have remained fairly constant in recent years, but there is good evidence to suggest that the CPUE has decreased. The landings have been maintained by increasing fishing effort and efficiency, moving into new areas and deeper water, and discarding fewer of the smaller fish. In 2000 ICES recommended a 50% cut in fishing effort and the European Commission proposed a precautionary TAC on this and other deep-water species in 2001. However, the fisheries ministers deferred the Commission’s proposals and the fisheries remain unregulated. Almost nothing is known about the stock structure of roundnose grenadier and, although knowledge of the biology is improving, the data for analytical, age-based stock assessments is lacking. It is probable that the stocks in the area of the Rockall Trough that is under national jurisdiction are part of a larger stock around the Rockall and Hatton Banks that is in international waters. The lack of any regulation in international waters will diminish the value of any management measures introduced in waters under coastal state jurisdiction. There is also the issue whether TACs are an appropriate management tool for mixed fisheries. The problems of discarding over-quota marketable species, so-called high grading, etc., are all well documented in shelf fisheries, but the additional problem of a changing catch composition with depth in deep-water fisheries is something that will be difficult to incorporate into a management scheme.

Scorpaenid Fisheries (Sebastes spp. and Sebastolobus spp.) (Figures 4 and 5) The trawl fisheries for redfishes are among the most important deep-water fisheries of the North Atlantic. In the north-western Atlantic the landings comprise a mixture of three species (Sebastes marinus, S. fasciatus, and S. mentella), but only S. mentella is a truly deep-water species. Because the landings are not separated into species, it is impossible to assess the importance of the deep-water component of the catch. The fishery peaked at almost 400 000 t in the mid-1950s and, except for an increase in the 1970s, has been steadily declining. The peaks in landings represent times when different nations entered the fishery but, at the present time, the fishery is dominated by Canada. It appears that this fishery is now in serious decline, which is perhaps not surprising given the high level of exploitation, the slow growth,



229

Figure 4 The reported landings (tonnes) of redfish (Sebastes spp.) caught by trawl in the north-east Atlantic and the north-west Atlantic (FAO statistical areas).

Figure 5 The reported landings (tonnes) of Pacific Ocean perch (Sebastes alutus) caught by trawl in the east and west Pacific (FAO statistical areas).

the high age at first maturity and the viviparous mode of reproduction. The mean length of fish caught by research vessels has been declining steadily. In the eastern North Atlantic the landings are comprised of two species (Sebastes marinus and S. mentella). Recently the landings of redfish by some countries have been partly apportioned into the two species on the basis of research surveys. S. mentella is the deep-water species, but these data do not give a true indication of the likely proportion of this species to the total catch of the whole region. However, there can be little doubt that the trend in recent years has been to increasingly exploit S. mentella in deeper water and for an overexploitation of all stocks. A longline fishery for so called ‘giant redfish’ developed in 1996 on the Reykjanes Ridge but only lasted for that year as a profitable fishery. In the North East Pacific the fishery for Pacific Ocean perch (Sebastes alutus) began in 1946 and, after a slow period of development, the landings peaked at over 460 000 t in 1965 and then steadily declined over the next decade so that they now they range between about 10 000 to 30 000 t (Figure 5). In the North West Pacific the fishery for Pacific ocean perch is on a smaller scale but peaked in the 1970s and has since declined. This species has a range from the outer shelf to the upper slope and most of the fishery is on the outer shelf at about 200–300 m. There are several scorpaenid fishes that extend from the outer shelf to the continental slope and, with the exception of the Pacific Ocean perch, the FAO statistics do not separate them at the species level. However, many of the stocks in shallower waters have become depleted and the fishery has progressively moved into deeper waters to exploit several species. These include in the east the long spine thornyheads (Sebastolus altivelis) and splitnose rockfish (Sebastes diplopoa) and in the west the

longfin thornyhead (Sebastolobus macrochir), the shortspine thornyhead (S. alascanus), the rougheyed rockfish (Sebastes aliutianus) and the osaga (S. iracundus).

The Orange Roughy (Hoplostethus atlanticus) (Figure 6) The orange roughy is probably the most frequently cited example of an exploited deep-water fish partly because it is an extreme example of all the features than can lead to over-exploitation; a life span of 4100 y, high age at first maturity (B25 y), and a relatively low fecundity. It also has many qualities that make it a highly marketable commodity. It tends to aggregate in large numbers around seamounts, pinnacles, and steep slopes and with sophisticated fishing techniques these stocks can be fished down

Figure 6 The reported landings (tonnes) of orange roughy (Hoplostethus atlanticus) caught by trawl in the south-west Pacific (New Zealand and eastern Australia), the eastern Indian ocean (Australia), the south-east Atlantic and the north-east Atlantic.


230


very rapidly: a type of fishing sometimes referred to as ‘mining’. The most important fishery for orange roughy is around New Zealand. This trawl fishery, at depths from about 700 to 1200 m, began in the late 1970s and expanded rapidly, with landings exceeding 60 000 t per annum in the late 1980s and early 1990s. Since then landings have declined to about 25 000 t. The dense spawning and feeding aggregations were systematically fished down and the fishery was maintained by the continual discovery of new fishing grounds, mainly on the extensive deep-water plateaus and rises that surround New Zealand. The present level of landings is composed of a reduced catch from established fisheries and higher catches from new areas. The Australian fishery for orange roughy extends along the slope of south-eastern Australia, around Tasmania, and extends up to New South Wales. Fishing began in 1982 as part of a general slope fishery and it was not until 1986 that dense aggregations were found off Tasmania and a directed fishery was established. Between 1986 and 1988 the landings of orange roughy ranged between about 4600 and 7200 t per annum. New non-spawning aggregations and a new spawning aggregation resulted in a dramatic increase in catches from 26 000 t in 1989 to 41 000 t in 1990. Since then the landings have declined. FAO reports the landings of this area together with the New Zealand Fishery (south-west Pacific). A smaller fish fishery is reported by FAO in the eastern Indian Ocean and landings have steadily decreased from initial levels of about 2000 t in 1989. In the north-eastern Atlantic, French trawlers began landing orange roughy from deep water to the west of the British Isles in 1991. Although there is still a degree of secrecy associated with this fishery, there is little doubt that it takes place at greater depths (down to about 1700 m) than the multispecies trawl fishery for roundnose grenadier (see above) and most probably in areas of steep slopes and seamounts. The landings peaked in 1992 and the fishery has subsequently declined in the Rockall Trough (ICES Sub-area VI). Landings from west of Ireland (ICES Sub-area VII) have remained fairly constant, probably as a result of the sequential discovery and fishing down of new aggregations. In the international waters of the north-eastern Atlantic, a Faroese vessel has been fishing on the Mid-Atlantic Ridge and other offshore banks and seamounts. There have also been spasmodic catches by Iceland on the Reykjanes Ridge. The New Zealand orange roughy fishery, because of its economic importance, has been the subject of considerable research in recent years. Nevertheless,

the rapid development of the orange roughy fishery far outpaced the scientific knowledge of the fish and the stocks. The initial quotas were essentially precautionary, to allow time for an assessment to be carried out. These assessments were based on random stratified trawl surveys but, because there was a lack of knowledge about the great age and low productivity and the stock discrimination of this fish, some inappropriate assumptions were made. As a result the quotas set in the early years of the fishery, mainly for the Chatham Rise and Challenger Plateau, were too high and the stocks were overexploited. In recent years there has been a marked improvement in the knowledge of the biology of the species, although the estimates of very high ages continue to be controversial. Improved stock assessment, using a variety of methods, and the application of genetics to stock discrimination have allowed more realistic TACs to be implemented. For example, on the Chatham Rise the TAC was decreased from 33 000 t in 1990 to 7000 t in 1996. By 1998 the stock was considered to be at 15–20% of virgin biomass but rebuilding. However, in other areas quotas continue to be reduced and on the Challenger Plateau, once a major fishery, the quota for the 2000/2001 fishing year has been reduced to 1 tonne, thereby effectively closing the fishery. The Australian fishery is also managed by quotas and these have been reduced drastically since the start of the fishery. The fishery in the north-east Atlantic remains totally unregulated.

Deep-water Sharks Deep-water sharks are landed as a by-catch of many deep-water fisheries and there are some targeted

Figure 7 The reported landings (tonnes) of Patagonian toothfish (Dissostichus eleginoides) caught in the south-west Pacific (New Zealand and eastern Australia), the south-east Pacific, the Indian and Atlantic Ocean sectors of Antarctica (Australia) and the south-west Atlantic.



Table 1

231

Deep-water fish species that are listed by area in FAO fishery statistics

(Continued )


232


fisheries. In some fisheries, such as in the bottomtrawl fishery for roundnose grenadier in the northeastern Atlantic, only some of the shark species that are caught are landed. The rest are discarded, sometimes after removal of the liver and/or the fins. The reported landings to FAO are, with a few exceptions, by grouped category and it is therefore difficult to estimate the true extent of the fishery. Deep-water sharks of small adult size or the juveniles of marketable species that are discarded are unlikely to survive. No records are kept of discards and there have been few scientific studies of discarding. Sharks are vulnerable to over-exploitation because of their longevity and slow growth. They have a high age at first maturity and a low reproductive rate per female as a result of the long embryonic development for oviparous and viviparous species, from several months up to about two years. The majority of deepwater species might have biennial or even triennial reproductive cycles. Several fisheries that target sharks, such as the longline fishery for gulper shark (Centrophorus granulosus) off mainland Portugal and the kitefin shark (Dalatias licha) at the Azores, have declined rapidly after a few years and are good examples of the vulnerability of these stocks. The CPUE of the combined catches of leafscale gulper shark (Centrophorus squamosus) and the Portuguese dogfish (Centroscymnus coelolepis) from the trawl fishery to the west of the British Isles has declined by about 50% in less than ten years. Longlining is often

presented as a more selective and environmentally friendly method of exploiting deep-water fish resources. However, many exploratory surveys have shown that there can be a high, perhaps unacceptable, by-catch of unwanted sharks.

Patagonian Toothfish (Dissostichus eleginoides) (Figure 7) The main deep-water fishery in the southern oceans is for the Patagonian toothfish. It is widely distributed in the area and it is caught by longline at depths down to 2000 m or more. The reported landings are probably unreliable as there is thought to be illegal fishing in the area. The significant landings are from the Atlantic (FAO Area 48) and Indian Ocean (FAO Area 58) sectors. In the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR) area the toothfish is managed by TAC with additional technical measures, some of which are designed to limit the accidental capture of seabirds attracted to the baited hooks as the lines are being set.

Other Deep-water Fisheries The above examples were chosen to illustrate some of the different aspects of deep-water fisheries and the associated problem of managing what are generally considered to be fragile resources. There are



many other deep-water species that are either targeted or are a by-catch of other deep-water fisheries. Table 1 gives some of the important deep-water species that are clearly identifiable in published FAO statistics by statistical area. This list excludes many species that are recorded under grouped categories or where most of the fishery takes place on the continental shelf.

Ecosystem Effects There is a growing concern, in part resulting from new technologies that allow direct observation, about the effects of fishing on the ecosystem and the probability that the deep-water ecosystems will be particularly susceptible to damage and will take a long time to recover. There is an expanding literature on the physical effects of fishing gear, especially trawls, on the sea bed in shallow water. In deeper water there are reports from photographic surveys of trawl marks in soft sediments, but little information on the effects at the biological level. Trawl damage to hard bottoms such as deep-water reefs and seamounts has been documented and reefs of the deepwater coral Lophelia off Norway and some seamounts off Australia and New Zealand are now protected. Discarding of species of no commercial value can represent a very high proportion of the catch. None of these discards will survive the trauma of being brought to the surface from great depths. Deep-water fishes generally have fragile skins and it is very probable that most fish entering the trawl and escaping through the meshes will not survive. The impacts of these discards and unseen mortalities on the ecosystem are unknown. The effect of the selective removal of top predators is also largely unknown.

See also Deep-Sea Fishes. Demersal Species Fisheries. Demersal Species Fisheries. Mesopelagic Fishes.

233

Open Ocean Fisheries for Large Pelagic Species. Salmon Fisheries, Pacific.

Further Reading Clark M (2001) Are deepwater fisheries sustainable? – The example of orange roughy (Hoplostethus atlanticus) in New Zealand. Fisheries Research 52: 123--136. Clark MR, Anderson OW, Francis RIC, and Tracey DM (2000) The effects of commercial exploitaton on orange roughy (Hoplostethus atlanticus) from the continental slope of the Chatham Rise, New Zealand, from 1979 to 1997. Fisheries Research 45: 217--238. Food and Agriculture Organization (1994) Review of the state of world fishery resources. FAO Fisheries Technical Paper, No. 335, p. 136. Rome: FAO. Food and Agriculture Organization (1997) Review of the state of world fishery resources: marine fisheries. FAO Fisheries Circular No. 920. 173 pp. Rome: FAO. Gordon JDM (1999) Management considerations of deepwater shark fisheries. In: Shotton R (ed.) Case studies of the management of elasmobranch fisheries. FAO Fisheries Technical Paper. No. 378, pp. 774--818. Rome: FAO. Gordon JDM (2001) Deep-water fisheries at the Atlantic frontier. Continental Shelf Research (In press). Hopper AG (ed.) (1995) Deep-water Fisheries of the North Atlantic Oceanic Slope. Dordrecht: Kluwer Academic Publishers. Koslow JA, Boehlert GW, Gordon JDM, et al. (2000) Continental slope and deep-sea fisheries implications for a fragile ecosystem. ICES Journal of Marine Science 57: 548--557. Lorance P and Dupouy H (2001) CPUE abundance indices of the main target species of the French deep-water fishery in ICES sub-areas V, VI and VII. Fisheries Research 52: 137--150. Merrett NR and Haedrich RL (1997) Deep-sea Demersal Fish and Fisheries. London: Chapman and Hall. Moore G and Jennings S (eds.) (2000) Commercial Fishing: The Wider Ecological Impacts. Randall DJ and Farrell AP (eds.) (1997) Deep-sea Fishes. London: Academic Press. Uchida RN, Hayasi S and Boehlert GW (eds.) (1986) Environment and Resources of Seamounts in the North Pacific. NOAA Technical Report NMFS 43. 105 pp. Seattle: United States Department of Commerce.


OPEN OCEAN FISHERIES FOR LARGE PELAGIC SPECIES sharks and other large pelagics are discussed only briefly, and whales and whaling are discussed elsewhere.

J. Joseph, La Jolla, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2031–2039, & 2001, Elsevier Ltd.

The Animals Introduction Open-ocean fisheries for large pelagic species target relatively large organisms that spend most of their lives in offshore waters, usually within about 100– 150 m of the surface. These large pelagic species include finfish, such as scombrids (tunas) (Figure 1), istiophorids (billfishes: spearfish, sailfish, and marlins), xiphiids (swordfish) (Figure 2), coryphaenids (mahimahi, or dolphinfish), and elasmobranchs (sharks), and also cetaceans (whales, dolphins, and porpoises). Many of these species do not spend their entire lives in the open ocean but, as their biological needs dictate, make sporadic forays into the coastal zone. Therefore, although they are captured mostly in the open ocean, they are sometimes taken near shore. This article concentrates on fisheries for tunas and billfishes, because they account for by far the greatest proportion of the total catch of all large pelagic species. (In this article catch includes fish discarded at sea, but landings do not.) The various types of fishing gear and vessels used to catch large pelagics, and the magnitude of those catches, are reviewed. Because of their great size, speed, and stamina, these large pelagics are much sought after by recreational fishers, so sportfishing is also included in this article. Fisheries for

Tunas, and many of the billfishes, comprise most of the world’s catch of large pelagic species. They have many characteristics that set them apart from most other types of fishes and which contribute to their nomadic lifestyle. They have very high metabolic rates, resulting in high energy and oxygen demands. Tunas must swim constantly in order to pass enough water over their gills to meet this high demand for oxygen. Unlike most other fishes they cannot pump water over their gills, so if they stopped swimming they would suffocate, and they would also sink because they are denser than the water that surrounds them. They possess highly developed circulatory systems that allow them to retain or dissipate heat as needed for efficient operation of their nervous, digestive, and locomotor systems, and are capable of maintaining their body temperature up to 151C above ambient. Their fusiform shape, highly developed swimming muscles, and crescent-shaped tail, which provides maximum forward thrust, enable them to swim at extraordinary speeds. Many of the other large pelagic species have a similar propensity for travel, but their adaptations to a nomadic life style are different. Tuna tend to form large schools, which makes it relatively easy to locate and catch them.

YELLOWFIN TUNA Thunnus albacares

Figure 1 Yellowfin tuna (Thunnus albacares) (Source: InterAmerican Tropical Tuna Commission).

234

Figure 2 Swordfish (Xiphias gladius) (Source: Inter-American Tropical Tuna Commission).


OPEN OCEAN FISHERIES FOR LARGE PELAGIC SPECIES

Methods of Catching the Fish Humans have most likely been fishing since the first days they walked the earth. The first marine organisms caught were probably taken by hand from tidepools and beaches. As their fishing skills developed, early humans fashioned hooks, harpoons, traps, and nets to capture their prey. Archaeological evidence indicates that ancient man harvested large pelagic species: skeletal remains found in caves indicate that giant Atlantic bluefin (Thunnus thynnus) were caught near modern-day Sweden more than 6000 years ago, and similar evidence shows that giant Pacific bluefin (Thunnus orientalis), weighing more than 250 kg, were taken by native Americans in the region between the southern Queen Charlotte Islands, British Columbia, and Cape Flattery, Washington, more than 5000 years ago. Just how early humans caught these giant fish is uncertain, but probably harpoons or handlines with baited hooks were used. As civilizations developed, so did trade in agricultural and natural resource products, including fish. Historical evidence indicates that nearly 3000 years ago the Phoenicians salted and dried tuna that they had caught, and traded it throughout the Mediterranean region. Harpoons

The harpoon is simply a spear modified for fishing by attaching a line and buoy for retrieving both spear and prey. Harpoons have been used since early times for capturing large pelagics, but were perfected for whaling. For many years harpoons were the primary means of capturing swordfish, which have a tendency to bask at the surface of the ocean, and they are still used today in many parts of the world for this purpose. In recent years they have been replaced by more efficient forms of fishing, but harpoon-caught swordfish still command premium prices, apparently because of their better quality. Some marlin (Makaira spp. and Tetrapturus spp.) and giant bluefin tuna are also taken with harpoons. Vessels used for harpoon fishing have a long, narrow platform projecting from the bow, on which the harpooner stands, and from which vantage point there is a better chance of harpooning the fish before it is aware of the approach of the vessel. Traps

Traps have been used extensively to catch tunas throughout the Mediterranean right up to the present time. The almadraba, a type of trap net used since the time of the Phoenicians, consists of corridors of netting, called leads, up to several kilometers long,

235

up which migrating fish swim until they reach holding chambers, where they are harvested. Because the fish need not be pursued, vessels are not required to catch the fish, but only to transport the catch from the traps to shore. Hook and Line

Hand lines The simplest form of hook-and-line fishing is a single hook attached to a hand-held line. The hooks, which can be baited with live, dead, or artificial bait, are generally set from a vessel or floating platform. All types and sizes of vessels, propelled by engine, sail or paddles, are used today to capture large pelagic fishes in various areas of the world. Hand-line fisheries, although primitive and, in a commercial sense, inefficient, are nevertheless widespread, and often spectacular to observe. In one such fishery, in Ecuador, the fishermen, frequently father and son, set out to sea shortly after midnight in sail-powered dugout canoes about 6 m long. Once on the fishing grounds, a single hook, baited with a small fish, is trailed several meters behind the canoe. When a fish, frequently large yellowfin tuna (Thunnus albacares) or black marlin (Makaira indica), which can weigh up to several hundred kilos, is hooked, the sails are reefed and the fishermen attempt to bring it to the boat. The struggle can last several hours, and once the fish is brought alongside it is clubbed in order to immobilize it. If it is too large to lift into the canoe, the fishermen enter the water, swamp the canoe, and roll the fish into it. They then bail out the vessel, hoist sail, and head for shore and the fish market. Scenes such as these are repeated in many artisanal fisheries throughout the world, but at diminishing rates as motorized fiberglass skiffs replace traditional canoes. Pole and line Pole-and-line fishing is used in many parts of the world to catch large pelagics, principally tunas. The technique, which was developed independently in several separate regions of the world, is similar to that used for handlines, with the difference that the hook and line are attached to the end of a pole, giving greater reach and better leverage. In the South Pacific, such fishing is done from canoes (Figure 3) and small motorized vessels using lures traditionally made from seashells (Figure 4). In Japan, commercial pole-and-line fishing for skipjack (Katsuwonus pelamis), yellowfin, and bluefin tunas was common in coastal waters for many centuries. During the twentieth century the Japanese developed larger pole-and-line vessels that were able to travel to all oceans of the world to fish for tunas. These vessels carry live bait in tanks of circulating sea water and


236


Figure 3 Pole-and-line fishing from a canoe in Tokelau (Courtesy of Robert Gillette, Suva, Fiji).

Figure 4 Pearl shell lure for catching large pelagic fish, from Tokelau (Courtesy of Robert Gillette, Suva, Fiji).

crews of more than 25; they can freeze their catches and stay at sea for several months. Some now carry automated machine-operated poles; a single crewman can tend several poles at once, thereby increasing vessel efficiency. In the Indian Ocean, the island nation of Maldives was ‘built on the backs of fish,’ as the majority of its people made their living or derived their sustenance from fishing. The major form of fishing is with poles and lines, targeting skipjack, yellowfin, and wahoo (Acanthocybium solandri), using small vessels originally powered by sail with compartments for live bait built into the hull through which water is circulated. This method of fishing, which uses both artificial and live bait and a hook of unique design, is distinctively different in detail from that of Japan. In southern California, after the introduction of tuna canning in the early twentieth century, poleand-line fishing was used to supply the growing demand for tuna. The first vessels were small, used ice

to preserve their catch, and fished within a few days of port. As the demand for tuna grew, larger, longerrange boats were built. This was the origin of the ‘tuna clipper,’ pole-and-line vessels that could pack several hundred tons of tuna in refrigerated fish holds, carry large amounts of live bait, and stay at sea for many months; they plied the eastern Pacific between California and Chile. Typically a vessel would catch bait before putting to sea to search for tuna. On sighting a school, the ‘chummer’ would throw the chum, or bait, to bring the fish alongside the vessel, and fishermen standing in racks at water level on the port side and stern of the vessel would catch the fish, usually using artificial lures (Figure 5). Once a fish was hooked, the fisherman lifted the pole, jerking the fish from the water and over his head. The pole would be stopped in the vertical position, and the fish would fall off the barbless hook onto the deck. If the school was feeding well, a fisherman could catch fish as quickly as he could get the hook into the water. If the chummer was successful in keeping a large school feeding alongside the boat, several tons could be taken in a short time. Because many fish are too large to be lifted by one fisherman on a single pole, a single hook might be attached to two, three and even four poles (Figure 6). For economic reasons, pole-and-line fishing is no longer predominant in the eastern Pacific, and only a few vessels operate on a regular basis. During the 1950s pole-and-line fishing was introduced off tropical West Africa by the French and Spanish; it is still practiced there today, but to a limited extent. Longlines Longlines, as the name implies, consist of a long mainline, kept afloat by buoys, from which



237

Most billfishes are taken by longlines. During the last two decades longlining for swordfish has become very popular, and accounts for most of the landings of that species. Longline vessels targeting swordfish operate in the Atlantic, Pacific, and Indian Oceans.

Figure 5 Pole-and-line fishing from a baitboat in Ecuador (Courtesy of Robert Olson, IATTC).

Trolling In this method of fishing a number of lines to which lures are attached are towed from outrigger poles and from the stern of a vessel. When fishing for large pelagics, the lures are towed at a speed of several knots. Most troll fisheries target albacore tuna (Thunnus alalunga), but many other species, including bluefin, yellowfin, and wahoo, are also caught. This form of fishing is used throughout the world, usually from vessels of less than 20 m in length. Nets

Figure 6 Catching a three-pole yellowfin tuna from a US baitboat (Courtesy of Pete Foulger, Harbor Marine Supplies, San Diego).

a number of branch lines, each terminating with a hook, are suspended (Figure 7). Longlines for large pelagics can have a mainline up to 125 km long, with as many as 500 buoys and 2500 hooks, and can take 8–12 h to set or retrieve. Longlines catch whatever fish happen to take the hook; a single set can capture several species of tunas, billfishes, and sharks. The distance between the buoys determines the depth of the hooks; they are normally suspended at depths between 100 and 150 m, where the water is cooler and large tunas are most likely to be. These large tunas, especially bigeye (Thunnus obesus), command high prices in the sashimi markets of Japan, and most of the larger longline fishing vessels of the world target this market. Longline vessels vary in size. Small vessels use much shorter lines, and normally operate in coastal waters, whereas the larger vessels roam the oceans of the world in search of their prey and, supplied by tender vessels, can stay at sea for extended periods. Japanese vessels account for most of the longline catches, followed by Taiwanese and South Korean vessels.

Gill nets Gill nets consist of a panel of netting, usually synthetic, held vertically in the water by floats along its upper edge and weights along its lower edge, in which fish become enmeshed when they attempt to swim through. The drift gill nets used to capture large pelagics in the open ocean consist of continuous series of such panels, sometimes more than 100 k long, and are very effective in catching the target species, but they also catch birds, sea turtles, and marine mammals. Because of these by-catches, and the fact that nets lost or abandoned at sea continue to catch fish (‘ghost fishing’), in the late 1980s the United Nations recommended banning the use of drift gill nets over 2.5 km long on the high seas. However, such nets are still used within the Exclusive Economic Zones (EEZs) of some nations, particularly for catching swordfish and sharks. Purse seines Purse seiners catch more tuna than all other types of vessels combined. Like a gill net, a purse seine is set vertically in the water, with floats attached to the upper edge; along the lower edge is a chain, for weight, and a series of rings, through which the pursing cable passes. Purse seines are constructed of heavy webbing, can be up to 1.5 km long and 150 m deep. When a school of tuna is sighted a large skiff, to which one end of the net is attached, is released from the stern of the fishing vessel (Figure 8). The vessel circles the school, paying out net, until it reaches the skiff, closing the circle. The pursing cable is winched aboard the vessel, closing the bottom of the net and trapping the fish; the net is then hauled back on board, concentrating the catch for loading. Purse-seine vessels fishing for large pelagics range from about 25 to 115 m in length, and can carry up to


238


Figure 7 Diagrammatic sketch of longline fishing.

STAGES OF A PURSE-SEINE SET

(1) Release of the Net Skiff

(2) Encirclement

(3) Pursing the Net

OPIC 82.90

(4) Rings Up; Net Pursed

(5) Net Retrieval

(6) Sack-up and Brailing

Figure 8 Setting a purse-seine tuna net (Source: Reproduced with permission from stequent and Marsac, 1991).

3000 tonnes of frozen fish, but most such vessels targeting the two principal species caught by this fishery, skipjack and yellowfin tuna, average about 75 m and 1200 tonnes (Figure 9). These vessels are capable of fishing all the oceans of the world and staying at sea for several months. Many carry helicopters for locating fish and directing the vessel while the net is being set, and are also equipped with

sophisticated electronic devices such as specialized radar for detecting flocks of birds at great distances, current meters, satellite communication gear for receiving data on weather and ocean conditions, global positioning systems, and a variety of sonars for detecting schools of fish underwater. Purse seiners make three types of sets: on free-swimming schools of tuna, on schools associated with flotsam such as parts



239

of trees, and on schools associated with marine mammals, mostly dolphins. Because the amount of flotsam is limited, fishermen build and deploy fishaggregating devices (FADs) to attract fish. FADs are usually fitted with electronic transmitters which allow the vessel to locate them easily.

at the rear. The forward end of the net is held open, and long reaches of net, or arms, serve to guide the fish toward the bag end. Most trawls are used to fish on or near the seabed, but recently pelagic trawls have been used to fish for tuna, primarily albacore.

Trawls A trawl is a conical net, towed through the water by a trawler vessel, with the apex, or bag end,

Catches

Figure 9 A modern purse-seine vessel of 1200 tonnes capacity; note the helicopter on board (Courtesy of Dave Bratten, Inter-American Tropical Tuna Commission).

The annual world landings of marine fish is about 85 Mt (million tonnes). Of this, approximately 4 Mt are large pelagics (Figure 10). The principal market species of tuna – skipjack, yellowfin, bigeye, albacore, and bluefin – account for about 3 Mt, but their economic value is out of all proportion to the volume of their landings. In fact, some species of tuna are among the most valuable fish: a single 300-kg bluefin tuna can sell for over US$ 75 000 in the Japanese sashimi market. About 65% of the world landings of all large pelagic species is taken from the Pacific Ocean, 20% from the Indian Ocean, and 15% from the Atlantic Ocean. Skipjack represents the greatest proportion in all three oceans, constituting about 50% of the total landings of the principal market species of tuna; yellowfin accounts for nearly 35%, and bigeye, albacore, and bluefin make up the rest.

1600 1400

1200

1 430 000

1 130 000

Kilotonnes

1000

800

600

400

200

335 000 195 000

155 000 60 000

0

Skipjack Yellowfin Bigeye Albacore Bluefin

80 000 90 000

50 000

Marlin Swordfish Sharks Mahimahi Sailfish Spearfish

Figure 10 Recent world landings of certain large pelagic species. (Data modified from the 1996 and 1997 FAO Yearbooks of Fishery Statistics.)


240


700 000

700 600 520 000

Kilotonnes

500 400 355 000 330 000

300 235 000 210 000 205 000

200

180 000

165 000 135 000 115 000

100 0 Japan

France 80 other Indonesia Rep. Korea Philippines Taiwan Spain USA Mexico Ecuador nations

Figure 11 Recent world landings of principal market species of tunas, ranked by country of capture. (Data modified from the 1996 and 1997 FAO Yearbooks of Fishery Statistics.)

The largest fishery for tuna in the world is in the western Pacific, which produces about 35% of the world landings, mostly skipjack and yellowfin. Other large fisheries are in the eastern Pacific, the western Indian Ocean, and the eastern Atlantic. In all of these fisheries vessels of many nations fish side by side, using all types of fishing gear. Prior to about 1975 most tuna-fishing vessels were longliners or baitboats, but now purse seiners predominate. In terms of landings, Japan is the leader, with 20% of the world total, followed by Taiwan, Indonesia, Spain, South Korea, the United States, the Philippines, Mexico, France, and Ecuador (Figure 11). These ten nations account for 80% of the total landings, the remainder being shared by about 80 other nations. Japan is also the principal consumer of large pelagics, accounting for about 30% of the world total, followed by Western Europe with about 25% and the United States with about 20%. In recent years the annual landings of billfish have been about 170 000 tonnes, about half of it swordfish. About 160 000 tonnes of sharks are landed annually, but of this total probably less than 20% are pelagic species. The actual catches are almost certainly higher, but many sharks (mostly blue sharks, Prionace glauca) are caught only for their fins and are not reflected in world catch statistics. Japan, once again, accounts for the highest landings of billfish and, with the exception of swordfish, also consumes most of the billfish caught. Both the catch and consumption of sharks are widespread among many nations. The annual commercial landings of the common dolphinfish (Coryphaena hippurus) are less than 50 000 tons.

Utilization Although tuna was first canned in Europe in the mid1800s, prior to the twentieth century nearly all tuna and other large pelagics were eaten either fresh, salted or dried, or used to make sauce. In 1903 tuna canning was introduced to the United States in San Pedro, California. The canned product was well received by American consumers, increasing the demand for fish and ushering in the modern tuna industry. Since then the consumption of tuna has steadily increased, and today about 60% of the world landings, nearly all caught by purse seiners, is canned, yielding some 150 million cases, or about 1.5 Mt, annually. The United States consumes almost 50 million cases, Western Europe just over 50 million cases, and the remainder is consumed mostly in Asia and Latin America. Second in importance is fresh tuna, for consumption either raw, as sashimi, mostly in Japan, or grilled. Katsuobushi, lightly fermented, smoked and dried skipjack tuna, used as a condiment in Japanese cuisine, is another important use of tuna. Of the billfishes, swordfish is consumed mostly as steaks, whereas blue marlin (Makaira spp.) and striped marlin (Tetrapturus audax) are frequently consumed fresh or smoked. Much of the catch of spearfish (Tetrapturus spp.) and sailfish (Istiophorus platypterus) is processed into imitation shrimp and crab, and other products. With the exception of a few species, sharks are generally held in lower esteem as a food fish, and command much lower prices than tunas and some of the billfish. In many cases only the fins, which fetch high prices in Asian markets as an ingredient for



soup, are utilized, the remainder of the animal being discarded. Prior to World War II some species of sharks were heavily fished because their livers were high in certain vitamins, but with the development of synthetic vitamins this use of sharks has declined. Currently, the demand for some sharks is increasing because of the purported curative properties of their cartilage. Most of the dolphinfish landed is consumed in the United States, fresh or fresh/frozen.

Mariculture Propagation of freshwater fish has been practiced for centuries, but it is only in recent times that the culture of marine fish (mariculture) has become important on a commercial scale. Over the last few years mariculture has contributed about 10% to the total annual landings of all marine fishes. The pelagic habitat, migratory nature, and large size of pelagic species make rearing them in captivity difficult, and only recently have scientists been able to spawn and rear tunas artificially. In Japan there has been limited success in spawning bluefin tuna in captivity, hatching the eggs, and rearing the young. At the InterAmerican Tropical Tuna Commission’s Achotines Laboratory in Panama, scientists have been able to spawn yellowfin tuna held in captivity in onshore tanks on a regular basis, and have also successfully reared the hatched eggs to the juvenile stage. It is anticipated that this research will eventually make it possible to complete the life cycle of yellowfin tuna in captivity, an essential precondition to rearing tuna commercially. In the meantime, attempts at what is known as bluefin ranching are being increasingly successful. This involves catching adult bluefin tuna and transporting them to pens in inshore areas, where they are held and fed until reaching a marketable size and condition. Because of the high value of bluefin in the sashimi market, ranching is a growing enterprise in many countries, including Australia, Croatia, Japan, Mexico and Spain.

Recreational Fishing Recreational fishing for large pelagics has been important since the early 1900s, and became more so after it was popularized by writers such as Zane Gray and Ernest Hemingway. The most sought-after species are the larger ones, like marlins, sailfish, and bluefin and yellowfin tuna. It was originally mostly a sport for the wealthy, because of the high cost of fishing boats and access to areas where these species abound. More recently, however, fleets of charter

241

vessels and party vessels, which take groups of recreational fishers for extended trips on the high seas in pursuit of large pelagic species at a reasonable cost, have become available. This, coupled with the lower cost of air travel, has brought big-game fishing within the grasp of many people, and recreational fishing has become an important component of the economy of a number of nations. However, this expansion of recreational fishing has brought it into conflict with commercial fishers, due primarily to competition for the same species, particularly marlins and bluefin tuna. Although there are no accurate estimates of how many fish the recreational fisheries take, the catches are small in comparison to commercial fisheries, but both the catches and the amount of money spent to make them are increasing.

Conservation of the Resource Large pelagics are a renewable resource whose abundance can be profoundly influenced by human activities, particularly fishing. These activities have been steadily increasing as the demand for food and recreation from the sea increases, and landings of tunas and related species have increased tenfold since 1945. These increases have resulted in overfishing of some species, particularly bluefin tuna, swordfish, and some sharks. Many of the other species of large pelagics, with the notable exception of skipjack tuna, are currently fully, or nearly fully exploited, and sustained increased landings cannot be expected. Many of these other species are in urgent need of conservation if they are to be protected from overexploitation. Widespread concern over the status of some of these stocks has led to action by coalitions of chefs throughout the United States to boycott the use of swordfish in their restaurants, and attempts by public interest groups to include sharks and bluefin tuna as species covered by the Convention on International Trade in Endangered Species (CITES). This increased fishing has also caused problems for other marine species taken incidentally as by-catch with the species targeted by the fishery. These by-catch species are often of no value to the fishers, and are thrown back to the sea, dead. By-catches occur in most fisheries for large pelagic species, and it is important to determine their impact on the populations from which they are taken and on the ecosystem to which they belong. A similar problem in some purse-seine fisheries for tuna, especially in the eastern Pacific, is the capture of dolphins. Many dolphins were killed in purse-seine nets between the inception of this type of fishing in the late 1950s and the early 1990s, but in recent years, thanks to the joint efforts of environmental organizations, the fishing


242


industry, and governments, the mortality of dolphins caused by tuna fishing has been reduced to very low levels. Populations of pelagic sharks, in addition to being heavily exploited for their flesh and/or fins, are also taken in large quantities as a by-catch in some fisheries. Because many species of sharks have low fecundity, are slow growing and later maturing, they are vulnerable to severe overexploitation which could lead to the collapse of some populations. Because tunas and other large pelagic species are highly migratory, the vessels that fish for them roam the oceans of the world, and many nations are involved in their harvest, international cooperation is essential for effective conservation of these species. Article 64 of the United Nations Convention on the Law of the Sea calls on nations to work jointly through appropriate international bodies to manage and conserve such highly migratory species. There are currently four such bodies, the Inter-American Tropical Tuna Commission, the International Commission for the Conservation of Atlantic Tunas, the Commission for the Conservation of Southern Bluefin Tuna, and the Indian Ocean Tuna Commission, and negotiations for creating a similar body for the western and central Pacific are near completion. These bodies are responsible for coordinating and conducting scientific research and, on the basis of that research, making recommendations to governments for the conservation of the tunas and tuna-like species. As a result of their efforts conservation measures are in effect for many of the large pelagic species. Such measures can take a variety of forms, including catch quotas that limit the amount of fish that can be caught or landed, closed areas and seasons, limits on the size of fish that can be landed, and restrictions on the types and amounts of gear that can be used to catch fish. As fishing pressure continues to increase, there will be a need to expand such measures to other areas and species.

See also Dynamics of Exploited Marine Fish Populations. Fishery Manipulation through Stock Enhancement or Restoration. Marine Fishery Resources, Global State of. Pelagic Fishes.

Further Reading Anganuzzi AA, Stobberup KA, and Webb NJ (eds.) (1996) Proceedings of the 6th Expert Consultation on Indian Ocean Tunas. Colombo, Sri Lanka: Indo-Pacific Tuna Development and Management Programme. Beckett JJ (ed.) (1998) Proceedings of the ICCAT Tuna Symposium. Collective Volume of Scientific Papers, Vol. 50. Madrid: International Commission for the Conservation of Atlantic Tunas. Cort JL (1990) Biologiá y pesca del atuń rojo. Institut. Espanõl Oceanographia, Pub. Esp. No. 4. Madrid. Doumenge F (1999) ha Storia Delle Pesche Tonniere (The history of tuna fisheries). Biol Mar Medit 6(2): 1--106. FAO Fisheries Department (1993) World Review of High Seas and Highly Migratory Fish Species and Straddling Stocks. FAO Fisheries Circular. No. 858. Rome: FAO. Fonteneau A (1997) Atlas of Tropical Tuna Fisheries – World Catches and Environment. Paris: Orstom editions. Goadby P (1987) Big Fish and Blue Water – Gamefishing in the Pacific. North Ryde, London: Angus and Robertson. Jones S and Kumaran M (1959) The fishing industry of Minicoy Island with special reference to the tuna fishery. Indian Journal of Fisheries VI(1): 30--57. Joseph J (1998) A review of the status of world tuna resources. In: Nambiar KPP and Pawiro S (eds.) Papers of the 5th World Tuna Trade Conference, pp. 8--21. Bangkok: INFOFISH. Joseph J and Greenough JW (1979) International Management of Tuna, Porpoise, and Billfish – Biological, Legal, and Political Aspects. Seattle: University of Washington Press. Joseph J, Klawe W, and Murphy P (1988) Tuna and Billfish – Fish without a Country. La Jolla: Inter-American Tropical Tuna Commission. Orbach MK (1977) Hunters, Seamen, and Entrepreneurs – The Tuna Seinermen of San Diego. Berkely: University of California Press. Shomura RS, Majkowski J, and Langi S (eds.) (1994) Interactions of Pacific Tuna Fisheries. FAO Fisheries Technical Paper 336, Vols 1 and 2. Rome: FAO. Stroud RH (ed.) (1989) Planning the Future of Billfishes: Research and Management in the 90s and Beyond vols 1 and 2. Savannah: National Coalition for Marine Conservation, Inc. Ward P and Elscot S (2002) Broadbile Swordfish: Status of World Fisheries. Australia: Bureau of Rural Sciences, Aquaculture, Fisheries and Forestry.


OPTICAL PARTICLE CHARACTERIZATION P. H. Burkill and C. P. Gallienne, Plymouth Marine Laboratory, West Hoe, Plymouth, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2040–2048, & 2001, Elsevier Ltd.

Particles and their Properties Particles are ubiquitous in ocean waters, where they are intimately involved in defining the optical properties, productivity, and biogeochemistry of our seas. Marine particles exist in a wide range of sizes and concentrations, and exhibit an inverse relationship between size and concentration, and a positive relationship between concentration and ambient nutrient concentration (the ‘trophic status’) in surface waters of the ocean. Particles range in size from the largest marine organisms (blue whales, c. 70 m length) down to the size that arbitrarily divides particles from dissolved materials. In biological oceanography, this is defined operationally as 0.2 mm. But here we will focus on particles that fall within the size range of the plankton. Plankton organisms range from viruses (c. 0.05 mm) up to larger zooplankton such as euphausiids (c. 2 cm). However, even within this size range, many particles are not living but instead contribute to the large pools of detritus that often predominate over living particles in the ocean.

Particle Characterization A wide array of techniques is available for the characterization of marine particles. Although most are based on optical properties, nonoptical techniques, such as the acoustic doppler current profiler (ADCP) and the multifrequency echosounder, can also be used to quantify and characterize particles such as large zooplankton and fish in sea water. Optical characterization techniques vary considerably in their resolution. At one extreme, satellite-based remote sensing can be used to quantify and characterize the marine phytoplankton across whole ocean basins. At the other extreme, microscope-based techniques and analytical flow cytometry resolve single particles. For the biologist, microscopy is the benchmark procedure for identification of plankton. This is true whether the particles of interest are viruses, which are typically analyzed by electron

microscopy, or bacteria, protozoa, or larger zooplankton, which are analyzed by light microscopy. As an adjunct to light microscopy, fluorescencebased techniques are used increasingly to characterize, and sometimes quantify, the chemical properties of cells. Such approaches can be extremely powerful, particularly when used in conjunction with fluorescently labeled molecular probes. Such probes can be tailored to target specific taxonomic groups. Although microscopy remains the benchmark, for the simple reason that ‘what you see, you believe,’ it is time consuming and costly. A wide range of techniques offer rapid analysis of particles. However, these tend to be ‘black box’ techniques and should always be used with appropriate controls. No single technique provides a panacea in particle analysis, and it is often useful to combine two or more complementry techniques. Rapid optical techniques for analyzing planktonsized particles may be based on scattered, fluorescent or transmitted light. Scattering and fluorescence methods are applicable to smaller particles (o500 mm equivalent spherical diameter (ESD)), where as larger particles, such as zooplankton, are usually analyzed by transmission techniques. Two techniques that have been developed rapidly in the last decade, are analytical flow cytometry (AFC) and optical plankton counting (OPC). AFC and OPC are particularly suitable for the analysis of smaller particles (viruses to protozoa) and larger particles (metazoa), respectively.

Analytical Flow Cytometry Technique Analytical flow cytometry (AFC) is a generic technique based on the multiparametric analysis of single particles at high speed. Originally developed for medical hematology and oncology, AFC is used increasingly in biological oceanography. Its strengths are derived from its quantitative capability, versatility, sensitivity, speed, statistical precision, and ability to identify and, in many instruments, sort particle subsets from heterogeneous populations. Its drawbacks are its cost and, for commercial instruments, the small volume of sample (c. 0.5 cm3) analyzed. Particle characterization and quantification in flow cytometry relies on cellular fluorescence and light scatter, and the power of the technique derives from the ability to make multiple measurements


243

244

OPTICAL PARTICLE CHARACTERIZATION

Beam stop

Sorted cells

Forward light scatter detector

Waste tank Light collection lens

Cell sorter catcher tube Flow cell

560nm DM ± 11.3 nm 640 nm LP filter Sample

Brewster window Orange fluorescence detector

Depolarized light scatter detector Horizontal polarizer

488 nm filter

585nm filter ± 21nm 650 nm LP filter

Laser beam Sheath fluid tank Red fluorescence detector

Side scatter detector

Figure 1 Operating principles of AFC in which samples containing the particles of interest are passed singly across a laser beam. Each particle scatters light and this is collected by forward and side light scatter detectors. Birefringent particles will tend to depolarize the vertically polarized laser light and this is measured at the appropriate detector. Fluorescence from each particle is collected and spectrally filtered so the wavelength of interest is detected by photomultiplier tubes. Output from each sensor is digitized and the data are transferred to a computer. Particle size and refractive index are determined by the light scatter and the chemical properties are determined by fluorescence. Particles exhibiting appropriate properties can be collected by sorting, whereas other particles pass to waste. (Figure produced by Glen Tarran, Plymouth Marine Laboratory.)

simultaneously on each cell at high speed. Typically, up to 5000 cells can be analyzed per second and sorting rates of 410 000 s1 with 498% purity, can be achieved. The principles of AFC (Figure 1) are based on hydrodynamically focusing a suspension that is streamed coaxially through a flow chamber so that individual particles pass singly through the focus of a high intensity light source. Suitable light sources include coherent wave lasers since these provide a very stable light beam that can easily be focused to small dimensions. Light flux from the highly focused light source generates enough fluorescence from individual cells for this to be measured in the few microseconds taken to traverse the light beam. As the particles traverse the beam, they scatter light and may also produce fluorescence. Cellular fluorescence arises from autofluorescent cells or cells stained with a fluorochrome, and this is collected by a high numerical aperture lens located orthogonally to the irradiation source and the sample stream. The light collected is spectrally filtered sequentially by dichroic mirrors which reflect specific wavelengths into photomultiplier tubes (PMT). The PMTs are optically screened by band-pass filters. The quantity of light incident upon each PMT, responding within a given color band, is then proportionally converted

into an electrical signal. The signal is amplified, digitized, and stored transiently in computer memory. The data are then displayed on a computer screen and stored onto disk as ‘list mode’ data. The list mode data can be considered to be analogous to a spreadsheet in which each row represents a particle, each column represents a different AFC sensor with values that represent quantitative optical signatures of each particle. The advantage of list mode data, as with data in any spreadsheet, is that it can be replayed, reanalyzed, and redisplayed. Commercial cytometers are usually equipped with light-scatter detectors that are situated in the narrow forward and an orthogonal angle, as well as two or three fluorescence PMTs. Although a single laser is normal, AFC instruments can be equipped with two or more lasers to increase the number of fluorochrome excitation wavelengths. Some cytometers may use arc-lamp excitation particularly if UV irradiation is required. There is also a move to the use of diode lasers for applications that demand low power use. Flow chambers may vary in their hydrodynamic, optical, mechanical, and electrical characteristics to achieve high sensitivity and good stream stability. AFC instruments often have quartz cuvette sensing zones to improve sensitivity and to allow the application of UV irradiation. Specialized



cytometers can also measure particle volumes based on the Coulter principle of electrical impedance alteration as particles flow through a restricted orifice. Other specialized instruments may generate images of particles in the sensing zone. Data processing and display procedures have been developed which handle fully crossed-correlated multidimensional data and this is achieved by microcomputers. Considerable developments have taken place in the last few years to apply sophisticated procedures such as multiparametric statistics or neural net generation, to identify and characterize particles from within heterogeneous mixtures. Many AFC instruments are able to sort cells and this is invaluable for identification, manipulation or as a gateway to other analysis procedures. High speed ‘sorting-in-air’ is based on developments in ink-jet printing. Two populations may be sorted from the sample stream that undergoes oscillation, driven by a piezo-electric crystal that is mechanically coupled to the flow chamber. The crystal, driven at 30–40 kHz, produces uniform liquid droplets of which a small percentage contain single cells. The ‘sort logic’ circuitry compares processed signals from the sensors with pre-set, operator-defined ranges. When the amplitude falls within the pre-set range, an electronic time delay of a few microseconds is activated. This triggers an electrical droplet-charging pulse at the moment the cell arrives at the droplet formation break-off point. The droplet-charging pulse causes a group of droplets to be charged, and subsequently, deflected by a static electric field into a collection vessel. Cells failing the pre-set sort criteria do not trigger droplet-charging, and so pass undetected into the waste collector. Other sorting procedures include one in which a collecting arm moves into the sample stream to pick up particles that meet the programmed sort criteria. Sorting is an essential adjunct to AFC and is crucial to verifying both satisfactory instrument and analytical protocol operation. Applications Oceanographic applications of AFC are now diverse and continue to expand rapidly. Although the fundamental principle of AFC remain constant, recent developments in optical sensitivity and design of AFC instruments have aided new applications. Fluorochrome chemistry and molecular biology are both richly endowed fields and developments in fluorescent assays of biochemical constituents, coupled with the ability to target individual taxa have proved invaluable for AFC applications in marine biology. Detection limits are adequate to measure cellular attributes of many planktonic cells.

245

Cellular fluorescence may be derived from two basic categories: 1. autofluorescence in which the fluorescent molecule of interest occurs naturally in the cell; 2. applied fluorescence in which the fluorescent dye is applied, or otherwise generated, and fluorescence is accumulated within the cell. Phytoplankton. Analysis of phytoplankton by AFC is based on the presence of chlorophyll, a highly autofluorescent compound that is found in all viable plants. Chlorophyll is the phytoplankton’s principal light-harvesting pigment, and absorbs light strongly in the blue and red regions of the visible spectrum. Blue light cellular absorption coincides with the emission of the argon ion laser at 488 nm that is commonly used in flow cytometers. Chlorophyll fluorescence is emitted in the far red (lem ¼ 680 nm), thereby offering a useful Stokes’ shift of some 200 nm. This window means that phytoplankton can be readily characterized and quantified by flow cytometers equipped with an argon laser (or other blue light source) and suitable spectral filtration (such as a 650 nm longpass filter) of fluorescent light emitted by cells onto a sensitive photomultiplier tube. As well as chlorophyll, some phycobiliproteins are also autofluorescent. One of these is phycoerythrin which is found in cyanobacteria and cryptophytes. Although phycoerythrin absorbs light in the green– blue end of the spectrum, the 488 nm emission of the argon is sufficiently close to excite this compound. In studies of phytoplankton, fluorescence from phycoerythrin is measured by a separate photomultiplier tube that is spectrally filtered to collect emissions at 585 nm. Based on this differentiation and coupled with light scatter measurements (the magnitude of which is roughly proportional to cell size), AFC can readily differentiate and quantify the phytoplankton groups shown in Table 1. In recent years, the application of powerful multivariate statistical and neural net procedures have been applied to further characterize algal taxa Table 1 Routine AFC analysis of phytoplankton based on optical characteristics Differentiation

AFC criteria

Phytoplankton Prochlorococcus Synechococcus Cryptophytes Coccolithophores

Chlorophyll autofluorescence Low chlorophyll and light scatter Low phycoerythrin and light scatter Phycoerythrin and light scatter High orthogonal light scatter and laser depolarization


246


from within the complex mixtures that are typical of sea water. Multivariate statistics that have been used include quadratic discriminant analysis and canonical variate analysis. The latter is a useful graphical technique for analyzing and displaying data, whereas quadratic discriminant analysis can discriminate over two-thirds of mixtures of 22 algal taxa, with classification rates 470%. Such approaches are more than two orders of magnitude faster than conventional flow cytometric analyses for discriminating and enumerating phytoplankton species. Artificial neural nets (ANN) have proved to be extremely powerful in increasing AFC capability for differentiating algal taxa. This approach is based on training an ANN to recognize the optical characteristics of individual taxa. This is achieved by presenting the net with AFC data derived from unialgal cultures. The unknown samples are then analyzed by the AFC under the same conditions and the data passed through the trained ANN. The net outputs identification probabilities for each cell analyzed. Several types of ANN have been used and it is now possible for nets to differentiate and recognize 470 taxa with high accuracy. Considerable developments are anticipated in this field in the coming years. As well as providing procedures for differentiation of phytoplankton from other particles, there are AFC protocols for quantifying cellular attributes of phytoplankton. These include the cellular concentrations of chlorophyll, phycoerythrin, protein and

DNA as well as enzymes such as ribulose-1,5bisphosphate carboxylase. AFC instruments are capable of great sensitivity and are able to quantify concentrations of cellular chlorophyll in phytoplankton in the range of about 1–2000 fg cell1. In practice, marine phytoplankton are typically analyzed using a fresh sample of sea water without pretreatment. The sample is analyzed at a constant rate so sample volume can be determined from analysis time. Chlorophyll-containing phytoplankton and those containing phycoerythrin are registered by the red and orange fluorescence emitted from single cells as they traverse the laser beam. Typical sample analysis time is generally 4–5 min. Examples of data generated by AFC protocols for the analyses of natural waters are shown in Figure 2. Bacteria. Bacteria, traditionally quantified by epifluorescence microscopy, can now be differentiated from other particles and analyzed by AFC (Figure 2). Both approaches are based on the intercalation of a fluorochrome with the cell’s nucleic acid. However, such intercalation is universal and often does not differentiate between autotrophic and heterotrophic bacteria. A range of fluorochromes have been used including 488 nm absorbing YOYO-1, YO-PRO-1, PicoGreen and SYBR Green as well as the more traditional UV excited bis-benzimide Hoechst 33342 or 40 ,6-diamidino-2-phenylindole (DAPI). Of these,

Figure 2 AFC characterization and differentiation of bacteria and phytoplankton in lab culture and in natural communities carried out at the Plymouth Marine Laboratory. Bacteria are measured using SYBR Green fluorescence and the different phytoplankton are measured by chlorophyll autofluorescence. Analysis can be verified by microscopic analysis of flow-sorted particles. (Figure produced by Glen Tarran, Plymouth Marine Laboratory.)



SYBR Green offers the practical advantage that autotrophic and heterotrophic bacteria can be differentiated readily. It is now possible to Quantify the cellular protein and DNA content of bacteria in natural waters. Such AFC techniques use the intensity of SYPRO-protein or DAPI–DNA fluorescence of individual marine bacteria. Cultures of various marine bacteria have been measured in the range 60–330 fg protein cell1, but the amount of natural bacterioplankton from the North Sea in August 1998 was shown to be only 24 fg protein cell1. The total DNA of natural bacteria has been estimated to be about 3 fg cell1 by AFC techniques. Changes in bacterioplankton community composition have also been assessed by molecular biological AFC techniques. The combination of AFC analysis and sorting combined with denaturing gradient gel electrophoresis of polymerase chain reaction (PCR)-amplified 18S rDNA fragments and fluorescence in situ hybridization has been shown to be a rapid method of analyzing the taxonomic composition of bacterioplankton. Experimental manipulation of natural water samples resulted in a bacterial succession from members of a Cytophaga flavobacterium cluster, through gamma-proteobacteria and finally alpha-proteobacteria. Protozoa. AFC-based techniques for the analysis of protozoa have, so far, been based on molecular probes. Ribosomal RNA species-specific probes to various members of the common heterotrophic flagellate genus, Paraphysomonas, have been developed (Figure 3). However, they have been restricted to laboratory applications since naturally occurring organisms exhibit cellular fluorescence levels that are often too low to distinguish from background. This observation may either be a reflection of poor probing efficiency or it may be due to the organisms low growth rates in situ. Viruses. Viruses are now thought to be one of the most abundant types of particles in the ocean. AFC protocols are now available for enumerating natural marine viruses based on staining with the nucleic acid-specific dye SYBR Green-I. Interestingly AFCbased counts are often higher than those obtained by microscopy, suggesting that further development work is needed. However, this AFC protocol reveals two, and sometimes three, virus populations in natural samples, whereas microscopy would only differentiate one pool of viruses. Cultures of several different marine virus families (Baculoviridae, Herpesviridae, Myoviridae, Phycodnaviridae, Picornaviridae, Podoviridae, Retroviridae, and

247

(A)

(B)

(C) Figure 3 Photomicrographs showing a mixture of four Paraphysomonas species after hybridization with a mixture of PV1 and EUK probes tagged with fluorescein and rhodamine, respectively. PV1 is specific for P. vestita whereas EUK labels all eukaryotes. The mixture was irradiated at (A) 488 nm to show P. vestita labeled with PV1, (B) 568 nm to show organisms labeled with EUK, and (C) both 488 and 568 nm to reveal both probes. Scale bar is 10 mm. (Reproduced from Rice et al. (1997) with permission from the Society for General Microbiology.)

Siphoviridae) have also been stained with a variety of highly fluorescent nucleic acid-specific dyes. Highest fluorescence is achieved using SYBR Green I, allowing DNA viruses with genome sizes between 48.5 and 300 kb (kilobases) to be detected. Small genome-sized RNA viruses (7.4–14.5 kb) are at the current limit of detection by AFC. Zooplankton and larval fish. Although commercially available AFC instruments are directly applicable


248


for the analysis of microbial cells, it is also possible to adapt the generic AFC concept for the analysis of metazoan organisms. This involves a scaling up of flow chambers and the associated fluidics system. The Macro Flow Planktometer built within the EU MAROPT project has been applied to organisms as large as larval fish. An inherent property of this system is that it incorporates ‘imaging in flow’ as part of the analysis. Images may be stored either photographically or electronically and then be available for subsequent image analysis. In-situ AFC. As we move towards operational oceanography, there is an increasing need for autonomous, in situ instrumentation. An important step towards this has been made recently with the development of CytoBuoy, an AFC instrument housed in a moored buoy and capable of wireless data transfer. CytoBuoy is one of the very few AFC instruments to have been designed and built purely for aquatic use. Its characteristics include enhanced optics and electronics designed to obtain maximal information on particle characteristics. Whereas standard cytometers reduce these to single peak or area ‘list-mode’ numbers, time resolved signals are preserved fully and transferred to the computer as raw data. Pulse shape signals aid identification considerably and allow, for the first time, a true measurement of particle length. The CytoBuoy concept has also been taken a step further and has been redesigned as a functional module of the UK Autosub autonomous underwater vehicle. Future trends. The generic capability of AFC lends itself to a variety of applications. The thrust in recent years has been towards greater taxonomic resolution and this has been aided by the application of molecular techniques particularly involving oligonucleotide probes. These have many advantages including the ability to be tailored to target particular groups of organisms. Here the level of taxonomic targeting may range from general (e.g. differentiation of classes of phytoplankton) down to individual species. Molecular probes may also be coupled with chemotaxonomic capability that can analyze cellular function such as specific enzyme production. Flow cytometry has opened up our ability to characterize marine particles with a greater degree of taxonomic resolution and further development of techniques such as ANN will increase this capability considerably. This should allow characterization of natural populations objectively in close to real-time. The quest for greater analytical resolution will undoubtedly continue. However, it is also possible that

a taxonomic identification watershed may soon be reached. So far, the focus of AFC protocols has generally conformed to traditional taxonomic criteria. But there may be another route that remains to be explored. This involves an approach to classifying particles based directly on flow cytometry variables of light scatter, fluorescence, time of flight etc. Such an approach might prove worthwhile because of its direct simplicity. How such an approach would compare to traditional taxonomic identification remains to be addressed, and that remains an exciting challenge for the future.

Optical Plankton Counting The Optical Plankton Counter (OPC) has been designed to analyze those zooplankton called the mesozooplankton whose size range conventionally spans 0.2–20 mm. Technique The OPC uses a collimated beam of light through an enclosed volume, received by a photosensor. When a particle interrupts this beam of light the sensor produces an electronic response proportional to the cross-sectional area of the particle (Figure 4). This response is digitized and this digital size is converted into ESD using a semiempirical formula. This is the diameter of a sphere having the same cross-sectional area as the particle being measured, and can be simply converted into volume. The OPC comes in two versions: a towed instrument for in situ use and a benchtop laboratory version. The towed version has a sampling tunnel of 22 cm 2 cm. The laboratory version has a glass flow cell 2 cm square. LED array 640 nm

Monitor photodiode

Lens + aperture

Target LED control

Reference Linear detector

FSK out Pulse

Figure 4 Schematic of the operating principle of the optical plankton counter. LED, light-emitting diode; FSK; frequency shift keying.



Applications The OPC is capable of large-scale, rapid and continuous sampling of zooplankton, providing a reliable measure of size distributed abundance and biovolume, between 0.25 and 20 mm ESD, at data rates up to 200 s1. The use of the in situ OPC on various towed platforms is now well established. The laboratory version is intended for characterizing preserved samples. It has also been deployed at sea in pump-through mode producing continuous real-time data on surface zooplankton abundance and size distribution, permitting near continuous sampling of epipelagic zooplankton across ocean basin scales. Figure 5 shows data from the OPC in this mode on the Atlantic Meridional Transect. The data in Figure 5 show one of 11 transects completed to date, each comprising a near-continuous 13 000 km transect of size distributed biovolume in the North and South Atlantic, illustrating the power of this kind of instrument. Data at this level of detail and spatial resolution acquired autonomously and continuously could not have been gathered over such large spatial scales by any other means. Operational considerations Bias and coincidence require calibration of the instrument against some other sampling device. Initial calibration uses spherical glass beads of known size. Several researchers have noted nonlinearity in this calibration at the extremes of the size range, and have suggested that the operational size range should be reduced. Sensor response time and coincidence limit the densities at which the OPC can operate. Coincidence occurs when more than one particle is in the beam at the same time. They register as one larger particle and abundance will, therefore, be underestimated, and biomass

249

overestimated. Coincidence in the towed OPC can be reduced in areas of high abundance by inserting into the sampling tunnel a transparent plate, reducing the sampling volume to one-fifth. There is a finite probability of coincidence occurring at all concentrations, increasing with abundance and flow rate through the instrument. It has been determined experimentally that at a count rate of 30 s1, more than 90% of particles will be counted. Smaller particles far outnumber larger ones, so coincidence will result in a loss of these smaller particles, and biovolume is underestimated to a lesser degree than abundance. Towed at 4 m s1, the standard OPC will pass 17.6 l s1 through the sampling tunnel. Coincidence will therefore begin to have a significant effect on estimated abundance above concentrations of 1700 m3, or 8500 m3 with the flow insert. Object orientation can also present problems in this type of counter. Elongated organisms present a very different cross-sectional area depending upon whether they are side-on or end-on. Biovolume may be considerably underestimated in the latter case. It has been shown that on average the cross-sectional area measured for a randomly oriented object will be greater than 70% of the true value. Biovolume may also be overestimated by the spherical model assumed in the OPC calibration – most zooplankton have a shape closer to that of an oblate spheroid. We use an ellipsoidal model based on cross-sectional area of the particle. From cross-sectional area the length and width of the ellipsoid can be calculated (assuming a length to width ratio typical for copepods). Several cases have been reported of OPCs producing counts many times higher than those from concurrent net samples. Comparison between OPC and a Longhurst Hardy Plankton Recorder (LHPR)

_3

Biovolume (mm3 m )

600 500 > 2000 µm 1000−2000 µm 500−1000 µm 250−500 µm

400 300 200 100 0 47

43

35

27

18

9 −1 − 8 −15 _ 24 _ 29 _ 35 _ 46 Latitude (˚N)

Figure 5 A 13 000 km transect of size-distributed epipelagic zooplankton biovolume in the North and South Atlantic from the OPC in continuous surface sampling mode on the Atlantic Meridional Transect.


250


showed that abundance and biovolume recorded by the OPC were consistently four times higher than for the LHPR (200 mm mesh). Using a 53 mm mesh showed that there had been significant undersampling of zooplankton o350 mm ESD by the LHPR. Studies like these show that care is required when comparing any two sampling methods. When comparing the OPC to net systems, careful selection of mesh size or OPC lower size threshold is required. Our own investigations indicate that the most suitable mesh size for a net used in comparison with the standard OPC is 125 mm (250 mm ESD, 3:1 ellipsoidal model, mesh size 75% of width of smallest animal to be quantitatively sampled). It must be emphasized that OPC cannot discriminate living from nonliving particles. All particles within its operating range will be detected, including detritus, marine aggregates, air bubbles, etc. A recent study found average abundance in the Faroe–Shetland Channel and north–western North Sea to be 2.5 times higher for an OPC than for net samples and up to 40 times higher at extremely low concentrations. This was put down to detrital particles/marine aggregates, which may in some cases be the most abundant particle type in the water, and are often too fragile to be sampled by nets. Detrital particles may be considered part of the plankton, being significant in marine food webs, but their presence in large numbers will bias comparisons between OPC and net derived estimates of abundance.

Particle Imaging Instruments Automated particle counters (electronic, acoustic, and optical) can give reliable data on zooplankton abundance and size distribution, but tell us little or nothing about the species present, except where dominant species are already known and well separated in size. Larger autotrophic particles may be counted as zooplankton, and in areas where high concentrations of detritus or marine aggregates are present, they are not discriminated from zooplankton. To make these discriminations, information on shape as well as size is needed, requiring the use of imaging devices. Photographic methods High-speed silhouette photography has been undertaken at sea without the need for sample preservation. Samples in a shallow tray on top of 8 10-inch (20 25 cm) film were exposed by a xenon strobe. Samples could then be counted and identified from the film, to provide a permanent record. A towed version of the system using concentrating nets ahead of a camera

system was subsequently developed. Although this system avoided some of the problems associated with the deployment and use of net systems, and preservation of samples, a human investigator still carried out counting and identification of the samples. Automated image processing of digitized photographic images could alleviate this problem, but duration and spatial resolution is limited by the amount of film that can be carried. Video methods The most advanced video method for producing abundance and size distribution indicators of good temporal and spatial resolution, together with near-real-time classification to taxonomic groups, is the video plankton recorder. This consists of a towed frame 3 m long with four cameras each having concentric fields of view (5–100 mm) covering the size range of interest (0.5–20 mm. ESD). The imaged volume is defined by the oblique intersection of the cone defining the camera field of view, and that produced by a collimated, strobed 80-watt xenon light beam pulse 1 ms duration, producing dark field illumination. Sample volume varies from 1 ml to 1 liter. Fiberoptic telemetry is used to send video data to the surface, where information is recorded to tape. Subsequent upgrading of this system is expected to permit preprocessing to be done in real time, and near real time production of size and species distributions. Concurrent calibration of the device is provided by an integrated LHPR system. The device can differentiate detrital material and zooplankton, unlike optical and acoustic systems. The video plankton recorder has been used extensively around George’s Bank in the north-west Atlantic. Video cameras used in scientific imaging typically produce 10–100 million bytes s1, and image-processing algorithms are highly computer intensive. For real-time acquisition systems to be used at sea in continuous sampling mode, assuming 1 m3 min1 as the required sampling volume, and at typical oceanic zooplankton abundance of 50–5000 m3, we need to be able to process 3000–300 000 animals per hour. Although currently available technology can resolve this problem to some degree, the solution is not yet likely to be economical in size or cost. ROV devices Remotely operated vehicles (ROVs) carrying video cameras have been used for in situ studies of zooplankton abundance and behavior. Gelatinous organisms are notoriously difficult to sample using traditional methods, and an ROV carrying a video camera has been successfully deployed for their study. ROVs are usually restricted to small-scale, observational studies, but



may be useful for the location of zooplankton patches. Small imaging volume may preclude the study of rarer taxa, and poor image quality may preclude the study of smaller taxa. Zooplankton also exhibit attraction/avoidance responses to the presence of ROV systems.

See also Acoustic Scattering by Marine Organisms. Artificial Reefs. Carbon Cycle. Fish Larvae. Fluorometry for Biological Sensing. Inherent Optical Properties and Irradiance. Krill. Microbial Loops. Plankton. Plankton Viruses. Platforms: Autonomous Underwater Vehicles. Population Genetics of Marine Organisms. Remotely Operated Vehicles (ROVs).

Further Reading Boddy L, Morris CW, and Wilkins MF (2000) Identification of 72 phytoplankton species by radial basis function neural network analysis of flow cytometric data. Marine Ecology Progress Series 195: 47--59. Brussaard CPD, Marie D, and Bratbak G (2000) Flow cytometric detection of viruses. Journal of Virological Methods 85: 175--182. Davis CS, Gallager SM, Marra M, and Stewart WK (1996) Rapid visualisation of plankton abundance and taxonomic composition using the video plankton recorder. Deep-Sea Research II 43: 1947--1970. Dubelaar GBJ and Gerritzen PL (2000) CytoBuoy: a step forward towards using flow cytometry in operational oceanography. Scientia Marina 64: 255--265. Gallienne CP and Robins DB (1998) Trans-oceanic characterisation of zooplankton community size structure using an Optical Plankton Counter. Fisheries Oceanography 7: 147--158.

251

Herman AW (1992) Design and calibration of a new optical plankton counter capable of sizing small zooplankton. Deep Sea Research 39(3/4): 395--415. Jonker R, Groben R, Tarran G, et al. (2000) Automated identification and characterisation of microbial populations using flow cytometry: the AIMS project. Scientia Marina 64: 225--234. Kachel V and Wietzorrek J (2000) Flow cytometry and integrated imaging. Scientia Marina 64: 247--254. Olson RJ, Zettler ER, and DuRand MD (1993) Phytoplankton analysis using flow cytometry. In: Kemp, et al. (eds.) Handbook of Methods in Aquatic Microbial Ecology, pp. 175--186. Boca Raton: Lewis. Reckermann M and Colijn F (2000) Aquatic flow cytometry: achievements and prospects. Scientia Marina 64(2): 119--268. Rice J, Sleigh MA, Burkill PH, et al. (1997) Flow cytometric analysis of characteristics of hybridization of speciesspecific fluorescent oligonucleotide probes to rRNA of marine nanoflagellates. Applied and Environmental Microbiology 63: 938--944. Rice J, O’Connor CD, Sleigh MA, et al. (1997) Fluorescent oligonucleotide rDNA probes that specifically bind to a common nanoflagellate. Paraphysomonas vestita. Microbiology 143: 1717--1727. Schultze PC, Williamson CE, and Hargreaves BR (1995) Evaluation of a remotely operated vehicle (ROV) as a tool for studying the distribution and abundance of zooplankton. Journal of Plankton Research 17: 1233--1243. Wood-Walker RS, Gallienne CP, and Robins DB (2000) A test model for optical plankton counter (OPC) coincidence and a comparison of OPC derived and conventional measures of plankton abundance. Journal of Plankton Research 22: 473--484. Zubkov MV, Fuchs BM, Sturmeyer H, Burkill PH, and Amann R (1999) Determination of total protein content of bacterial cells using SYPRO staining and flow cytometry. Applied and Environmental Microbiology 65: 3251--3257.


ORGANIC CARBON CYCLING IN CONTINENTAL MARGIN ENVIRONMENTS R. A. Jahnke, Skidaway Institute of Oceanography, Savannah, GA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Continental margin environments, including estuarine, salt marsh, mangrove forest, coral reef, continental shelf, and continental slope ecosystems, are the interface zone between oceanic and terrestrial realms with gateways also to the atmosphere and solid Earth (sediment) reservoirs. Transports between these carbon reservoirs exert critical control on the global carbon cycle and will be a major factor in mitigating anthropogenically driven climate change. This interface is also where man most directly interacts with and exploits the oceans. Much of the global economy’s goods and services are transported along and across coastal zones; most of the world’s fisheries are found within continental margin ecosystems; and much of the waste and materials mobilized from humans and their industrial and agricultural activities are delivered to the coastal zone. In addition, threefourths of the world’s human population will reside on the coastal plain by 2025. Rising population not only increases human pressures on and interactions with coastal ecosystems but also increases the susceptibility of humans and their property to oceanrelated natural hazards such as inundation by tsunamis and hurricanes. Advancing understanding of coastal marine processes is, therefore, of fundamental importance. The following section focuses specifically on the cycling of organic carbon in this critically important environment.

Processes that Characterize Coastal Ecosystems Many transport and exchange processes are either intensified within or unique to ocean boundaries and exert considerable influence on the structure, composition, and characteristics of continental margin habitats. Examples of some of these processes are schematically depicted in Figure 1. Many of the processes highlighted in the figure emphasize interactions with the seafloor which rises to intersect the sea surface at the shoreline. This is not to minimize the importance of processes in the water column,

252

some of which are included as well, but rather reflects the recognition that at the most fundamental level, it is the presence of the seafloor that differentiates continental margin ecosystems from their oceanic counterparts. The important inputs to the margin depicted in Figure 1 include surface freshwater inputs; groundwater exchanges; atmospheric inputs, which while not unique to coastal systems are generally intensified due to proximity to terrestrial sources of airborne materials; and oceanic inputs represented by wind-driven upwelling but also including inputs by currents and upwelling driven by other processes such as eddies associated with boundary currents. Important to the overall character of coastal systems are processes that internally exchange and transform materials, promoting the biogeochemical cycling of carbon, nutrient elements, and associated trace elements. Examples include benthic solute exchange; sediment resuspension; intensified water column mixing due to friction with the seafloor; internal wave–seafloor interactions; turbulent interactions with the surface and bottom boundary layers; and water column and benthic interactions with surface gravity waves. These material exchanges support transformative processes such as biological production and respiration and abiotic processes such as mineral precipitation and dissolution. Processes within margin systems also provide efficient conduits for the transport of materials to the open ocean surface, intermediate and deep water column, and the seafloor. Contributing to these important fluxes are transports associated with offshore surface and subsurface advective plumes; cascading downslope highdensity flows; particle fluxes associated with the above advective transports; gravitational settling; and downslope benthic boundary layer lateral transport. As complex as this figure is, numerous processes are not displayed, such as ice-driven exchanges and organism migrations, and Figure 1 is admittedly still a simplistic view of the margin system. The processes depicted in the figure structure margin ecosystems and support high rates of metabolic activity and vertical and lateral material fluxes. Furthermore, it must be recognized that the intensity, mix, and effectiveness of the processes vary temporally in response to atmospheric and hydrographic forcings. The magnitudes and interactions of these processes and timing of responses are not well known.


ORGANIC CARBON CYCLING IN CONTINENTAL MARGIN ENVIRONMENTS

253

Figure 1 Examples of important transport processes that are unique or intensified along continental margins. Reproduced with permission, & Skidaway Institute of Oceanography.

The magnitude and mix of dominant processes vary across this coastal/margin interface zone between the land and ocean system. In most instances, variations are gradual and dividing this system into subsystems is a difficult and occasionally arbitrary process. Nevertheless, for discussion and quantification purposes, it is useful to divide this complex system into simpler subsystems. In the following, the coastal ecosystems closest to and with the greatest interaction with the shoreline are discussed as a group and include estuarine, salt marsh, and mangrove forest ecosystems. This section is then followed by a discussion of continental shelf and slope ecosystems.

Shore Zone Ecosystems (Estuaries, Salt Marshes, and Mangrove Forests) Perhaps the most common and best-studied coastal ecosystem is the estuary. An estuary is typically considered to be a semi-enclosed region where fresh water, predominantly from a river but possibly also including groundwater input, and salt water mix.

The generalized pattern of circulation is that of fresh and brackish waters advecting off shore at the surface, saline waters being drawn shoreward along the bottom and mixing of these waters at the boundary between them. The thickness of the saline water along the bottom thins progressively upstream, eventually pinching out and resulting in the commonly described ‘salt wedge’ estuary (depicted in Figure 1). This circulation concentrates nutrients within the estuary and maintains water-column stability, thereby supporting high rates of primary production. While this simple picture describes most estuaries, there are natural variations. For example, for some large rivers, the freshwater discharge is sufficient to ‘push’ the salt wedge out onto the open continental shelf. Estuaries are important pathways for transferring materials from the land to the ocean. The efficiency with which materials such as suspended sediments and nutrients are passed through the estuary and delivered to the continental margin depends on a variety of factors including the magnitude of river flow,


254


morphology of the estuary, and residence time of waters within the estuarine system. In addition, driven by the strong physical-chemical gradients and intense metabolic rates, important transformations of materials may occur within the estuary. One critically important example is denitrification, conversion of nitrogen from a ‘fixed’ form such as nitrate or ammonium that is biologically available to nitrogen gas which can be utilized by only a very select group of microorganisms, the nitrogen fixers. Because nitrogen is a required nutrient element for life and overall biological primary production is limited by the availability of fixed nitrogen in many regions, the rate of denitrification in estuaries can be a major factor controlling estuarine biological production. Because estuaries are often excellent harbors, providing access to ocean- and river-based transportation and supporting productive fisheries, many of the largest cities are also located adjacent to estuaries. Human influences on the coastal environment are often concentrated and accentuated within estuarine systems. Influences include altering ecosystems by removing specific top predators through fishing activities, introducing exotic species through transportation and dumping of ballast waters in ships, eutrophication of the estuarine system by increased nutrient loading from sewage, farm runoff, and other nonpoint sources of nutrients, and the introduction of contaminants (e.g., trace metals and pesticides). Through these activities and interactions, the cycling of carbon within many, or most, estuarine systems has already been altered from natural rates. Other ecosystems at the shore zone include mangrove forests and salt marshes. Mangroves are found along sheltered tropical and subtropical shorelines. While the mangrove trees are the obvious characteristic of these ecosystems and a major contributor to total biological primary production, other plants such as macroalgae, periphyton, and phytoplankton also contribute significantly. The proportion of the contribution from each is controlled by such factors as the total surface area covered by each plant type, water turbidity, and nutrient delivery pathways and magnitude. Supported by contributions from these different plants, mangrove ecosystems exhibit some of the highest rates of primary production observed for any ecosystem on earth (Table 1). Salt marshes occur over a wider latitudinal range than mangroves, extending from the subpolar regions to the tropics. Marshes, sometimes called tidal marshes, occur between the high- and low-tide levels and require a significant tidal range. In general, they occur in low-energy, depositional environments, along estuaries and on the protected side of barrier islands. The most obvious characteristic of salt

Table 1 Total global area and primary production of shore zone ecosystems Environment

Surface area (106 km2)

Primary production rate (mol C m 2 yr 1)

Total production (1012 mol C yr 1)

Estuaries Mangroves Salt marshes

1.4 0.2 0.4

22 232 185

31 46 74

marshes is the dominance of cord grass, the most common being Spartina alterniflora and Spartina patens, and the black needle rush, Juncus roemerianus. These unique plants can tolerate large variations in salinity and account for much of the productivity of the marsh system although other plants, such as benthic diatoms living at the sediment surface, also contribute significantly to total productivity (Table 1). Together, these shoreline ecosystems provide critical habitats for many life forms including juvenile forms of fish, shellfish, and birds. As such, understanding, managing, and maintaining the overall health of these ecosystems are priorities of local municipalities and governmental agencies. The contribution of these ecosystems to the cycling of organic carbon within the local continental margin system depends critically on the balance between the production and respiration rates. Because of the magnitude of the rates, even a small difference would represent a significant net carbon transfer. The difficulty of determining the difference between production and respiration is compounded by natural variability; added complexity because organic carbon can be temporarily stored on the seafloor and then periodically transported by large storms; and the fact that human activities continue to alter the system. Based on numerous comparisons, it is currently thought that in most estuarine systems respiration is greater than production because of added organic input from the river while nutrient inputs to salt marshes and mangrove forests support production that exceeds respiration. Therefore, these latter ecosystems may export organic matter to adjacent continental shelves and other ecosystems. As shown in the next section, however, while the impacts of this carbon cycling are important to local ecosystems, they are relatively small when compared to transfers on the global scale.

Continental Shelf and Slope Ecosystems (The Continental Margin) Continental shelf and slope environments (hereafter referred collectively as the continental margin) occur



everywhere at the continent–ocean boundary. To facilitate the discussion, these environments can be divided into broad categories such as eastern and western boundary current, polar, subpolar, monsoonal, and tropical margins. Where there is significant exchange, inland seas are linked to the margin across which the major material exchange occurs. In the following, these broad categories will be used to generalize rates of organic carbon cycling and exchange. In addition to the regional distinctions noted above, local geomorphology of the margin also influences the magnitude and mixture of processes controlling carbon and associated bioactive element transfers. In the broadest sense, margins can be classified into two fundamental types, slope-dominated and shelf-dominated margins (Figure 2). These margin types are differentiated by (1) the location of the majority of primary production (seaward of the shelf break for slope-dominated, shoreward for shelf-dominated), (2) (a) High productivity + particle flux

Photic zone

Shelf

Slope

Rise

(b) High productivity + particle flux Photic zone

Shelf

Slope

Rise

Figure 2 Generalized representation of major distinctions between (a) a slope-dominated margin and (b) a shelfdominated margin. From Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and TalaueMcManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer.

255

the relative importance of nutrient input from upwelling versus benthic nutrient regeneration, and (3) the extent to which turbulence and water column mixing are supported by friction with the seafloor. These margin types are punctuated with other features such as river plumes and canyons while variations in factors such as local meteorological forcing and tidal ranges and currents add spatial variability, but nevertheless, all margin regions can be classified within this simple two end-member framework. At slope-dominated margins (Figure 2(a)), the majority of the primary production occurs seaward of the shelf break. Generally, but not exclusively, these margins exhibit relatively narrow continental shelves. Examples include eastern margins of the Pacific from Vancouver Island to central Chile, and the eastern margin of the Atlantic from the Iberian Coast to Cape Town. Residence time of water on the narrow, open shelf is relatively short. When upwelling, often wind-driven, occurs, a significant portion of upwelled waters is driven offshore, supporting high production rates seaward of the shelf break, over the continental slope and rise. In addition, benthic boundary layer processes act to resuspend particles facilitating downward lateral transport of materials off the shelf. Combined, these processes result in large transfers of particulate organic matter seaward of the shelf break, often for distances of 200–300 km. This may occur within the benthic boundary layer, within surface plumes extending offshore or in subducted intermediate layers in offshore plumes. It is difficult to generalize the width of this margin zone but 150–200 km seaward of the shelf break in the surface and 200–300 km in the benthic boundary layer are reasonable estimates. At shelf-dominated margins, on the other hand, the majority of primary production occurs shoreward of the shelf break. In general, these margins have broad shallow shelves (Figure 2(b)) although when bounded by western boundary currents, some narrowed-shelf systems may be classified as shelf-dominated margins. Examples are the eastern seaboard of the US south of Cape Cod, the broad shelf adjacent to southern Brazil, Uruguay, and Argentina, East China Sea, and North Sea. Due to their shallow nature, sufficient energy is transmitted to the seafloor to frequently resuspend sediments. Many of these systems are also characterized by low relief and low sediment flux from the adjacent land. The combination of winnowing away of the fine-grained sediments and slow replenishment leaves the majority of these shelves with nonaccumulating, coarse-grained sandy sediments. Recent studies have revealed, however, that while these sandy shelf sediments are generally depauperate in organic matter, high metabolic rates


256


are observed. This seeming paradox is most likely the result of efficient advective transport of substrates and metabolites into and out of these high permeability sediments. Friction with the seafloor and the shallow nature of these systems enhance water column mixing with fronts commonly found near the shelf edge. Nutrient inputs are dominated by dynamic processes at the shelf break that bring nutrient-rich ocean waters onto the shelf but may also be augmented via atmospheric deposition, river outflow, and groundwater discharge. On the western side of ocean basins, these systems are often bounded by western boundary currents such as the Gulf Stream, Kuroshio, and Brazil Currents and eddy interactions induce upwelling of nutrient-rich waters onto the shelf. Once upwelled, nutrient-rich waters either move onto the shelf via lateral exchange Table 2

processes or are subducted, restricting high biological productivity to the shelf break and shoreward. The high productivity and particle flux region in these systems is shoreward of the shelf break and sinking particles are intercepted by the seafloor and efficiently remineralized. Because shelf-dominated systems are generally wide with a shallow water column, exchange between ocean waters and shelf waters is reduced while shelf water residence times are relatively long, further promoting remineralization of particulate organic matter in these environments. The overall length and areas characterizing each major region type are listed in Table 2. The total length of the shelf break as shown in Figure 3 is 1.5 105 km. The focus here is on margin–open ocean exchange and the shelf break is defined as the

Summary of areas of major margin regions Shelf area (106 km2)

Rise area (106 km2)

Surface expression (106 km2)

Slope and rise area (106 km2)

Margin type

Shelf width (km)

Arctic – P

Shelf

253

13 897

3.51

2.78

1.39

3.51

4.17

Antarctic – P

Slope

134

16 332

2.19

3.27

1.63

5.46

4.90

NW Atlantic – SP NW Atlantic – WBC W Atlantic – T SW Atlantic – WBC SW Atlantic – SP NE Atlantic – SP NE Atlantic – EBC E Atlantic – T SE Atlantic – EBC Atlantic subtotal

Shelf Shelf Shelf Shelf Shelf Shelf Slope Shelf Slope

417 136 180 305 693 536 89 43 69

5 393 11 332 3 413 5 501 1 731 4 362 5 201 4 179 3 161 44 273

2.25 1.54 0.62 1.68 2.33 2.34 1.68 0.18 0.22 12.83

1.08 2.27 0.68 1.10 0.35 0.87 1.04 0.84 0.63 8.85

0.54 1.13 0.34 0.55 0.17 0.44 0.52 0.42 0.32 4.43

2.25 1.54 0.62 1.68 2.33 2.34 2.72 0.18 0.85 14.51

2.16 3.40 1.02 1.65 0.52 1.31 1.56 1.25 0.95 13.82

W Indian – M W Indian – T W Indian – WBC E Indian – M E Indian – T E Indian – EBC E Indian – SP Indian subtotal

Slope Slope Shelf Shelf Slope Slope Shelf

65 34 50 113 53 96 112

7 708 2 236 3 713 5 475 4 405 2 624 3 363 29 524

0.50 0.08 0.18 0.62 0.23 0.25 0.38 2.24

1.54 0.45 0.74 1.10 0.88 0.52 0.67 5.90

0.77 0.22 0.37 0.55 0.44 0.26 0.34 2.95

2.04 0.52 0.18 0.62 1.11 0.78 0.38 5.64

2.31 0.67 1.11 1.64 1.32 0.79 1.01 8.86

NW Pacific – SP NW Pacific – WBC W Pacific – T SW Pacific – WBC NE Pacific – SP NE Pacific – EBC E Pacific – T SE Pacific – EBC SE Pacific – SP Pacific subtotal

Shelf Shelf Shelf Shelf Slope Slope Slope Slope Slope

471 269 226 377 57 47 48 35 65

6 180 5 069 9 514 5 340 3 815 8 356 2 122 5 672 539 46 607

2.91 1.36 2.15 2.01 0.22 0.40 0.10 0.20 0.03 9.38

1.24 1.01 1.90 1.07 0.76 1.67 0.42 1.13 0.11 9.32

0.62 0.51 0.95 0.53 0.38 0.84 0.21 0.57 0.05 4.66

2.91 1.36 2.15 2.01 0.98 2.07 0.53 1.33 0.14 13.48

1.85 1.52 2.85 1.60 1.14 2.51 0.64 1.70 0.16 13.98

150 633

30.15

30.13

15.06

42.59

45.19

Global totals

Shelf break length (km)

Slope area (106 km2)

Region

For abbreviations, see Figure 3. From Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and Talaue-McManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer.



P

P

SP

257

P

SP

SP

SP

EBC

WBC WBC EBC T

T

M T

T

T EBC

M

T

EBC

WBC

WBC

WBC

SP SP P

SP P

P P P

Figure 3 Location of shelf break and distribution of major region designations for global integration. WBC, western boundary current; EBC, eastern boundary current; SP, subpolar; P, poloar; T, tropical; M, monsoonal. From Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and Talaue-McManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer.

top of the continental slope that is contiguous with the main abyssal ocean basins. This results in a shelf break length that is considerably shorter than the 3–4 105 km length one would estimate if a higherresolution coastline and bathymetry were used but probably better reflects the length of boundary across which ocean–margin exchange must occur. Note that fully enclosed inland seas such as the Baltic Sea, Mediterranean Sea, Hudson Bay, Arabian Sea, Red Sea, Persian Gulf, and the Sea of Japan are shoreward of the shelf break and are, therefore, included in the shelf area as defined in this simple, two-end-member framework. The total shelf area estimated from mean shelf width and shelf break length is 30.2 106 km2. The surface expression of the margin ecosystem is defined to be the shelf area for shelf-dominated margins and the shelf plus slope areas for slopedominated margins and is considerably larger than the shelf area alone (42.6 106 km2). With this definition, 12.6 106 km2 or 30% of the surface expression of the margin is in water depths greater than 200 m. The inclusion of a significant deep-water area within the margin system is an important difference between the description here and previous efforts but is consistent with the processes displayed in Figures 1 and 2. Note that all of these values are thought to be

minimum estimates for several reasons. At many margin regions such as eastern boundary current systems and the Argentine shelf, satellite imagery reveals elevated chlorophyll and presumed productivity at much greater distances from the continent than are used here. There is no definitive value or threshold between high coastal productivity and low gyre productivity at which to insert a boundary. The point is to identify regions that are sufficiently different from the open ocean because of the dominance of margin processes that they require specific study. The definition of the surface expression adopted here meets this criterion. Another reason these values are believed to be minimum estimates is that islands and the Canadian and Indonesian archipelagos have been excluded. Various estimates suggest that there are roughly 18 000 islands in the oceans. Many of these occur adjacent to shorelines and would not provide significant additional continental shelf break length or margin area. Additionally, many islands are clustered together and therefore share shelf and slope areas and shelf-break length. However, even if only 180 or 1% are considered separate islands or groups of islands protruding to the ocean’s surface and if it is assumed as a minimum, each is an infinitely small point surrounded by a 30-km-wide shelf and a


258


200-km-wide slope, these 180 islands would contribute 33 840 km of shelf break length, 0.59 106 km2 of shelf area, and 29.4 106 km2 of slope area. Thus, omission of islands is a serious deficiency in the present estimates of coastal carbon cycling that future efforts must address. All of the major ecosystem types contribute significantly to shelf break length (Table 3). Values range from c. 3.1 104 km for western boundary current systems to 1.3 104 km for monsoonal systems. A greater range is observed in the contribution of each to total global continental shelf area. Because of the preponderance of shallow marginal seas (e.g., North Sea, Bering Sea, and Sea of Okhotsk) in regions classified as subpolar, the subpolar margins account for 1.05 107 km2 or 35% of total shelf area. Globally, the polar and subpolar regions contain 54% of continental shelf area. Eastern boundary current and monsoonal systems contain only a small portion of the global shelf area (9.1% and 3.7%, respectively). A slightly different pattern is observed in the relative ranking of the surface areas of the margin types. Polar and subpolar systems still contain the largest amounts of sea surface area, 21% and 26%, respectively. However, now eastern boundary current and monsoonal systems provide 17.8% and 6.1%, respectively. This increase in the relative sea surface area reflects the inclusion of the offshore area that is still dominated by margin processes. Shelf- and slope-dominated margins each contribute significantly to total shelf break length (88 462 km and 62 171 km, respectively) while shelf-dominated margins account for 80% of global shelf area. However, because surface processes that extend over adjacent deeper waters are considered in defining margin boundaries, both margin classifications contri-

Table 3

bute relatively equally to margin sea surface area (24.1 106 km2 (shelf-dominated); 18.5 106 km2 (slope-dominated)). A total of 9.7 Pg C yr 1 (1 Pg ¼ 1015 g) is estimated for continental margin production (Table 4). This is c. 21% of the current estimates of global production (c. 48 Pg C yr 1) and is significantly larger than the total contributed by estuaries, salt marshes, and mangrove forests. Highest productivities are reported for the monsoonal and eastern boundary current regions (Table 4) with the lowest values attributed to the polar region. Interestingly, when productivities are integrated over each major ecosystem type, the contributions to total production from subpolar, western boundary current, eastern boundary current, and monsoonal regions are all roughly similar, between 1.9 and 2.7 Pg yr 1 (Table 5). The western boundary current system is lowest of the group but still significant due to contributions from broad shelves while the eastern boundary current system is highest. Also, similar contributions to total production are estimated for shelf-dominated and slope-dominated systems (5.0 vs. 4.6 Pg C yr 1, respectively). Of the estimated productivity, 3.5 Pg C yr 1 or 36% occurs over the slope where water depths are greater than 200 m. High export efficiency over the slope could provide an important conduit for the transfer of carbon to the deep sea via the biological pump. Direct exchange of CO2 with the atmosphere suggests a net uptake of 0.29 Pg C yr 1 ( 2.4 1013 mol yr 1; note that fluxes from the atmosphere to the ocean are expressed as negative fluxes). Consistent trends among the major ecosystem types are observed and summarized in Table 5. The majority of the influx is estimated to occur in the polar and

Summary of areas by major ecosystem types

Ecosystem type

Shelf break length (km)

Shelf area (106 km2)

Slope area (106 km2)

Rise area (106 km2)

Surface expression Slope and rise area (106 km2) (106 km2)

Polar Subpolar Western boundary current Eastern boundary current Tropical Monsoonal

30 229 25 583 30 955

5.70 10.46 6.77

6.05 5.08 6.19

3.02 2.54 3.10

8.97 11.33 6.77

9.07 8.15 9.27

25 014

2.74

5.00

2.50

7.74

7.50

25 869 13 183

3.36 1.12

5.17 2.64

2.59 1.32

5.11 2.66

7.76 3.95

Shelf-dominated Slope-dominated

88 462 62 171

24.06 6.10

17.69 12.43

8.85 6.22

24.06 18.53

26.53 18.65

150 633

30.15

30.13

15.06

42.59

45.19

Total

From Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and Talaue-McManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer.



259

Table 4 Primary production, air–sea and margin–ocean carbon exchange, and slope and rise of organic carbon deposition for margin regions Region

Primary production (g C m 2 yr 1)

Total production (1012 g C yr 1)

Air–sea exchange (mol CO2 m 2 yr 1)

Total air–sea exchange (1012 mol CO2 yr 1)

Deposition rate (mol m 2 yr 1)

Total deposition (1012 mol OC yr 1)

Arctic – P

59

207.1

2.4

8.49

0.11

0.46

Antarctic – P

80

436.5

0.8

4.53

0.56

2.74

NW Atlantic – SP NW Atlantic – WBC W Atlantic – T SW Atlantic – WBC SW Atlantic – SP NE Atlantic – SP NE Atlantic – EBC E Atlantic – T SE Atlantic – EBC Atlantic subtotal

175 350 350 350 292 175 300 150 450

393.8 539.0 215.6 588.0 680.4 409.5 816.1 27.0 382.6 4051.9

1.7 þ 2.5 0.0 þ 1.0 1.5 1.3 0.4 0.0 0.4

3.83 þ 3.85 0.00 þ 1.68 3.50 3.04 1.09 0.00 0.34 6.26

0.34 0.17 0.30 0.25 0.39 0.22 0.44 0.51 0.51

0.55 0.58 0.31 0.41 0.20 0.29 0.69 0.64 0.48 4.14

W Indian – M W Indian – T W Indian – WBC E Indian – M E Indian – T E Indian – EBC E Indian – SP Indian subtotal

365 150 175 300 150 225 175

745.9 78.6 32.2 186.0 167.1 174.6 65.6 1450.0

þ 1.4 0.0 þ 1.0 0.4 0.0 0.0 1.8

þ 2.86 0.00 þ 0.18 0.25 0.00 0.00 0.68 þ 2.12

0.59 0.30 0.30 0.59 0.30 0.23 0.21

1.36 0.20 0.33 0.97 0.40 0.18 0.21 3.66

NW Pacific – SP NW Pacific – WBC W Pacific – T SW Pacific – WBC NE Pacific – SP NE Pacific – EBC E Pacific – T SE Pacific – EBC SE Pacific – SP Pacific subtotal

175 350 188 150 300 345 225 500 300

509.3 476.0 404.2 301.5 294.6 713.2 118.4 665.2 42.8 3525.2

2.0 1.0 þ 0.1 þ 1.0 1.8 0.0 0.0 0.0 1.8

5.82 1.36 þ 0.22 þ 2.01 1.77 0.00 0.00 0.00 0.26 6.98

0.30 0.17 0.25 0.12 0.15 0.51 0.59 0.59 0.22

0.56 0.26 0.71 0.19 0.17 1.28 0.38 1.00 0.04 4.59

Grand total

24.14

9670.6

15.59

For abbreviations, see Figure 3. From Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and Talaue-McManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer.

Table 5 type

Summary of primary production, air–sea exchange, and slope and rise of organic carbon deposition by major ecosystem

Ecosystem type

Total primary production (1012 g C yr 1)

Total air–sea exchange (1012 mol CO2 yr 1)

Total slope and rise deposition (1012 mol OC yr 1)

Polar Subpolar Western boundary current Eastern boundary current Tropical Monsoonal

644 2396 1937 2751 1011 932

13.02 18.88 þ 6.36 1.43 þ 0.22 þ 2.61

3.20 2.02 1.78 3.63 2.63 2.33

Shelf-dominated Slope-dominated

5035 4636

19.02 5.12

6.67 8.92

Total

9671

24.14

15.59

From Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and Talaue-McManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer.


260


subpolar regions. Contrastingly, near-zero net exchange rates are generally reported for tropical and eastern boundary current regions and effluxes are reported for most western boundary current regions. Based on the values reported here, the majority (79%) of the air–sea exchange at margins is attributed to shelf-dominated margins. An important caveat for these conclusions, however, is to note that for many of the regions, no direct measurements have been reported and the values used here have been extrapolated from other, presumably similar, locations. The observed trends must be verified by future studies. In principle, the above estimate of CO2 uptake from the atmosphere defines the major role margins may play in mitigating the impacts of the release of anthropogenic CO2 to the atmosphere. However, along the margin, lateral exchange with terrestrial pools may provide another pathway by which CO2 input to the ocean may be augmented. Simply, productive coastal zone ecosystems such as salt marshes and mangrove forests may take up carbon directly from the atmosphere and release it to coastal waters as dissolved organic and inorganic carbon. Waterexchange processes may transfer it offshore, significantly enhancing the coastal–open ocean carbon flux. Recent studies at shelf-dominated margins, such as the North Sea, Georgia Bight, and East China Sea, in fact suggest that the net carbon transfer from the margin to the open ocean is significantly enhanced relative to what would be predicted from air–sea exchange alone. More measurements are required to yield a global-scale estimate but it is likely that the transfer of carbon from the margins to the open ocean is significantly larger than that estimated from air–sea exchange alone listed above. Integrating over the regions defined here suggests that 15.6 1012 mol of organic carbon is deposited on the adjacent slope and rise sediments. This value accounts for 47% of the estimated total deepseafloor carbon deposition (3.3 1013 mol organic carbon yr 1). Assuming that there is no dramatic difference in the transfer efficiency of sinking organic matter between the base of the main thermocline and the seafloor, this pattern of deposition should reflect the relative strength of the biological pump transfer of carbon to the deep ocean. Thus, continental margin ecosystems likely supply roughly one-half of the carbon transferred by the biological pump to the deep ocean. This conclusion is consistent with the global productivity assessment provided earlier if the margin export efficiency averages 3 times the open ocean average. As noted earlier, because a significant portion of the margin production occurs over deep water where sinking particles will not be intercepted by the seafloor, higher export efficiencies are likely.

Studies of how the biological pump may change with future climates must, therefore, include these quantitatively important environments.

Summary and Concluding Remarks Intense carbon cycling in coastal ecosystems is driven by a complex mixture of physical and biological processes. The impact of ocean boundaries on the global carbon budget is not well known but likely provides major pathways for transferring carbon among the important, climatically relevant reservoirs. With the caveat that significant uncertainty remains in the estimates of many of the important carbon transfers, results to date imply that 21% of global primary production, a likely higher proportion of the export production, a significant proportion of the net oceanic uptake of CO2, and nearly half of the biological pump transfer of carbon to the deep sea reservoir occur at the ocean boundaries, driven by coastal ecosystems. Ocean margins must be a priority in future marine research given the magnitude of these values, the concentration of human-induced pressures on coastal systems, and the role margins may play in the mitigation of and oceanic response to climate change.

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.

Further Reading Gattuso J-P, Frankignoulle M, and Wollast R (1998) Carbon and carbonate metabolism in coastal aquatic ecosystems. Annual Review of Ecology and Systematics 29: 405--434. Jahnke RA (in press) Global synthesis. In: Lui K-K, Atkinson L, Quinones R, and Talaue-McManus L (eds.) Global Change, The IGBP Series: Carbon and Nutrient Fluxes in Continental Margins: A Global Synthesis, 450pp. New York: Springer. Robinson AR and Brink K (eds.) (2005) The Sea, Vol. 13: The Global Coastal Ocean: Multiscale Interdisciplinary Processes. Cambridge, MA: Harvard University Press. Walsh JJ (1988) On the Nature of Continental Shelves, 520pp. London: Academic Press. Wollast R (1998) Evaluation and comparison of the global carbon cycle in the coastal zone and in the open ocean. In: Brink KH and Robinson AR (eds.) The Sea, vol. 10, pp. 213--252. New York: Wiley.


ORIGIN OF THE OCEANS K. K. Turekian, Yale University, New Haven, CT, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2055–2058, & 2001, Elsevier Ltd.

Introduction The oceans and the salts dissolved in them probably all formed early in the Earth’s history. Although planetary degassing is occurring, as evidenced by studies of rare gases in rocks, the oceans, and the atmosphere, there is also a loss of water to the mantle via subduction. Supply from and loss to the mantle may be roughly equal at present.

Acquiring an Ocean The oceans can be conceived of either as a feature established in the earliest history of the planet or as the result of a continuing supply of water and the constituents of sea water by degassing from the Earth’s interior. In 1951 W.W. Rubey, in his Presidential Address to the Geological Society of America, took the latter point of view. His arguments were colored by the knowledge available at that time of the age of the Earth and the age of the oldest rocks. Based on the analysis of lead isotopes in galenas, it was determined in the late 1940s that the Earth was about 3.2 billion years old. Some continental rocks, presumed to be relicts of ancient terrains, dated at about the same time also gave ages of about 3.2 billion years. On this basis it was assumed that the oldest rocks preserved a record of the dawn of Earth history. The contemporary oceans and oceanderived sediments contain chemical species in quantities far in excess of those available from the weathering of crustal rocks (Table 1). These components were called ‘excess volatiles’ by Rubey, but Harold Urey suggested that they were really better characterized as ‘excess solubles’. If these species all arrived with an early ocean, the early ocean would have had a radically different composition. It would dissolve rocks and also precipitate compounds different from those depositing from the present ocean. If that were indeed the case, Rubey argued, the initial rocks should show the effects of a sudden supply of ocean water and hydrochloric and sulfuric acids, and carbon dioxide that ultimately dissolved rocks and formed the saline sea. The ancient rocks, however, do not look appreciably different from

younger rocks; thus the absence of a difference in composition indicates that the oceans with their attendant excess anionic species have grown slowly with time. Indeed, it was argued that if a small fraction of the flux of water from fumaroles and hot springs were primary (from the Earth’s interior) rather than meteoric or surface recycled water, then, over time, the oceans could be added to the surface from the interior so that the volume was increasing with time. The discovery in 1955 that the Earth as a member of the solar system was really about 4.55 billion years old, and that the oldest rocks were considerably younger, ruled out having a record of the earliest days of the Earth’s existence. In addition, from measurements of the hydrogen and oxygen isotopes of hot springs and in some cases tracking radioactive tritium from nuclear tests in hot springs, it was clear that all or most of the water in hot springs, and fumaroles was meteoric, and therefore determining a primary water flux was virtually impossible. There is evidence, however, that there is planetary degassing, as revealed in the flux of radiogenic 40Ar (Figure 1) and primordial 3He (Figure 2) to the atmosphere. When these fluxes are used to model the flux of other gases (or their condensation products), two results are obtained. Gases that behave like 36Ar (the nonradiogenic argon isotope) appear to have arrived at the Earth’s surface in the earliest days of Earth history, while carbon dioxide and its condensation products, limestone and organic compounds, are being recycled via the processes associated with plate tectonics. One can assume that other chemically reactive analogues like water, behave in the same way. Like 36Ar water may have been at the Earth’s surface early in its history, and like carbon dioxide it is being recycled. Table 1 Components of the oceans, atmosphere, and sedimentary rocks not derivable by weathering of primary silicate rocks Chemical species

Amount on Earth’s surface not derived by weathering ( 1020g)

Water Total carbon as carbon dioxide Chlorine Nitrogen Sulfur

16 600 910 300 42 22

After Rubey (1951).


261

262

ORIGIN OF THE OCEANS

3.4

3

1

15 20

1

25

2.2 lid So

2.0

Depth (km)

2.4

42

2

10

2.6

Argon-40 (10 atoms)

5

He (%) 4 3

5

K

Pr o

2.8

40

3.0

6

STN. 7 0

d

du

3.2

by n io y t c ca e

rth Ea

1.8

30

2

35

30

3

25

1.6 1.4

4

20

East Pacific Rise 0 1000 km 1000 km

1.2 1.0

h osp Atm

0.8

ere

5 140°W

130°W

120°W

110°W

100°W

90°W

Longitude

0.6 0.4 0.2 0 0

0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Initial

9

Time (10 years)

Present

Figure 1 The degassing of radiogenic 40Ar (from the radioactive decay of 40K) from the solid Earth can be expressed by the 40 40 ðtÞ=dt ¼ lK0 expðltÞ pexpðbtÞ XSE ðtÞ, where equation XSE 40 (t) is Ar in the solid Earth, l is the radioactive decay constant X40 SE for 40K, a is the first-order degassing constant, and b is a measure of the exponential decrease in the efficiency of degassing over time as the result of the decrease in heat production from radioactive decay and loss of initial heat. K0 is the amount of 40K in the planet 4.55 109 years ago that will decay to 40Ar. The lowest curve represents the growth of 40Ar in the atmosphere over the past 4.5 billion years. The application of the values of a and b derived from the 40Ar record when applied to the nonradiogenic isotope of argon (36Ar) indicates that the nonradiogenic rare gases in the atmosphere largely had to be supplied to the atmosphere early in the Earth’s history. a ¼ 1:4 109 ; b ¼ 1:2 109 ; K0 ¼ 3:85 1042 . Present-day flux of 40Ar to atmosphere 1.5 1031 atoms per year.

We are thus left with the possibility that the oceans, including both the water and the soluble salts found therein, were present at the Earth’s surface right at the beginning and subsequently have been subjected to recycling between crust and mantle via the plate tectonic cycle.

The Loss of Water from the Earth’s Surface If the oceans were effectively at the surface at the beginning, what processes could diminish the

Figure 2 A section across the East Pacific Rise in the South Pacific showing 3He in excess over that expected from the atmospheric 3He dissolved in sea water when scaled to the more common 4He. The difference between the dissolved isotopic ratio and atmospheric ratio is represented by d3 He in percentage difference. 3He is assumed to be primordial and indicates degassing of the mantle at the divergent plate boundaries featured as oceanic ridge systems. STNstation number of the sampling. (Reprinted with permission from Lupton and Craig 1981, copyright American Association for the Advancement of Science.)

original inventory? There are two ways to lose water from the Earth’s surface: (1) the photolytic dissociation of water vapor in the atmosphere with subsequent loss of hydrogen to space and the utilization of the released oxygen to oxidize reduced iron and sulfur compounds from the mantle; and (2) the loss of water, as well as of the oxidized components, to the mantle by way of hydrated minerals (such as clay, minerals, micas, and amphiboles) and minerals with oxidized iron and sulfur. Although the first process is occurring today, it is not very efficient because of recombination with the abundant oxygen in the atmosphere as well as the fact that the supply of solar hydrogen to Earth may compensate for the loss of hydrogen. In the early history of the solar system, the flux of high-energy photons from the sun is thought to have been much greater than it is today, resulting in a very efficient conversion of water to hydrogen and oxygen. The stream of hydrogen from the atmosphere to outer space may have entrained heavier gases and caused the mass fractionation of the rare gases relative to each other and the fractionation of the isotopes of individual rare gases relative to solar composition.



2 MgSiO3 þ Fe þ 2 H2 O -Mg2 SiO4 þ Fe2 Sio4 þ 2 H2

14 MFL 13

25%

22

Ne/ Ne

12

50%

11 75% 10

0.03

½1

Sources of Water Water reached Earth by one of two processes. The first is by way of the capture of water from the solar nebula as it cooled down. There is evidence from neon isotopes (Figure 3) that the initial imprinting of Earth with gases was likely to have been of solar composition. If this were the case, then the amount of water trapped by the accreting planet would have been large and provided the fundamental source of water. A second source is cometary impacts. Comets have often been described as dirty snowballs. They are rich in organic compounds and ice. Carbonaceous chondrites may be related to comets and, if so, their rare gas isotopic signatures are characteristically different from solar abundances and isotopic compositions owing to adsorption processes. Therefore, if water tracks the rare gases and if the neon isotope signature of mantle rocks indicates a solar origin for the terrestrially accumulating gases rather than a cometary origin, then most of the oceans owe

Diamonds MORB Somoa Loihi Other OIB

A

9 0.02

To oxidize the mass of the Earth with this initial make-up to the present oxidized level would require the dissociation of 60 present-day ocean volumes. Clearly, the results indicate hydrogen loss and planetary oxidation, although the exact starting material is not precisely defined. In either case, the oxidation of the planet is the consequence of the photodissociation of water vapor and loss of hydrogen to space. In order to fractionate the rare gases by entrainment in the hydrogen streaming from Earth, this process must have occurred almost instantaneously in the early history of the Earth. Therefore, it is reasonable to assume that both water and rare gases were at the Earth’s surface virtually at the time of formation of the Earth.

M

P

20

We can estimate how much water was dissociated by considering the oxidation of the mantle. If the Earth started out with the oxidation state of chondrites (although not necessarily of total chondritic composition), it would take the oxygen from about one present-day ocean volume to reach the oxidation state of the mantle inferred from mantle-derived rocks. Alternatively, if we start with a more reduced ensemble, suggested by condensation calculations from a solar nebula, we will have primarily enstatite (MgSiO3) and metallic iron as the original source to be oxidized. The oxidation reaction would then be as shown in eqn [1].

263

0.05

0.04 21

0.06

0.07

0.08

22

Ne/ Ne

Figure 3 The distribution of neon isotopes in mantle-derived rocks, indicating the presence of an atmospheric component, a radiogenic component adding 21Ne (produced by neutrons from uranium fission acting on oxygen and magnesium), and a solar component. It is this latter that indicates that gases in the mantle were derived from the capture of solar material in the early history of the Earth. M ¼ MORB (midocean ridge basalts); P ¼ plume or ocean island basalts (OIB); A ¼ atmosphere. Solar neon is represented by the horizontal line at 20Ne/22Ne ¼ 12.5; MFL is the mass fractionation line. The presence of solar neon in ocean basalts was first identified by Craig and Lupton (Craig H and Lupton JE (1976) Earth and Planetary Science Letters 31: 369– 385). (Reprinted with permission from Farley and Poreda (1993).

their origin to direct capture of gases in the solar nebula cooled sufficiently to allow the production of water molecules. There have been claims that absorption spectra of sunlight indicate house-sized blocks of cometary ice entering our atmosphere. If the claim were substantiated the calculated rate of influx of these blocks of ice would be sufficient to provide the present volume of the oceans. However, these claims have not been independently substantiated and the consensus is that such blocks of ice are not responsible for the observations. Unless new measurements support this suggestion, we are constrained to accept the hypothesis of an initial ‘watering’ of the planet rather than a gradual accumulation.

The Composition of the Oceans If indeed the oceans were present virtually from the day the Earth was formed, there is a further question: Was it salty like the present ocean? The salts dissolved in the ocean were probably dissolved from the original materials composing the Earth if low-temperature condensation compounds such as the


264


chlorides, sulfates, and carbonates were present. Indeed, salts of these anions have been leached from primitive meteorites such as the carbonaceous chondrites as well as some other meteorites that have undergone some reheating. On the basis of theoretical calculations, similar compounds should have been present in the accreting Earth. With leaching of these compounds by the original water released from the Earth early in its history, it can be assumed that the oceans were always salty. Indeed, the argument that oceans over several hundred million years ago might actually have been twice as saline as the contemporary ocean is based on the fact that there are many geological salt deposits and deep brines, often associated with oil fields. The assertion is based on a reasonable premise that salt deposits became important about 400 million years ago. The salt prior to that time had to be stored in the oceans, thus increasing its salinity by a factor of about two higher than the contemporary ocean.

production by radioactive nuclides, both of which are waning with time. Therefore, the rate of supply of water to the mantle is now diminishing and there may actually be a release of the water stored in the mantle from previous times. As in the case of carbon dioxide, we may be in a steady-state of water supply from the mantle and return of water to the mantle, thereby maintaining the size of the oceans. At any rate, changes in the volume of water will probably not be large in future.

See also Conservative Elements. Elemental Distribution: Overview. Hydrothermal Vent Fluids, Chemistry of. Mid-Ocean Ridge Geochemistry and Petrology. Volcanic Helium.

Further Reading The Future of the Oceans If most of the water found on Earth was primarily in the oceans, with some dissolved in a molten mantle existing in the early days of the history of the Earth, then as we have seen, there was a decrease in the size of the original oceans. This decrease occurred because of photolysis of water vapor and subsequent loss of hydrogen from the atmosphere or the entrapment of water in hydrated minerals that were then subducted into the mantle. The rate of photolytic loss must be considerably smaller at present compared to that on the early Earth because of the decrease in the extreme ultraviolet flux from the Sun. Also, the rate of subduction of the hydrated crust must be less now than it was early in the Earth’s history because the driving forces for mantle convection and thus plate tectonics are gravitational heat from accumulation and fractionation and heat

Craig H (1963) The isotopic geochemistry of water and carbon in geothermal areas. In: Tongiorgi E (ed.) Nuclear Geology on Geothermal Areas, Proceedings of the First Spoleto Conference, Spoleto, Italy, pp. 17--53. Pisa: V. Lischi Figli. Farley KA and Poreda RJ (1993) Mantle neon and atmospheric contamination. Earth and Planetary Science Letters 114: 325--339. Kump LR, Kasting JF, and Crane RC (1999) The Earth System. London: Prentice-Hall. Lupton JE and Craig H (1981) A major helium-3 source at 151S on the East Pacific Rise. Science 214: 13--18. Lupton JE and Rubey WW (1951) Geologic history of sea water: an attempt to state the problem. Geological Society of America Bulletin 62: 1111--1147. Turekian KK (1990) The parameters controlling planetary degassing based on 40Ar systematics. In: Gopalan K, Gaur VK, Somayajulu BLK, and Macdougall JD (eds.) From Mantle to Meteorites, pp. 147--152. Delhi: Indian Academy of Science.


OVERFLOWS AND CASCADES G. F. Lane-Serff, University of Manchester, Manchester, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction When dense water enters an ocean basin it often does so at a shallow depth passing through a narrow strait or over a sill from a neighboring basin or marginal sea, or flowing down the continental slope from a shelf sea. The dense water flow is affected by the Earth’s rotation and so does not flow directly downslope but instead turns to flow partly along the slope. Large-scale, continuous flows from one ocean basin or large sea down into another are referred to as ‘overflows’, while smaller-scale, intermittent flows of dense water from shelf seas down continental slopes are referred to as ‘cascades’. In both cases, the flows can be summarized as gravity currents on slopes in a rotating system (see Rotating Gravity Currents). The deep ocean is divided into many sub-basins by ridges, so that overflows play an important role in the global thermohaline circulation, providing the mechanism by which dense waters created in the polar regions flow from one ocean basin down into another. The flow speeds in overflows can be 10–100 times faster than most other deep-ocean flows, generating turbulence and entraining ambient seawater into the overflow. Thus the mixing of deep waters is often dominated by the mixing that happens at overflows. The combination of density contrast, slope, and rotation produces complicated dynamics that are still not fully understood. The effects of the overflow are not limited to the waters immediately around the overflow, but can extend up through the water column producing, in some cases, eddies that can be observed at the sea surface, hundreds of meters above the overflow. These flows can be characterized in terms of the density contrast, flow rate, slope, water depth, and effect of the Earth’s rotation.

Observations Waters of increased density are formed at a wide range of locations throughout the globe. In low latitudes, waters of increased density are formed through evaporation increasing the salinity, such as in the Mediterranean Sea and the Red Sea. At mid-latitudes,

wintertime cooling on continental shelves leads to the formation of cold, dense waters that flow down the continental slope into the deep ocean. In polar regions cooling again increases the density, but this can be further enhanced by brine rejection (and thus increased salinity) during ice formation. The salinity of the Mediterranean is approximately 2 parts per thousand more than the Atlantic. The resulting density difference drives an exchange flow through the Strait of Gibraltar, with fresher Atlantic water entering the Mediterranean at the surface while the denser, more saline Mediterranean water flows into the Atlantic beneath. The flow of Mediterranean water through the strait is approximately 0.7 Sv (1 Sv ¼ 106 m3 s 1). The dense water descends the slope in the eastern Gulf of Cadiz into the Atlantic, with the flow deflected to the right by Coriolis forces and ambient Atlantic water entrained into this fast-moving flow (Figure 1). The path of the Mediterranean water is strongly influenced by the local bathymetry, especially by canyons that cut the continental slope and provide ‘shortcuts’ into deeper water. With the strong entrainment of Atlantic water the overflow reaches a neutral level at a depth of approximately 1000 m, where the density of the overflow matches that of the ambient seawater. The Mediterranean waters then flow along the continental slope at this level, although rotating lenses of water are shed from the flow, perhaps through the strong changes in direction of the flow caused by the capes along the flow path. These relatively well-mixed lenses of high-salinity water (known as ‘meddies’) have been observed to propagate around the North Atlantic as coherent features for considerable periods of time. On the northwest European shelf, wintertime cooling results in colder waters on the shelf than in the neighboring deeper waters. In the deeper waters, the cooling produces a surface mixed layer that is deeper than the waters on the shelf. For the same heat loss, the shallower layer of water on the shelf becomes colder than the deeper surface layer offshelf. Furthermore, during the spring the off-shelf, deep waters are replaced more quickly by warmer water from the south while the flows of waters onto and off the shelf are relatively slow. As the surface waters on the shelf begin to warm and stratify, the deeper waters on the shelf retain their lower temperatures, becoming significantly colder than any nearby waters. These waters tend to flow off-shelf in intermittent cascades. While cascades are smaller in


265

266

OVERFLOWS AND CASCADES

(a) 36⬚ 30′

20

1 ms⫺1

(b)

0

G

E

F

D C BA 38

36.0

Depth (m)

300

F

1000

Salinity (PSU)

500

B

.0

35.6

700 36.0

900

E

.6

37

.4 38

600

36⬚ 00′ N

H 0 100

36.8

36.4

A 1100

D C 35⬚ 30′ W 7⬚ 30′ W

7⬚ 00′ W

6⬚ 30′ W

6⬚ 00′ W

Figure 1 The Mediterranean outflow. (a) The peak outflow velocities in the region just to the west of the Strait of Gibraltar, with depths given in meters. Note how the flow is initially downslope before turning and flowing along the slope. (b) Salinity section along the axis of the outflow, with the letters A–F corresponding to the lines of stations in (a) (positions G and H lie beyond the regions covered by (a)). Reproduced from Price JF, Baringer MO, Lueck RG, et al. (1993) Mediterranean outflow mixing and dynamics. Science 259: 1277–1282.

magnitude and shorter in duration than overflows, they obey the same dynamics, turning under the influence of the Earth’s rotation to flow along rather than directly downslope. Cascades are also observed flowing off the shelf to the east of Bass Strait, between Tasmania and the mainland of Australia. These cascades reach a flow rate of up to 3 Sv, and have a typical duration of 2–3 days. More generally, throughout the world’s oceans, typical cascade fluxes (measured at the shelf edge) are c. 0.05–0.08 Sv per 100 km of shelf. Dense waters formed in the Arctic Ocean flow south through the Denmark Strait, which lies between Iceland and Greenland, forming one of the largest overflows with a flux of 2.7 Sv. The sill depth at Denmark Strait is approximately 700 m and the overflow descends down into the North Atlantic reaching a depth of between 1000 and 1500 m before flowing along the slope. The overflow remains at approximately this depth for at least 500 km. It forms cyclonic eddies in the overlying water which have been tracked by deep-drogued buoys (Figure 2). Measurements from these, and from moored current meters, show that these eddies have a regular frequency and a strong vorticity. Similar overflows are observed in the Southern Hemisphere. Dense ice shelf water (with a temperature below the surface freezing point) flows out of the Filchner Depression in the southeast corner of the Weddell Sea. This flow (with a flux of 0.5 Sv) mixes as

it flows down into the Weddell Sea, eventually contributing to the formation of Antarctic Deep and Bottom Waters. A flow in the opposite direction can be found in the southwest corner of the Weddell Sea, where high-salinity shelf water, formed by brine rejection during sea ice formation in open water, flows southward underneath the Ronne Ice Shelf. In the case of flow under an ice shelf, the dynamics of the flow is complicated by presence of the ice shelf above.

Flow Characteristics Parameters

Overflows and cascades are formed by sources of dense water on slopes in a rotating system. The source and the system into which the source is flowing can be characterized by five main parameters (Figure 3): 1. the source flux (Q), volume per unit time; 2. the density contrast between the source water and the ambient seawater (Dr); 3. the tangent of the slope angle (b); 4. the local value of the Coriolis parameter (f ); and 5. the depth of ambient seawater above the source (H). Although the overflow may initially be constrained by flowing through a narrow strait or channel, the width and depth of the overflow are not generally



42⬚

40⬚

W 36⬚

38⬚

34⬚

32⬚

3 .4

65⬚

64⬚

5 .8

8.3

500

267

30⬚

A 1500

65⬚

2500

2000

64⬚

0

15

.4

63⬚

4 9. .5 10 4 . 11 .6 12

500 100

63⬚

23.8 20.5

N

B

C

62⬚

N

62⬚

3.8 5.9 4.3 8.0

61⬚

61⬚

8.9 3000

12.9 16.4

60⬚ 42⬚

60⬚ 40⬚

38⬚

36⬚ W

34⬚

32⬚

30⬚

Figure 2 Trajectories of satellite-tracked buoys trapped in cyclonic eddies on the East Greenland continental slope, downstream of the Denmark Strait overflow. Tick marks give time in days, depths are given in meters. The flow here is mainly along the slope, with a superimposed cyclonic (anticlockwise) motion. Reproduced from Krauss W (1996) A note on overflow eddies. Deep Sea Research Part I 43(10): 1661–1667.

f Q Surface

H + Δ

tan =

Figure 3 The basic parameters and behavior of overflows. A source (flow rate Q) of relatively dense fluid (r þ Dr, where the ambient seawater has density r) flows onto a slope (of tangent b) in water of depth H. The local value of the Coriolis parameter (which varies with latitude) is f. Initially the fluid flows down the slope, turning to the right (in the Northern Hemisphere) under the influence of the Earth’s rotation until the main flow is flowing along the slope. A thin viscous sublayer continuously drains the main flow, taking fluid at an angle down the slope.

independent parameters. The shape of the overflow (and thus its width and height) adjusts in response to the effects of gravity and rotation as it flows over the slope. Where the ambient seawater is stratified, the

last parameter (the depth of ambient water above the source, H) may be replaced by a vertical length scale dependent on the strength and nature of the stratification (see below). Strong currents in the ambient seawater into which the overflow is flowing have an effect on the overflow, as do irregularities in the slope. In particular, channels and canyons can direct the dense fluid downslope much more rapidly than a similar flow on a smooth slope.

Basic Behavior A basic description of the behavior of overflows can be given based on the results of laboratory experiments, field observations, and numerical models. As the dense water leaves the channel or flows over the sill that marks the source at the top of the slope, it initially flows directly down the slope under the influence of gravity. The effects of the Earth’s rotation deflect the current (to the left in the Southern Hemisphere, to the right in the Northern Hemisphere) so that the flow curves to eventually flow mainly along the slope, maintaining its depth. The distance over which the adjustment from downslope


268


scales have been proposed as the appropriate Rossby radius, the first based on the dynamics of a rotating gravity current and the second based on the injection of fluid of the same density as the ambient seawater. These depend on the source flux, the local value of Coriolis parameter (which varies with latitude), and (for the first type) the density contrast between the dense current and the ambient seawater, with R1 ¼ ð2Qg0 Þ

Figure 4 Columns of ambient fluid taken into deeper water stretch and become thinner. Viewed from a stationary observer outside the rotating Earth, columns that appear stationary on the Earth are actually rotating (with vertical vorticity f ). When the column becomes thinner it must spin even faster to conserve angular momentum (as ice skaters spin faster when they draw their arms in), giving it cyclonic vorticity relative to the Earth. (Cyclonic is anticlockwise in the Northern Hemisphere, clockwise in the Southern Hemisphere.)

1=4 3=4

f

Scalings We make use of the parameters introduced above, but it is useful to express density contrast in terms of the reduced gravity g0 ¼ (Dr/r)g, where g is the acceleration due to gravity and r is the density of the ambient seawater. Two different horizontal length

R2 ¼ ðQ=f Þ1=3

Typically both these scalings give a Rossby radius of order 10 km for the cases described above. If a column is taken into deeper water by a distance R over a slope of tangent b, then the column will increase in height by an amount bR. This increase in height divided by the original height, H, gives the relative stretching of the water columns and thus gives useful nondimensional parameters (depending on the version of R used): G ¼ bR1 =H

to along-slope flow takes place scales with the Rossby radius of deformation (discussed below). The along-slope flow maintains an inviscid geostrophic balance, but is continuously drained by a viscous Ekman layer at its base. The viscous draining flow takes fluid from the base of the current at an angle down the slope. The inviscid along-slope flow is not always steady. The overflow carries columns of overlying ambient seawater out into deeper water so that these columns of ambient seawater will be stretched (Figure 4). This stretching produces cyclonic vorticity in the columns as they conserve their angular momentum. The vorticity in the overlying water breaks up the dense overflow into a series of domes. The cyclonic eddies in the ambient water lie above the domes of dense overflow water and the eddy-dome structures propagate along the slope (but the dense fluid is still continuously drained by the viscous sublayer). The downslope motion of columns of ambient fluid may occur in the initial downslope flow and adjustment near the source, or as a result of instabilities in the along-slope flow further from the source (see the section ‘Instabilities and mixing’).

and

or

G ¼ bR2 =H

In practice, these parameters are very similar in magnitude and it has yet to be established which is the most useful in describing the behavior of overflows. Comparisons with experiments show that an increase in these parameters (i.e., increased stretching) gives an increase in the frequency of eddy production. Where the ambient seawater is stratified (e.g., by a strong thermocline above the source) this can put an effective ‘lid’ on the vertical influence of the flow and the total depth H is replaced by a height scale derived from the stratification (this would be the height of the thermocline above the source for the simple example mentioned earlier, but a height scale can be obtained for more complicated stratification too). In Antarctica a more physical lid is provided by floating ice shelves, and eddies generated by dense high-salinity water flowing under the ice shelves are affected by the sudden changes in ambient water depth at the front of ice shelves, as well as more gradual depth changes beneath them. Where eddies are formed that extend up to the underside of an ice shelf this will have an effect on heat transfer and thus melting, although the scale of the contribution of these eddies to ice shelf melting has yet to be quantified. The appropriate scales for the speed and thickness of the viscous draining layer can also be estimated. This makes it possible to estimate the along-slope distance over which the initial flux is entirely drained into the viscous Ekman layer:


Y¼

Qf 3=2 pffiffiffiffiffiffi g0 b 2v


where v is the molecular viscosity for laboratory experiments (which have smooth, laminar boundary layers) or a vertical eddy viscosity for oceanographic flows (which have turbulent boundary layers). Dividing Y by a Rossby radius gives a nondimensional parameter that can be used to describe the importance of the draining flow, for example:

269

The draining flow is expected to dominate when this parameter is small. In laboratory experiments it has been found that the speed at which the eddies propagate along the slope is determined by Y/R, with slower propagation speeds for low values of Y/R and no eddies at all for sufficiently small values (Y/Ro1, approximately).

Streamtube models give a simple mathematical formulation that can be used to analyze some of the aspects of overflows. However, the averaging of properties throughout the flow (especially momentum) leads to a somewhat misleading picture of the flow. While the continuously drained along-slope flow (described earlier) does have a mean center position that moves gradually downslope, thinking of the flow in that way does not help comparisons with observations. In many overflows the effect of the bottom friction is not distributed throughout the overflow, but confined to a lower boundary layer, giving the flow sketched in Figure 3. However, there are some strong, turbulent overflows (e.g., Mediterranean outflow) in which properties are well mixed and the streamtube approach provides a useful model for direct comparison with oceanographic observations and numerical models.

Modeling and Interpretation

Laboratory Models

Y=R1 ¼

Q3=4 f 9=4 23=4 g05=4 v1=2 b

Streamtube Models

In developing a mathematical description of overflows there has been considerable use of ‘streamtube’ models. These are ‘integral’ models (meaning that the results can be obtained by integrating the equations of motion over planes perpendicular to the flow) that describe the flow in terms of mean properties as a function of the distance (e.g., s) along the flow centerline (Figure 5). Thus the overflow is described by a mean centerline speed, U(s), density contrast, Dr(s), and shape parameters (e.g., width and height w(s) and h(s)). With prescribed parameterizations for the entrainment of ambient seawater into the flow and for the frictional drag at the base of the flow, the path of the overflow centerline (x(s), y(s)) can be calculated.

w (s)

h (s) U (s) s

Figure 5 Streamtube models represent the overflow in terms of a mean centerline position with mean overflow properties described as functions of the distance (s) along the centerline.

Laboratory experiments allow a wide range of parameters and flows to be examined in detail and much of the insight into the behavior of overflows has come from laboratory modeling. Experiments are conducted in tanks mounted on rotating tables to simulate the effects of the Earth’s rotation (see Figure 6 for an example). Typically the scale of the laboratory tanks is of the order of p1 m, but rotating tables with diameters of up to 13 m are used. In order to apply the results from the laboratory experiments to oceanographic flows it is necessary to scale the parameters carefully. The inviscid aspects of the flow can be simulated with confidence in the laboratory (e.g., by matching the values of the nondimensional stretching parameters) so that the alongslope flow and eddy formation processes are likely to be well represented. However, in the laboratory experiments the viscous draining layer is smooth and usually laminar and driven by molecular viscosity, while in the corresponding oceanographic flows the Ekman layer is a turbulent boundary layer. It is therefore necessary to identify an effective vertical eddy viscosity in the oceanographic flows to allow comparison of the draining flow with the laboratory experiments. Reasonable choices of the turbulent eddy viscosity have allowed successful comparisons between the laboratory experiments and oceanographic measurements, but a better understanding of the turbulent oceanic bottom boundary layer would allow more reliable predictions. In addition to flows where the ambient fluid is homogeneous, experiments have also been conducted where the flow is directed into a stratified ambient fluid. As mentioned earlier, the effect of


270


Figure 6 A sequence of photographs of a laboratory experiment (plan view). Dense fluid (dyed) is released on an axisymmetric hill. The inviscid core flows around the hill at constant depth (darkest fluid) while a thin layer of fluid drains down the slope (toward the edge of the image). The dense fluid breaks up into a series of domes with ‘subplumes’ extending downslope from the domes. Reproduced from Baines PG and Condie S (1998) Observations and modelling of Antarctic downslope flows: A review. In: Jacobs SS and Weiss RF (eds.) Antarctic Research Series, Vol. 75: Ocean, Ice, and Atmosphere – Interactions at the Antarctic Continental Margin, pp. 29–49. Washington, DC: American Geophysical Union.

stratification is to reduce the effective vertical height scale, thus replacing the fluid depth H with a smaller height. This increases the stretching parameters and enhances eddy production. There is some evidence that the propagation speed of the eddies along the slope is reduced by the presence of stratification, perhaps because of internal wave generation. In experiments with a homogeneous ambient fluid, any mixed fluid created at the interface between the overflow and the ambient fluid is still denser than the ambient fluid and eventually rejoins the current. Where the ambient fluid is stratified, mixed fluid may have the same density as ambient fluid above the final depth reached by the main overflow and thus may detrain from the overflow at shallower depths. This distributed injection of overflow water into a stratified ambient has been observed in detail in nonrotating experiments but has yet to be quantified in rotating experimentsNon-Rotating Gravity Currents.

Denmark Strait overflow was treated. Of the numerical models in that comparison, the isopycnic model (which treats the ocean as made up of a series of layers of uniform density) seems the most promising, since it had too little mixing (when compared with oceanographic observations), whereas the standard level model (with ‘step-like’ bottom topography) and sigma-coordinate model (with terrain-following coordinates) had too much mixing. Numerical models of idealized geometries (similar to the laboratory models described above) with high resolution have shown similar results to the laboratory models, with both the formation of eddies and draining downslope flow represented provided that the resolution is high enough. These simpler models give useful insights into how numerical ocean models might be improved. In addition to resolution and mixing, the way the bottom drag is treated is also important. The inclusion of special benthic boundary layer models in future ocean models may address this problem.

Numerical Models

In numerical ocean models the ocean is generally divided into a series of boxes, the horizontal dimensions of which are at least 10 km and often much larger (100 km or so for many climate simulations). These boxes are too large to represent overflows accurately since typical overflow widths are of the order of 10 km or less. There are also problems with the accurate representation of mixing at overflows, which often has a dramatic effect on the wider ocean circulation. For example, in a recent project to compare three different models of the North Atlantic, the different models showed very different behavior for meridional overturning and water mass formation. These differences were found to be strongly dependent on the way the mixing of the

Instabilities and Mixing

One of the striking features of overflows is their ability to generate strong cyclonic vortices (or eddies) in the overlying water. These have been observed above the Denmark Strait overflow (Figure 2) and even above overflows beneath the Ronne Ice Shelf (Antarctica). These eddies fall into two main groups: strong eddies that form very close to the source (referred to as PV for potential vorticity eddies) and those that form after the along-slope flow has been established (BI for baroclinic instability eddies). In flows with PV eddies, much of the dense fluid is concentrated in the domes moving along the slope (although with a thin viscous draining layer), while the BI eddies are accompanied by more substantial



‘subplumes’ of dense fluid extending downslope from the flow. The BI eddies have a weaker effect on the overlying fluid and appear to have a weaker effect on mixing. Both types of eddies are observed in laboratory experiments and numerical models, and the instabilities leading to BI eddies have been analyzed mathematically. From the mathematical studies an ‘interaction’ parameter has been defined: m ¼ d=bRH where d is the depth of the overflow current as it flows along the slope and RH is a Rossby radius based on the total depth of the fluid: pffiffiffiffiffiffiffiffi g0 H RH ¼ f Thus bRH is the vertical scale corresponding to horizontal motions of scale RH on a slope of tangent b, and the interaction parameter compares this vertical scale to the depth of the current. Although the mathematical analysis is formally valid for small values of m, laboratory experiments suggest that BI eddies occur for large m, while PV eddies occur for small m. While these eddy types have been identified, and the conditions under which they occur in laboratory experiments have been approximately defined, the clear characterization of eddies in oceanographic flows has yet to be attempted. The viscous draining flows and mixing behavior of real overflows have also not been quantified in much detail, although shallow layers of relatively unmixed fluid extending downslope with a more mixed flow remaining at constant depth have been observed (e.g., downstream of the Faeroes–Shetland channel in the Northeast Atlantic).

Summary The flow of dense fluid down slopes occurs at a range of scales in the oceans, from small, temporary cascades from shallow shelf seas to the large, continuous flows into major ocean basins. These flows play an important role in controlling the large-scale thermohaline circulation, but their representation in numerical ocean models (especially those used for climate prediction) is poor because of problems with resolution, mixing, and bottom drag. Overflows

271

display a rich behavior and can have a strong effect on the overlying seawater. This behavior is being illuminated through laboratory experiments, numerical and mathematical models, and increasingly detailed and sophisticated field observations.

See also Bottom Water Formation. Flows in Straits and Channels. Fluid Dynamics, Introduction, and Laboratory Experiments. Meddies and SubSurface Eddies. Non-Rotating Gravity Currents. Ocean Circulation: Meridional Overturning Circulation. Rotating Gravity Currents. Topographic Eddies. Turbulence in the Benthic Boundary Layer.

Further Reading Baines PG and Condie S (1998) Observations and modelling of Antarctic downslope flows: A review. In: Jacobs SS and Weiss RF (eds.) Antarctic Research Series, Vol. 75: Ocean, Ice, and Atmosphere – Interactions at the Antarctic Continental Margin, pp. 29--49. Washington, DC: American Geophysical Union. Griffiths RW (1986) Gravity currents in rotating systems. Annual Review of Fluid Mechanics 18: 59--89. Hansen B and Osterhus S (2000) North Atlantic–Nordic Seas exchanges. Progress in Oceanography 45(2): 109--208. Ivanov VV, Shapiro GI, Huthnance JM, Aleynik DL, and Golovin PN (2004) Cascades of dense water around the world ocean. Progress in Oceanography 60: 47--98. Krauss W (1996) A note on overflow eddies. Deep Sea Research Part I 43(10): 1661--1667. Lane-Serff GF and Baines PG (2000) Eddy formation by overflows in stratified water. Journal of Physical Oceanography 30(2): 327--337. Price JF and Baringer MO (1994) Outflows and deep-water production by marginal seas. Progress in Oceanography 33(3): 161--200. Price JF, Baringer MO, Lueck RG, et al. (1993) Mediterranean outflow mixing and dynamics. Science 259: 1277--1282. Roberts MJ, Marsh R, New AL, and Wood RA (1996) An intercomparison of a Bryan-Cox-type ocean model and an isopycnic ocean model. Part 1: The subpolar gyre and high-latitude processes. Journal of Physical Oceanography 26(8): 1495--1527. Simpson JE (1997) Gravity Currents: In the Environment and the Laboratory, 2nd edn. Cambridge, UK: Cambridge University Press.


OXYGEN ISOTOPES IN THE OCEAN "

K. K. Turekian, Yale University, New Haven, CT, USA

18

O=16 Osample

18 O=16 O

# 1 1000

standard

Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2066–2067, & 2001, Elsevier Ltd.

The oxygen isotope signature of sea water varies as a function of the processing of water in the oceanic cycle. The two chemical parameters, salinity and oxygen isotope ratio, are distinctive for various water types. The oxygen-18 to oxygen-16 ratio is represented in comparison to a standard. The notation is dO18 which is defined as follows:

Figure 1 shows the results for the world oceans from Craig and Gordon (1965). During the GEOSECS program the oxygen isotope ratios of seawater samples were also determined. The features resemble those in Figure 1. The GEOSECS data are available in the shore-based measurements volume of the GEOSECS Atlas (1987) published by the US National Science Foundation. The tracking of fresh water from streams draining into the ocean at different latitudes has been used to

.60 ng = 0 mixi S 1% /d ic d tlant =_ 2 21N 47W A P . N 20N 49W 25W 37W 27N 23W

1.5 34N 10W

12N 45W

1.0 0° 11W 17N 55W

23S 12W

23N 8W

d/dS = 0.11 E = P = _ 4%

N. Atlantic Deep Water

0

18

O (% )

0.5

Pacific and Indian Deep Water

_0.5

Antarctic Bottom Water

d/dS ≈ 0 (freezing)

_1.0

Equatorial Surface Waters (Lusiad) N. Atlantic Surface Waters (Zephyrus) N. Atlantic Deep Water Weddell Sea (Eltanin Cruise 12)

_1.5

33.0

33.5

34.0

34.5

35.0

35.5 Salinity (% )

36.0

36.5

37.0

37.5

38.0

Figure 1 Oxygen-18–salinity relationships in Atlantic surface and deep waters. dE and dP refer to the isotopic composition of evaporating vapor and precipitation, respectively. (From Craig and Gordon, 1965.)

272


OXYGEN ISOTOPES IN THE OCEAN

study several coastal oceanic regimes. The work of Fairbanks (1982) is one of the earliest of these efforts.

See also Cenozoic Climate – Oxygen Isotope Evidence. River Inputs.

Further Reading Craig H and Gordon LI (1965) Deuterium and oxygen-18 variations in the ocean and marine atmosphere. Stable

273

Isotopes in Oceanographic Studies and Paleotemperatures, Consiglio Nazionale Delle Ricerche, Laboratorio di Geologia Nucleare-Pisa, 122 pp. (Also in Symposium on Marine Chemistry, Publ. 3., Kingston, Graduate School of Oceanography, University of Rhode Island, 277--374). Fairbanks RG (1982) The origin of continental shelf and slope water in the New York Bight and Gulf of Maine: 16 evidence from H18 2 O/H2 O ratio measurements. Journal of Geophysical Research 87: 5796--5808. GEOSECS Atlantic, Pacific, and Indian Ocean Expeditions, Volume 7, Shorebased Data and Graphics (1987) National Science Foundation. 200pp.


OYSTERS – SHELLFISH FARMING I. Laing, Centre for Environment Fisheries and Aquaculture Science, Weymouth, UK & 2009 Elsevier Ltd. All rights reserved.

As I ate the oysters with their strong taste of the sea and their faint metallic taste that the cold white wine washed away, leaving only the sea taste and the succulent texture, and as I drank their cold liquid from each shell and washed it down with the crisp taste of the wine, I lost the empty feeling and began to be happy and to make plans. (Ernest Hemingway in A Moveable Feast).

A Long History Oysters have been prized as a food for millennia. Carbon dating of shell deposits in middens in Australia show that the aborigines took the Sydney rock oyster for consumption in around 6000 BC. Oyster farming, nowadays carried out all over the world, has a very long history, although there are various thoughts as to when it first started. It is generally believed that the ancient Romans and the Greeks were the first to cultivate oysters, but some maintain that artificial oyster beds existed in China long before this. It is said that the ancient Chinese raised oysters in specially constructed ponds. The Romans harvested

immature oysters and transferred them to an environment more favorable to their growth. Greek fishermen would toss broken pottery dishes onto natural oyster beds to encourage the spat to settle. All of these different cultivation methods are still practiced in a similar form today.

A Healthy Food Oysters (Figure 1) have been an important food source since Neolithic times and are one of the most nutritionally well balanced of foods. They are ideal for inclusion in low-cholesterol diets. They are high in omega-3 fatty acids and are an excellent source of vitamins. Four or five medium-size oysters supply the recommended daily allowance of a whole range of minerals.

The Oyster Oysters are one among a number of bivalve mollusks that are cultivated. Others include clams, cockles, mussels, and scallops.

Current Status Oysters form the second most important group, after cyprinid fishes, in world aquaculture production.

Figure 1 A dish of Pacific oysters.

274


OYSTERS – SHELLFISH FARMING

In 2004, 4.6 million tonnes were produced (Food and Agriculture Organization (FAO) data). Asia and the Pacific region produce 93.4% of this total production. The FAO of the United Nations lists 15 categories of cultivated oyster species. The Pacific oyster (Crassostrea gigas) is by far the most important. Annual production of this species now exceeds 4.4 million metric tonnes globally, worth US$2.7 billion, and accounts for 96% of the total oyster production (see Table 1). China is the major producer. The yield from wild oyster fisheries is small compared with that from farming and has declined steadily in recent years. It has fallen from over 300 000 tonnes in 1980 to 152 000 tonnes in 2004. In contrast to this, Pacific oyster production from aquaculture has increased by 51% in the last 10 years (see Figure 2).

Within the shell is a fleshy layer of tissue called the mantle; there is a cavity (the mantle cavity) between the mantle and the body wall proper. The mantle secretes the layers of the shell, including the inner nacreous, or pearly, layer. Oysters respire by using both gills and mantle. The gills, suspended within the mantle cavity, are large and function in food gathering (filter feeding) as well as in respiration. As water passes over the gills, organic particulate material, especially phytoplankton, is strained out and is carried to the mouth.

Table 1 The six most important oyster species produced worldwide

General Biology As might be expected, given a long history of cultivation, there is a considerable amount known about the biology of oysters. The shell of bivalves is in two halves, or valves. Two muscles, called adductors, run between the inner surfaces of the two valves and can contract rapidly to close the shell tightly. When exposed to the air, during the tidal cycle, oysters close tightly to prevent desiccation of the internal tissues. They can respire anaerobically (i.e., without oxygen) when out of water but have to expel toxic metabolites when reimmersed as the tide comes in. They are known to be able to survive for long periods out of water at low temperatures such as those used for storage after collection.

Oyster species

Number of producing countries

Major producing countries (% of total)

Production (metric tonnes)

Pacific oyster

30

4 429 337

1

China (85), Korea (5), Japan (5), France (2.5) USA (95), Canada (4.5) Philippines (100)

1

Australia (100)

5600

17

Spain (50), France (29), Ireland (8) Cuba (99)

5071

American oyster Slipper oyster Sydney oyster European oyster Mangrove oyster

4

3

4 500 000 4 000 000 3 500 000 3 000 000 2 500 000 2 000 000 1 500 000 1 000 000 500 000


2

8

Figure 2 The increase in world production, in metric tonnes, of the Pacific oysters (1950–2004).

20 0

4

19 9

0

19 9

6

19 9

2

19 8

8

19 8

4

19 7

0

19 7

6

19 7

2

19 6

8

19 6

19 5

4

0 19 5

110 770 15 915

1184

Tonnages are 2004 figures (FAO data). All are cupped oyster species, apart from the European oyster, which is a flat oyster species.

5 000 000

19 50

275

276


Oysters have separate sexes, but they may change sex one or more times during their life span, being true hermaphrodites. In most species, the eggs and sperm are shed directly into the water where fertilization occurs. Larvae are thus formed and these swim and drift in the water, feeding on natural phytoplankton. After 2–3 weeks, depending on local environmental conditions, the larvae are mature and they develop a foot. At this stage they sink to the seabed and explore the sediment surface until they find a suitable surface on which to settle and attach permanently, by cementation. Next, they go through a series of morphological and physiological changes, a process known as metamorphosis, to become immature adults. These are called juveniles, spat, or seed. Cultivated species of flat oysters brood the young larvae within the mantle cavity, releasing them when they are almost ready to settle. Fecundity is usually related to age, with older and larger females producing many more larvae. Growth of juveniles is usually quite rapid initially, before slowing down in later years. The length of time that oysters take to reach a marketable size varies considerably, depending on local environmental conditions, particularly temperature and food availability. Pacific oysters may reach a market size of 70–100 g live weight (shell-on) in 18–30 months.

Methods of Cultivation – Seed Supply Oyster farming is dependent on a regular supply of small juvenile animals for growing on to market size.

These can be obtained primarily in one of two ways, either from naturally occurring larvae in the plankton or artificially, in hatcheries. Wild Larvae Collection

Most oyster farmers obtain their seed by collecting wild set larvae. Special collection materials, generically known as ‘cultch’, are placed out when large numbers of larvae appear in the plankton. Monitoring of larval activity is helpful to determine where and when to put out the cultch. It can be difficult to discriminate between different types of bivalve larvae in the plankton to ensure that it is the required species that is present but recently, modern highly sensitive molecular methods have been developed for this. Various materials can be used for collection. Spat collected in this way are often then thinned prior to growing on (Figure 3). In China, coir rope is widely used, as well as straw rope, flax rope, and ropes woven by thin bamboo strips. Shells, broken tiles, bamboo, hardwood sticks, plastics, and even old tires can also be used. Coatings are sometimes applied. Hatcheries

Restricted by natural conditions, the amount of wild spat collected may vary from year to year. Also, where nonnative oyster species are cultivated, there may be no wild larvae available. In these circumstances, hatchery cultivation is necessary. This also allows for genetic manipulation of stocks, to rear and maintain lines specifically adapted for certain

Figure 3 A demonstration of the range of oyster spat collectors used in France.



traits. Important traits for genetic improvement include growth rate, environmental tolerances, disease resistance, and shell shape. Methods of cryopreservation of larvae are being developed and these will contribute to maintaining genetic lines. Furthermore, triploid oysters, with an extra set of chromosomes, can only be produced in hatcheries. These oysters often have the advantage of better growth and condition, and therefore marketability, during the summer months, as gonad development is inhibited. Techniques for hatchery rearing were first developed in the 1950s and today follow well-established procedures. Adult broodstock oysters are obtained from the wild or from held stocks. Depending on the time of year, these may need to be bought into fertile condition by providing a combination of elevated temperature and food (cultivated phytoplankton diets) over several weeks (see Figure 4). Selection of the appropriate broodstock conditioning diet is very important. An advantage of hatchery production is that it allows for early season production in colder climates and this ensures that seeds have a maximum growing period prior to their first overwintering. For cultivated oyster species, mature gametes are usually physically removed from the gonads. This involves sacrificing some ripe adults. Either the gonad can be cut repeatedly with a scalpel and the

277

gametes washed out with filtered seawater into a part-filled container, or a clean Pasteur pipette can be inserted into the gonad and the gametes removed by exerting gentle suction. Broodstock can also be induced to spawn. Various stimuli can be applied. The most successful methods are those that are natural and minimize stress. These include temporary exposure to air and thermal cycling (alternative elevated and lowered water temperatures). Serotonin and other chemical triggers can also be used to initiate spawning but these methods are not generally recommended as eggs liberated using such methods are often less viable. Flat oysters, of the genera Ostrea and Tiostrea, do not need to be stimulated to spawn. They will spawn of their own accord during the conditioning process as they brood larvae within their mantle cavities for varying periods of time depending on species and temperature. The fertilized eggs are then allowed to develop to the fully shelled D-larva veliger stage, so called because of the characteristic ‘D’ shape of the shell valves (Figure 5). These larvae are then maintained in bins with gentle aeration. Static water is generally used and this is exchanged daily or once every 2 days. Throughflow systems, with meshes to prevent loss of larvae, are also employed. Cultured microalgae are added into the tanks several times per day, at an appropriate

Figure 4 The SeaCAPS continuous culture system for algae. An essential element for a successful oyster hatchery is the means to cultivate large quantities of marine micro algae (phytoplankton) food species.


278


Figure 5 Pacific oyster D larvae. At this stage shell length is about 70 microns.

daily ration according to the number and size of the larvae. The larvae eventually become competent to settle and for this, surfaces must be provided. The area of settlement surface is important. Types of materials in common usage to provide large surface areas for settlement include sheets of slightly roughened polyvinyl chloride (PVC), layers of shell chips and particles prepared by grinding aged, clean oyster shell spread over the base of settlement trays or tanks, bundles, bags, or strings of aged clean oyster shells dispersed throughout the water column, usually in settlement tanks or various plastic or ceramic materials coated with cement (lime/mortar mix). For some of these methods, the oysters are subsequently removed from the settlement surface to produce ‘cultchless’ spat. These can then be grown as separate individuals through to marketable size and sold as whole or half shell, usually live. Provision of competence to settle Pacific oyster larvae for remote setting at oyster farms is common practice on the Pacific coast of North America. Hatcheries provide the mature larvae and the farmers themselves set them and grow the spat for seeding oyster beds or in suspended culture. In other parts of the world, hatcheries set the larvae, as described above, and grow the spat to a size that growers are able to handle and grow. Oyster juveniles from hatchery-reared larvae perform well in standard pumped upwelling systems and survival is usually good, although some early losses may occur immediately following metamorphosis. Diet, ration, stocking density, and water flow rate are all important in these systems (Figure 6). They are only suitable for initial rearing of small seed. As the spat

grow, food is increasingly likely to become limiting in these systems and they must be transferred to the sea for on-growing. The size at which spat is supplied is largely dictated by the requirements and maturity of the growout industry. Seed native oysters are made available from commercial hatcheries at a range of sizes up to 25–30 mm. The larger the seed, the more expensive they are but this is offset by the higher survival rate of larger seed. Larger seed should also be more tolerant to handling. Ponds

Pond culture offers a third method to provide seed. This was the method that was originally developed in Europe in early attempts to stimulate production following the decline of native oyster stocks in the late nineteenth century. Ponds of 1–10 ha in area and 1–3 m deep were built near to high water spring tides, filled with seawater, and then isolated for the period of time during which the oysters are breeding naturally, usually May to July. Collectors put into the ponds encourage and collect the settlement of juvenile oysters. There is an inherent limited amount of control over the process and success is very variable. In France, where spat collectors were also deployed in the natural environment, it was relatively successful and became an established method for a time. In the UK, spat production from ponds built in the early 1900s was insufficiently regular to provide a reliable supply of seed to the industry and the method was largely abandoned in the middle of the twentieth century in favor of the more controlled conditions available in hatcheries.



279

Figure 6 Indoor (a) and outdoor (b) oyster nursery system, in which seawater and algae food are pumped up through cylinders fitted with mesh bases and containing the seed.

Methods of Cultivation – On-growing Various methods are available for on-growing oyster seed once this has been obtained, from either wild set larvae, hatcheries, or ponds. Oysters smaller than 10 g need to be held in trays or bags attached to a superstructure on the foreshore until they are large enough to be put directly on to

the substrate and be safe from predators, strong tidal and wave action, or siltation (Figure 7). Tray cultivation of oysters can be successful in water of minimal flow, where water exchange is driven only by the rise and fall of the tide and gentle wave action. In such circumstances it may even be possible in open baskets.


280


Figure 7 The traditional bag and trestle method for on-growing Pacific oysters.

Figure 8 Open baskets for on-growing, as developed in Tasmania for Pacific oysters.

In more exposed sites, systems developed in Australia and employing plastic mesh baskets attached to or suspended from wires (see Figure 8) are becoming increasingly popular. It is claimed that an oyster with a better shape, free from worm (Polydora) infestation, will result from using these systems. Polydora can cause unsightly brown blemishes on the inner surface of the shell and decrease marketability of the stock. Less labor is required with these systems but they are more expensive to purchase initially.

In all of the above systems, the mesh size can be increased as the oysters grow, to improve the flow of water and food through the animals. Holding seed oysters in trays or attached to the cultch material onto which they were settled, and suspended from rafts, pontoons, or long lines is an alternative method in locations where current speed will allow. The oysters should grow more quickly because they are permanently submerged but the shell may be thinner and therefore more susceptible



to damage. Early stages of predator species such as crabs and starfish can settle inside containers, where they can cause significant damage unless containers are opened and checked on a regular basis. In the longer term the oysters will perform better on the seabed, although they can be reared to market size in the containers. Protective fences can be put up around ground plots to give some degree of protection to smaller oysters from shore crabs. Potting crabs in the area of the lays is another method of control. The steps in the oyster cultivation process are shown as a diagram in Figure 9. The correct stocking density is important so as not to exceed the carrying capacity of a body of water. When carrying capacity is exceeded, the algal population in the water is insufficient and growth declines. Mortalities also sometimes occur.

Yields in extensively cultivated areas are usually about 25 tonnes per hectare per year but this can increase to 70 tonnes where individual plots are well separated. In France, premium quality oysters are sometimes finished by holding them in fattening ponds known as claires (see Figure 10), in which a certain type of algae is encouraged to bloom, giving a distinctive green color to the flesh. In the UK, part-grown native flat oysters from a wild fishery are relayed in spring into areas in which conditions are favorable to give an increase in meat yield over a period of a few months, prior to marketing in the winter. There are proposals to establish standards for organic certification of mollusk cultivation but many oyster growers consider that the process is intrinsically organic.

Adult oysters in the sea Immature

Mature Spawn in wild

Collect spat on settlement material (cultch)

Remove from cultch

Brood stock conditioning Larvae Larvae

Ponds

Spat

Hatchery Spat

Adult oysters for the market

On cultch

In trays

Spat Spat On-growing in the sea (spat to adults)

Nursery

281

On-bottom

Adult oysters for the market

Figure 9 The steps and processes of oyster farming.


282


Figure 10 Oyster fattening ponds, or Claires, in France.

Pearl Oysters

Table 2

There is an important component of the oyster farming industry, located mainly in Japan, Australia, and the South Sea Islands, devoted to the production of pearls. The process involves inserting into the oyster a nucleus, made with shell taken from a North American mussel, and a tiny piece of mantle cut from another oyster. The oysters are cultivated in carefully tended suspended systems while pearls develop around the nucleus. Once the pearls have been taken out of the oysters, they are seeded anew. A healthy oyster can be reseeded as many as 4 times with a new nucleus. As the oyster grows, it can accommodate progressively bigger pearl nuclei. Therefore, the biggest pearls are most likely to come from the oldest oysters. There has been a significant increase in demand for pearl jewelry in the last 10 years. The US is the biggest market, with estimated sales of US$1.5 billion of pearl jewelry per year (Table 2).

Site Selection for On-growing A range of physical, biological, and chemical factors will influence survival and growth of oysters. Many of these factors are subject to seasonal and annual variation and prospective oyster farmers are usually advised to monitor the conditions and to see how well oysters grow and survive at their chosen site for at least a year before any commercial culture begins. Seawater temperature has a major effect on seasonal growth and may be largely responsible for any differences in growth between sites. Changes in

Estimated values of pearl production (2004) Market share Production in value (%) value (2004)

White South Sea cultured pearls (Australia, Indonesia, the Philippines, Myanmar) Freshwater cultured pearls (China) Akoya cultured pearls (Japan, China) Tahitian cultured pearls (French Polynesia) Total

35

US$220 million

24

US$150 million US$135 million US$120 million

22 19

US$625 million

Source: Pearl World, The International Pearling Journal (January/February/March 2006).

salinity do not affect the growth of bivalves by as much as variation in temperature. However, most bivalves will usually only feed at higher salinities. Pacific oysters prefer salinity levels nearer to 25 psu, conditions typical of many estuaries and inshore waters. Growth rates are strongly influenced by the length of time during which the animals are covered by the tide. Growth of oysters in trays stops when they are exposed to air for more than about 35% of the time. However, this fact can be used to advantage by the cultivator who may wish, for commercial reasons, to slow down or temporarily stop the growth of the stock. This can be achieved by moving the stock higher up the beach. It is a practice that is routinely



adopted in Korea and Japan for ‘hardening off’ wildcaught spat prior to sale. Other considerations are related to access and harvesting. It is important to consider the type of equipment likely to be used for planting, maintenance, and harvesting, particularly at intertidal sites. Some beaches will support wheeled or tracked vehicles, while others are too soft and will require the use of a boat to transport equipment.

Risks Oyster farming can be at risk from pollutants, predators, competitors, diseases, fouling organisms, and toxic algae species. Pollutants

Some pollutants can be harmful to oysters at very low concentrations. During the 1980s, it was found that tributyl tin (TBT), a component of marine antifouling paints, was highly toxic to bivalve mollusks at extremely low concentrations in the seawater. Pacific oysters cultivated in areas in which large numbers of small vessels were moored showed stunted growth and thickening of the shell, and natural populations of flat oysters failed to breed (see Figure 11). The use of this compound on small vessels was widely banned in the late 1980s, and since then the oyster industry has recovered. The International Maritime Organization has since announced a ban for larger vessels as well. When marketed for consumption, oysters must meet a

283

number of ‘end product’ standards. These include a requirement that the shellfish should not contain toxic or objectionable compounds such as trace metals, organochlorine compounds, hydrocarbons, and polycyclic aromatic hydrocarbons (PAHs) in such quantities that the calculated dietary intake exceeds the permissible daily amount. Predators

In some areas oyster drills or tingles, which are marine snails that eat bivalves by rasping a hole through the shell to gain access to the flesh, are a major problem. Competitors

Competitors include organisms such as slipper limpets (Crepidula fornicata), which compete for food and space. In silt laden waters, they also produce a muddy substrate, which is unsuitable for cultivation. They can be a significant problem. An example is around the coast of Brittany, where slipper limpets have proliferated in the native flat oyster areas during the last 30 years. In order to try to control this problem, approximately 40 000–50 000 tonnes of slipper limpets are harvested per year. These are taken to a factory where they are converted to a calcareous fertilizer. Diseases

The World Organisation for Animal Health (OIE) lists seven notifiable diseases of mollusks and six of

Figure 11 Compared with a normal shell (upper shell section), TBT from marine antifouling paints causes considerable thickening of Pacific oyster shells (lower section). These compounds are now banned.


284


these can infect at least one of the cultivated oyster species. Some of these diseases, often spread through national and international trade, have had a devastating effect on oyster stocks worldwide. For example, Bonamia, a disease caused by the protist parasitic organism Bonamia ostreae, has had a significant negative impact on Ostrea edulis production throughout its distribution range in Europe. Mortality rates in excess of 80% have been noted. The effect this can have on yields can be seen in the drastic (93%) drop in recorded production in France, from 20 000 tonnes per year in the early 1970s to 1400 tonnes in 1982. A major factor in the introduction and success of the Pacific oyster as a cultivated species has been that it is resistant to major diseases, although it is not completely immune to all problems. There are reports of summer mortality episodes, especially in Europe and the USA, and infections by a herpes-like virus have been implicated in some of these. Vibrio bacteria are also associated with larval mortality in hatcheries. The Fisheries and Oceans Canada website lists 53 diseases and pathogens of oysters. Juvenile oyster disease is a significant problem in cultivated Crassostrea virginica in the Northeastern United States. It is thought to be caused by a bacterium (Roseovarius crassostreae). Fouling

Typical fouling organisms include various seaweeds, sea squirts, tubeworms, and barnacles. The type and degree of fouling varies with locality. The main effect is to reduce the flow of water and therefore the supply of food to bivalves cultivated in trays or on ropes and to increase the weight and drag on floating installations. Fouling organisms grow in response to the same environmental factors as are desirable for good growth and survival of the cultivated stock, so this is a problem that must be controlled rather than avoided. Toxic Algae

Mortalities of some marine invertebrates, including bivalves, have been associated with blooms of some alga species, including Gyrodinium aureolum and Karenia mikimoto. These so-called red tides cause seawater discoloration and mortalities of marine organisms. They have no impact on human health.

Stock Enhancement If we accept the FAO definition of aquaculture as ‘‘The farming of aquatic organisms in inland and

coastal areas, involving intervention in the rearing process to enhance production and the individual or corporate ownership of the stock being cultivated,’’ then in many cases stock enhancement can be included as a type of oyster farming. Natural beds can be managed to encourage the settlement of juvenile oysters and sustain a fishery. Beds can be raked and tilled on a regular basis to remove silt and ensure that suitable substrates are available for the attachment of the juvenile stages. Adding settlement material (cultch) is also beneficial. In some areas of the world, there has been a dramatic reduction in stocks of the native oyster species. This is attributed mainly to overexploitation, although disease is also implicated. Native oyster beds form a biotope, with many associated epifaunal and infaunal species. Loss of this habitat has resulted in a major decline in species richness in the coastal environment. Considerable effort has been put into restoring these beds, using techniques that might fall under the definition of oyster farming. A good example is the Chesapeake Bay Program for the native American oyster Crassostrea virginica. Overfishing followed by the introduction of two protozoan diseases, believed to be inadvertently introduced to the Chesapeake through the importation of a nonnative oyster, Crassostrea gigas, in the 1930s, has combined to reduce oyster populations throughout Chesapeake Bay to about 1% of historical levels. The Chesapeake Bay Program is committed to the restoration and creation of aquatic reefs. A key component of the strategy to restore oysters is to designate sanctuaries, that is, areas where shellfish cannot be harvested. It is often necessary within a sanctuary to rehabilitate the bottom to make it a suitable oyster habitat. Within sanctuaries aquatic reefs are created primarily with oyster shell, the preferred substrate of spat. Alternative materials including concrete and porcelain are also used. These permanent sanctuaries will allow for the development and protection of large oysters and therefore more fecund and potentially diseaseresistant oysters. Furthermore, attempts have been made at seeding with disease-free hatchery oysters.

Nonnative Species The International Council for the Exploration of the Sea (ICES) Code of Practice sets forth recommended procedures and practices to diminish the risks of detrimental effects from the intentional introduction and transfer of marine organisms. The Pacific oyster, originally from Asia, has been introduced around the world and has become invasive in parts of Australia and New Zealand, displacing the native rock oyster



in some areas. It is also increasingly becoming a problem in northern European coastal regions. In parts of France naturally recruited stock is competing with cultivated stock and has to be controlled. The possibility of farming the nonnative Suminoe oyster (Crassostrea ariakensis) to restore oyster stocks in Chesapeake Bay is being examined and there are differing opinions about the environmental risks involved, with concerns of potential harmful effects on the local ecology. It should also be noted that aquaculture has been responsible for the introduction of a whole range of passenger species, including pest species such as the slipper limpet and oyster drills, throughout the world.

Food Safety Bivalve mollusks filter phytoplankton from the seawater during feeding; they also take in other small particles, such as organic detritus, bacteria, and viruses. Some of these bacteria and viruses, especially those originating from sewage outfalls, can cause serious illnesses in human consumers if they remain in the bivalve when it is eaten. The stock must be purified of any fecal bacterial content in cleansing

285

(depuration) tanks (Figure 12) before sale for consumption. There are regulations governing this, which are based on the level of contamination of the mollusks. Viruses are not all removed by normal depuration processes and they can cause illness if the bivalves are eaten raw or only lightly cooked, as oysters often are. These viruses can only be detected by using sophisticated equipment and techniques, although research is being carried out to develop simpler methods. Finally, certain types of naturally occurring algae produce toxins, which can accumulate in the flesh of oysters. People eating shellfish containing these toxins can become ill and in exceptional cases death can result. Cooking does not denature the toxins responsible nor does cleansing the shellfish in depuration tanks eliminate them. The risks to consumers are usually minimized by a requirement for samples to be tested regularly. If the amount of toxin exceeds a certain threshold, the marketing of shellfish for consumption is prohibited until the amount falls to a safe level.

See also Crustacean Fisheries. Molluskan Fisheries.

Further Reading

Figure 12 A stacked depuration system, suitable for cleansing oysters of microbiological contaminants. The oysters are held in trays in a cascade of UV-sterlized water.

Andersen RA (ed.) (2005) Algal Culturing Techniques. New York: Academic Press. Bueno P, Lovatelli A, and Shetty HPC (1991) Pearl oyster farming and pearl culture, Regional Seafarming Development and Demonstration Project. FAO Project Report No. 8. Rome: FAO. Dore I (1991) Shellfish – A Guide to Oysters, Mussels, Scallops, Clams and Similar Products for the Commercial User. London: Springer. Gosling E (2003) Bivalve Mollusks; Biology, Ecology and Culture. London: Blackwell. Helm MM, Bourne N, and Lovatelli A (2004) Hatchery culture of bivalves; a practical manual. FAO Fisheries Technical Paper 471. Rome: FAO. Huguenin JE and Colt J (eds.) (2002) Design and Operating Guide for Aquaculture Seawater Systems, 2nd edn. Amsterdam: Elsevier. Lavens P and Sorgeloos P (1996) Manual on the production and use of live food for aquaculture. FAO Fisheries Technical Paper 361. Rome: FAO. Mann R (1979) Exotic species in mariculture. Proceedings of Symposium on Exotic Species in Mariculture: Case Histories of the Japanese Oyster, Crassostrea gigas (Thunberg), with implications for other fisheries. Woods Hole Oceanographic Institution, Cambridge Press. Matthiessen G (2001) Fishing News Books Series: Oyster Culture. New York: Blackwell.


286


Spencer BE (2002) Molluscan Shellfish Farming. New York: Blackwell.

Relevant Websites http://www.crlcefas.org – Cefas is the EU community reference laboratory for bacteriological and viral contamination of oysters. http://www.oie.int – Definitive information on oyster diseases can be found on the website of The World Organisation for Animal Health (OIE). http://www.pac.dfo-mpo.gc.ca – Further information on diseases can be found on the Fisheries and Oceans Canada website. http://www.cefas.co.uk – The online magazine ‘Shellfish News’, although produced primarily for the UK industry, has articles of general interest on oyster farming. It can be found on the News/Newsletters area of the Cefas web site. A booklet on site selection for bivalve cultivation is also available at this site.

http://www.vims.edu – The Virginia Institute of Marine Science has information on oyster genetics and breeding programs and the proposals to introduce the non-native Crassostrea ariakensis to Chesapeake Bay. http://www.was.org – The World Aquaculture Society website has a comprehensive set of links to other relevant web sites, including many national shellfish associations. http://www.fao.org – There is a great deal of information on oyster farming throughout the world on the website of The Food and Agriculture Organization of the United Nations (FAO). This website has statistics on oyster production, datasheets for various cultivated species, including Pacific oysters, and online manuals on the production and use of live food for aquaculture (FAO Fisheries Technical Paper 361), Hatchery culture of bivalves (FAO Fisheries Technical Paper. No. 471) and Pearl oyster farming and pearl culture (FAO Project Report No. 8).


PACIFIC OCEAN EQUATORIAL CURRENTS adjustment of the currents to changing forcing is associated with a special class of internal wave motions termed linear equatorially trapped waves. There are several different types of waves with rich meridional and vertical structure, governed by dispersion relationships that tie zonal wavelength, meridional structure, and wave period together. The fastest waves cross the Pacific in only 2–3 months. With greater distance from the Equator, zonal propagation speeds become slower. The key feature is that these waves can transmit the signals of wind forcing to and from remote locations. The Pacific equatorial surface currents are primarily wind-driven. Local forcing of the equatorial currents is dominated by surface wind stress (as opposed to heat and/or fresh water fluxes) and its variability. Variable wind forcing results in vertical pumping of the thermocline and subsequent dynamic adjustment, including radiation of equatorially trapped waves. The currents are not forced solely by local winds, however, because equatorially trapped waves carry wind-forcing signals across the entire basin, and similar boundary-trapped waves transmit information about forcing between the equator and higher latitudes. An important portion of the

R. Lukas, University of Hawaii at Manoa, Hawaii, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2069–2076, & 2001, Elsevier Ltd.

Introduction An essential characteristic of Pacific equatorial ocean currents is that they span the width of the Pacific basin (15 000 km at the Equator), linking to eastern and western boundary flows (Figures 1 and 2). While they have long zonal scales, the relatively strong near-equatorial flows have complex vertical and meridional structures. They exhibit energetic variability on timescales from days to years. In particular, the currents of the equatorial Pacific are considerably altered during El Ninõ/Southern Oscillation (ENSO) events. These flows are subject to the distinctive physics associated with the equatorward decrease of the vertical component of the earth’s rotation vector. The associated vanishing of the horizontal Coriolis force at the Equator results in relatively strong currents for a given wind stress or pressure gradient. Rapid

140˚E

Kuroshio MUC 14 41 27 MC Mindanao

160˚E

17

20

14 ah

lm a er

Banda Sea

a Se

NGCC New NGC Guin UC ea

tral

SEC 35

on

34

NSCC 15 EUC

a

Transports _ 106 m3 s 1

SSCC

5

10

10

10

Per u

SEC

Equator South America

cur

ren

t

SEC

UC

14

20˚S

~ ~

140˚E

160˚E

20˚N

PC

LC

Surface Subsurface

eric

10

EAC

Australia

Am

5

Se

a

80˚W

10

NECC

SEC

So

lom

Torres Strait

GBRUC

20˚S 120˚E

15

NECC

12 SEC NSCC 15 EUC 8 SSCC 4 SECC

11

Ha

Equator

Cen

NEC

Celebes Sea

HE

100˚W

12 California current

27

NEC

ME

120˚W

~ ~

120˚E 20˚N

120˚W

100˚W

80˚W

Figure 1 Schematic illustration of the major equatorial currents in the Pacific Ocean and their connections to eastern and western boundary currents. (Note the break in the longitude axis.) Surface currents are indicated by solid lines; subsurface currents are indicated by dashed lines, with deeper currents having lighter weight. Approximate average individual current transports (106 m3 s1) are provided where known with some confidence. The equatorial surface currents are the North Equatorial Current (NEC), the South Equatorial Current (SEC), the North Equatorial Countercurrent (NECC), and the South Equatorial Countercurrent (SECC). The subsurface equatorial currents discussed here are the Equatorial Undercurrent (EUC), the Northern Subsurface Countercurrent (NSCC), and the Southern Subsurface Countercurrent (SSCC). Eastern boundary currents are the Peru Current and the Peru–Chile Undercurrent (PCUC). Western boundary surface currents are the Kuroshio, the Mindanao Current (MC), the New Guinea Coastal Current (NGCC) and the East Australia Current (EAC). Subsurface flows along the western boundary are the Mindanao Undercurrent (MUC), the New Guinea Coastal Undercurrent (NGCUC), and the Great Barrier Reef Undercurrent (GBRUC). The Mindanao Eddy (ME) and Halmahera Eddy (HE) are indicated.


287

288

PACIFIC OCEAN EQUATORIAL CURRENTS

Annual mean

10 m currents

20˚N NEC

15˚N 10˚N NECC

5˚N

SEC

EQ 5˚S

SECC

10˚S 15˚S 20˚S 140˚E

160˚E

160˚W

180˚

0

0.1

0.2

0.3

140˚W

0.4

120˚W

0.5

ms

100˚W

_1

Figure 2 Map of long-term mean surface flow in the tropical Pacific. White lines and arrows indicate the direction of flow, while the colors indicate the speed of flow as given by the color bar. Current names are abbreviated as in Figure 1.

equatorial circulation is forced by winds and buoyancy forces (surface heat and fresh water fluxes) far from the equator. The basic spatial structures and temporal variation of Pacific Ocean equatorial currents are presented here. Because systematic current measurements are available at only a few, widely separated locations, spatial variability is addressed primarily with the ocean assimilation/reanalysis from the US National Oceanic and Atmospheric Administration (NOAA) National Centers for Environmental Prediction, which combines a general circulation model of the ocean with atmospheric and ocean data to estimate the state of the tropical Pacific Ocean every week. Direct current measurements from several sites along the Equator maintained by the NOAA Pacific Marine Environmental Laboratory are used primarily to address temporal variability.

Mean Flow Interior Flows

Zonal geostrophic flow, where meridional pressure gradients are balanced by Coriolis forces (due to flow on the rotating earth), dominates over meridional flow in the long-term annual-average Pacific equatorial circulation. In the surface layer (upper 50 m or so), currents directly driven by the generally westward Trade Winds typically flow poleward, superimposed on these strong zonal flows (Figure 2). The divergence of poleward-flowing surface currents is most pronounced along the Equator, leading to depth-dependent pressure gradients and strong vertical flow (called equatorial upwelling).

Surface currents The time-averaged surface flows (Figure 2) are dominated by the westward South Equatorial Current (SEC; between about 31N and 201S) and the North Equatorial Current (NEC; between about 101N and 201N). A persistent North Equatorial Countercurrent (NECC) flows eastward across the basin in the narrow band between about 51N and 101N. A weaker eastward-flowing South Equatorial Countercurrent (SECC) extends eastward from the region of the western boundary, but this flow only intermittently reaches the central and eastern Pacific. North–south profiles of surface currents at three different longitudes across the Pacific basin clearly show the structure of these major zonal flows (Figure 3). The NEC, NECC, and SEC are strongest in the central Pacific where the Trade Winds are strongest. The SEC and NECC are considerably weaker in the west than in the east, reflecting the greater variability of the winds in the western Pacific. Very near the Equator in the central and eastern Pacific, there is a minimum in the speed of the SEC, and the relatively narrow filament of SEC north of the equator is stronger than the flow south of the Equator, except in the west. The SECC occurs between about 31S and 101S, but generally dissipates west of the dateline. The meridional flows are considerably weaker than the zonal flow (Figure 3). The average currents have a northward component north of the Equator, and southward to the south of the Equator. The transition occurs rapidly very close to the Equator at all three longitudes, due to east-to-west trade winds and the change in sign of the Coriolis force at the



V

U

0.3 165˚E 140˚W 110˚W

0.2

289

0.18 0.15

165˚E 140˚W 110˚W

0.12 Current speed (m s 1)

0.09

_

_

Current speed (m s 1)

0.1 0 _ 0.1 _ 0.2

0.06 0.03 0 _ 0.03 _ 0.06

_ 0.3

_ 0.09

_ 0.4

_ 0.12 20˚S

10˚S

0˚

10˚N 20˚N

(A)

20˚S

10˚S

0˚

10˚N 20˚N

(B)

Figure 3 North–south profiles of zonal (A) and meridional (B) current in the western (1651E), central (1401W) and eastern (1101W) Pacific. Positive current is eastward and northward. Note the difference in the velocity scales between the two panels.

Equator. This divergence causes equatorial upwelling, which is responsible for colder surface temperatures along the Equator than just to the north or south. Subsurface structure The zonal flow of the NECC is unusual among the surface currents in that it has a subsurface maximum near 50 m (Figure 4). This is due to its flowing eastward against the prevailing Trade winds. The directly wind-driven flow vanishes below the mixed layer (usually shallower than 100 m), and currents below are zonallyoriented except near the eastern and western boundaries. The time-averaged subsurface flows are dominated by the eastward Equatorial Undercurrent (EUC; Figures 4–6), which is the strongest equatorial Pacific current, reaching speeds of about 1 m s1 between 1201W and 1401W (Figures 5 and 6). The EUC is found within the very strong equatorial thermocline just below the westward SEC (Figures 4 and 5). The strong mean shear above the core of the EUC gives rise to strong vertical mixing. Note also the strong meridional shear of the zonal flow on either side of the EUC, and between the SEC and NECC. Much weaker Northern and Southern Subsurface Countercurrents (NSCC and SSCC) flow eastward below the poleward flanks of the EUC (Figure 4). The westward Equatorial Intermediate Current (EIC) is found directly below the EUC across the Pacific. Both the EUC and EIC slope upward toward the east

(Figure 5), tending to follow shoaling isopycnal surfaces. On the Equator, alternating deep equatorial jets (not shown) are found below the EIC. The poleward wind-driven meridional flows are mostly confined to the upper 50 m (Figure 4). The central Pacific section (Figure 4E) shows meridional flow nearly symmetric with respect to the equator, with divergent poleward flow in the near-surface, and convergent equatorward flow near the core of the EUC. This classical picture is not seen in the eastern and western sections, due to a cross-equatorial component of the surface winds, especially in the east. Boundary Flows

Surface The westward flow of the NEC impinges on the Philippines where it splits near 141N into a northward-flowing Kuroshio Current and southward Mindanao Current (MC) along the western boundary (Figures 1 and 2). The MC has surface speeds exceeding 1 m s1, and the flow reaches to depths of 300–600 m. The upper 100 m flow splits at the south end of Mindanao Island, with a significant portion entering the Sulawesi Sea, and most of the rest retroflecting into the NECC. A small portion recirculates around the persistent Mindanao Eddy. A portion of the flow in the Sulawesi Sea transits through the Makassar Strait and ultimately through the Indonesian Seas and into the Indian Ocean, while the rest returns to the


290


Mean V at 165˚E

Mean U at 165˚E

0 Depth (m)

100 200 300 400 500 (A) Mean V at 140˚E

Mean U at 140˚E

0 Depth (m)

100 200 300 400 500 (B) Mean U at 110˚E

0

Mean V at 110˚E

Depth (m)

100 200 300 400 500 20˚S (C) _ 20 _ 10

10˚S

0

Eq

10

20

30

10˚N

40

50

60

70

20˚N _ 80 cm s 1

20˚S (F)

10˚S

_ 10 _ 8 _ 6 _ 4 _ 2 0

Eq

2

4

10˚N

6

8

20˚N

_ 10 12 14 16 18 cm s 1

Figure 4 Vertical section showing the zonal (A–C) and meridional (D–F) components of mean flow in the upper 500 m of the western (A, D), central (B, E) and eastern (C, F) Pacific Ocean. Color bars give magnitude of flow (note that zonal and meridional scales are not the same); positive values are eastward and northward. Zero values are indicated by white contours.

Pacific to join the NECC (Figure 6). Deeper portions of the MC are also split between the Indonesian Throughflow and the Pacific equatorial circulation, with the latter flowing into the EUC and NSCC. The SEC impinges on the north-eastern coast of Australia and the complex of islands including New Guinea. Similarly to the NEC, it splits into poleward and equatorward boundary flows near 141S (Figure 1). The poleward branch is the East Australia Current, and the northward branch flows under a shallow southward surface flow as the Great Barrier Reef Undercurrent, eventually becoming the New Guinea Coastal Current (NGCC) and New Guinea Coastal Undercurrent (NGCUC), which follow a convoluted path around topographic features ending up with westward flow along the north coast of New Guinea. In this region, the surface flow of the NGCC reverses seasonally with the Asian winter monsoon westerly winds.

Low-latitude boundary currents in the eastern Pacific are not nearly as strong as along the western boundary, but they play a significant role in closing the circulation of the Pacific Ocean (Figure 1). The Peru Current flows northward along the west coast of South America, ultimately turning offshore into the SEC. North of the Equator, the NECC flows into the Gulf of Panama and retroflects around the Costa Rica Dome into the NEC, joining southward flow from the California Current. Subsurface Along the western Pacific boundary (Figure 1), the NGCUC flows westward along the north coast of New Guinea with a maximum near 200 m, but extending to at least 800 m. The upper thermocline waters contribute a small fraction to the Indonesian Throughflow, with the rest retroflecting around the persistent Halmahera Eddy to form (with contributions from the MC) the



flowing eastward in the NSCC and SSCC is not well known.

Mean U at 0˚

0

291

SEC EUC

200 Depth (m)

Variability 400

The variability of equatorial currents is complex, spanning a broad range of time and space scales. This variability is largely forced by changing winds; an important fraction of these wind changes are due to sea surface temperature changes in the equatorial zone, these being associated with changes in the currents and winds. These coupled variations include the well-known El Ninõ phenomenon, but also include the annual cycle.

EIC

600

800

1000 120˚E _ 20 _ 10

160˚E 0

10

160˚W 20

30

40

80˚W

120˚W 50

60

70

_ 80 90 cm s 1

Annual Cycle Figure 5 Zonal flow speed in the upper 1000 m in a vertical section along the Equator. Speeds are given in the color bar, with positive values eastward. The vertical lines indicate the longitudes of long-term current meter measurements (see Figure 7).

Although the annual cycle of wind forcing near the equator is not as extreme as in mid latitudes, the Pacific Trade Winds vary enough to force significant changes to the currents discussed above. Because of equatorial wave dynamics and coupling with the atmosphere, the relationship of the current variability to the winds is quite complex. Figure 7 shows the long-term mean plus annual cycle of zonal and meridional current in the nearsurface layer and at depth within the EUC for locations in the western, central and eastern Pacific. (Here, the annual cycle is the sum of annual and semiannual harmonics analyzed from direct current measurements.) The range of zonal current variation is about one order of magnitude larger than the meridional variations. The annual harmonic dominates the annual cycle of surface flow, except in the western Pacific where monsoon winds cross the

eastward-flowing EUC (Figure 6). The deeper portions of the NGCUC flow across the Equator and are traced into the weak Mindanao Undercurrent (MUC) which flows northward below the MC, carrying Antarctic Intermediate Water into the North Pacific. In the east, the Peru–Chile Undercurrent flows poleward along the west coast of Ecuador, Peru and Chile, basically an extension of the EUC past the Galapagos Islands that then turns southward after converging at the coast of Ecuador. Some of these waters join the westward flows of the Peru Current and SEC through upwelling. The fate of waters

20˚N NEC

15˚N 10˚N

NECC

5˚N EUC

0 5˚S

SEC

10˚S 15˚S 20˚S 140˚E

160˚E

180˚ 0

0.2

160˚W 0.4

0.6

140˚W 0.8

120˚W

100˚W

_1

1.0 m s

Figure 6 Map of long-term mean flow in the tropical Pacific on the isopycnal surface sY ¼ 24:5 kgm3 , which lies within the highspeed core of the Equatorial Undercurrent. White lines and arrows indicate the direction of flow, while the colors indicate the speed of flow as given by the color bar. Current names are abbreviated as in Figure 1.


292


Equatorial surface zonal current 5

40 30

165˚E 140˚W 110 ˚W

Equatorial surface meridional current 165 ˚E 140˚ W 110 ˚W

4 3

20

2 1

_

Speed (cm s 1)

10 0 0 _ 10

_1

_ 20

_2

_ 30

_3

_ 40 _ 50

_4 _5 J F M A M J J A S O N D J F M A M J

(A)

J F M A M J J A S O N D J F M A M J (B)

Equatorial thermocline zonal current 110

Equatorial thermocline meridional current 10 165˚E 200 m 140˚W 120 m 110˚W 120 m

100 8

_

Speed (cm s 1)

90

80

165˚E 200 m 140˚W 120 m 110˚ W 120 m

6

4

70

60 2 50 0 40 _2

30 J F M A M J J A S O N D J F M A M J

J F M A M J J A S O N D J F M A M J (D)

(C)

Figure 7 Mean annual variation of zonal (A, C) and meridional (B, D) currents at 10 m depth (A, B) and in the thermocline (C, D) on the Equator at three locations across the Pacific Ocean indicated in Figure 5. Because the thermocline is deeper in the western Pacific, currents are presented for a corresponding depth there. Note the different speed scales for each panel.

equator twice each year. Also, zonal current in the eastern equatorial thermocline and meridional current in the central equatorial thermocline show strong semiannual signals.

Strong annual variation of zonal surface current is sufficient to reverse the direction of the flow on the equator, especially during the Northern Hemisphere spring (Figure 7A). In the east, this feature occurs



nearly every year, and has been erroneously described as a ‘surfacing’ of the EUC. In the central Pacific, such reversals are mainly observed during strong El Ninõ events, and its appearance in the annual cycle here may be due to the occurrence of several El Ninõ events during the record that was analyzed (1984–1998). It is noteworthy that the maximum eastward deviation of the annual cycle appears progressively later toward the west, thought to be coupled with westward propagation of the annual cycle of zonal wind and sea surface temperature. The annual cycle in the strong eastward flow of the EUC within the thermocline (Figure 7C) is not large enough to reverse the current direction. (On interannual timescales, however, the flow may reverse— see below.) El Ninõ/Southern Oscillation (ENSO)

The strongest variability of the zonal equatorial currents is associated with El Ninõ episodes that

293

occurred in 1986–87, 1990–91, 1993, and 1997–98, and with La Ninã episodes of 1984, 1988, and 1996. Current meter records that have had their mean and annual cycles removed are presented in Figure 8, showing that El Ninõ variations are large enough to reverse the westward flow of the SEC, especially in the western equatorial Pacific. Strong eastward surface flow in the warm water pool of the western equatorial Pacific (e.g., 1997) has been implicated in the warming of the sea surface in the central and eastern equatorial Pacific. Also, the eastward flow of the EUC is reversed during some of these events (e.g., in Figure 8B at 1401W during 1997). This disappearance of the EUC was first observed during the strong 1982–83 El Ninõ event. The current systems off the Equator are also affected by ENSO, again through a combination of local wind forcing and remotely-forced baroclinic waves. The NECC strengthens in the early phases of El Ninõ. The off-equatorial portion of the SEC weakens during El Ninõ, and strengthens during La Ninõ. The interannual variations of the NEC show a

200

_

U Speed (cm s 1)

100

165˚E _ 140˚W, _100 cm s 1 _ 110˚W, _ 200 cm s 1

0 100 200 300 400 1982

1984

1986

1988

1990

1992

1994

1996

1998

1986

1988

1990

1992

1994

1996

1998

(A)

50

_

U Speed (cm s 1)

0

140˚ _ 110˚W, _ 200 cm s 1

50 100 150 200 250 300

1982

1984

(B) Figure 8 Time series of observed zonal currents, with mean and annual cycle removed, at three locations (indicated in Figure 5) along the Equator in the surface layer (A) and in the thermocline (B). Warm El Ninõ events are indicated by red shading along the time axis; blue shading indicates cold La Ninõ events. Missing observations are indicated by gaps. Note that time series have been offset from each other for clarity.


294


50

˚ ˚ ˚

165 E 140 W _ 100 cm s 1 _ 110 W _ 200 cm s 1

0

_

V (cm s 1)

50

_

100 150 200 250 300

1982

1984

1986

1988

1990

1992

1994

1996

1998

Figure 9 Time series of observed surface layer meridional currents, with mean and annual cycle removed, at three locations along the Equator. Missing observations are indicated by gaps. Note that time series have been offset from each other for clarity.

correlation with the NECC, but there are also strong variations with longer timescales than ENSO. High-frequency Current Variability

Equatorially trapped waves with periods of a few days to a couple of months are quite energetic and ubiquitous (Figures 8 and 9). Zonal current fluctuations are dominated by intraseasonal fluctuations associated with the atmospheric intraseasonal (30– 60 days) oscillation; eastward-propagating equatorially trapped Kelvin waves play an important role in transmitting this variability from the western Pacific to the eastern Pacific (e.g., Figure 8A during 1997). Meridional current fluctuations tend to have more energy at higher frequencies than the zonal current variations (compare Figures 8 and 9). Here, dominant periods are in the range 20–30 days. The amplitude of meridional current variability is largest at the surface in the east, and becomes somewhat smaller in the central Pacific, and much smaller in the west. The dominant mechanism is the tropical instability wave, which arises from the strong shears of the zonal flows discussed earlier. Modulation of the amplitudes of these waves occurs seasonally (largest amplitude in the northern fall) and interannually (small amplitude during El Ninõ) as the shears are affected by the annual cycle and ENSO.

See also Data Assimilation in Models. East Australian Current. Ekman Transport and Pumping. Elemental Distribution: Overview. El Ninõ Southern Oscillation (ENSO). El Ninõ Southern Oscillation (ENSO) Models. Kuroshio and Oyashio Currents. Single Point Current Meters. Upper Ocean Time and Space Variability. Wind Driven Circulation.

Further Reading Godfrey JS, Johnson GC, McPhaden MJ, Reverdin G, and Wijffels S (2001) The tropical ocean circulation. In: Siedler G and Church J (eds.) Ocean Circulation and Climate, pp. 215--246. London: Academic Press. Hastenrath S (1985) Climate and Circulation of the Tropics. Dordrecht: D. Reidel. Lukas R (1986) The termination of the Equatorial Undercurrent in the Eastern Pacific. Progress in Oceanography 16: 63--90. Neumann G (1968) Ocean Currents. Amsterdam: Elsevier. Philander SGH (1990) El Ninõ, La Ninã, and the Southern Oscillation. San Diego: Academic Press.


PALEOCEANOGRAPHY E. Thomas, Yale University, New Haven, CT, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction: The Relevance of Paleoceanography Paleoceanography encompasses (as its name implies) the study of ‘old oceans’, that is, the oceans as they were in the past. In this context, ‘the past’ ranges from a few decades through centennia and millennia ago to the very deep past, millions to billions of years ago. In reconstructing oceans of the past, paleoceanography needs to be highly interdisciplinary, encompassing aspects of all topics in this encyclopedia, from plate tectonics (positions of continents and oceanic gateways, determining surface and deep currents, thus influencing heat transport) through biology and ecology (knowledge of present-day organisms needed in order to understand ecosystems of the past), to geochemistry (using various properties of sediment, including fossil remains, in order to reconstruct properties of the ocean waters in which they were formed). Because paleoceanography uses properties of components of oceanic sediments (physical, chemical, and biological) in order to reconstruct various aspects of the environments in these ‘old oceans’, it is limited in its scope and time resolution by the sedimentary record. Ephemeral ocean properties cannot be measured directly, but must be derived from proxies. For instance, several types of proxy data make it possible to reconstruct such ephemeral properties as temperature (see Determination of Past Sea Surface Temperatures) and nutrient content of deep and surface waters at various locations in the world’s oceans, and thus obtain insights into past thermohaline circulation patterns, as well as patterns of oceanic primary productivity. The information on ocean circulation can be combined with information on planktic and benthic microfossils, allowing a view of interactions between fluctuations in oceanic environments and oceanic biota, on short but also on evolutionary timescales. Paleoceanography offers information that is available from no other field of study: a view of a world alternative to, and different from, our present world, including colder worlds (‘ice ages’; see PlioPleistocene Glacial Cycles and Milankovitch Variability) as well as warmer worlds (see Paleoceanography: the Greenhouse World). Paleoceanographic data thus serve climate modelers in providing data on

boundary conditions of the ocean–atmosphere system very different from those in the present world. In addition, paleoceanography provides information not just on climate, but also on climate changes of the past, on their rates and directions, and possible linkages (or lack thereof) to such factors as atmospheric pCO2 levels (see Plio-Pleistocene Glacial Cycles and Milankovitch Variability) and the location of oceanic gateways and current patterns. Paleoceanography therefore enables us to gauge the limits of uniformitarianism: in which aspects is the present world indeed a guide to the past, in which aspects is the ocean–atmosphere system of the present world with its present biota just a snapshot, proving information only on one possible, but certainly not the only, stable mode of the Earth system? How stable are such features as polar ice caps, on timescales varying from decades to millions of years? How different were oceanic biota in a world where deepocean temperatures were 10–12 1C rather than the present (almost ubiquitous) temperatures close to freezing? Such information is relevant to understanding the climate variability of the Earth on different timescales, and modeling possible future climate change (‘global warming’), as recognized by the incorporation of a paleoclimate chapter in the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (2007). Paleoceanography thus enables us to use the past in order to gain information on possible future climatic and biotic developments: the past is the key to the future, just as much and maybe more than the present is the key to the past.

Paleoceanography: Definition and History Paleoceanography is a relatively recent, highly interdisciplinary, and strongly international field of science. International Conferences on Paleoceanography (ICP) have been held every 3 years since 1983, and the locations of these meetings reflect the international character of the paleoceanographic research community (Table 1). The flagship journal of paleoceanographic research, Paleoceanography, was first published by the American Geophysical Union in March 1986. In the editorial in its first volume, its target was defined as follows by editor J.P. Kennett: Paleoceanography publishes papers dealing with the marine sedimentary record from the present ocean basins and


295

296

PALEOCEANOGRAPHY

Table 1 International paleoceanographic conferences (ICP meetings) Meeting

Year

Location

ICP1 ICP2 ICP3 ICP4 ICP5 ICP6 ICP7 ICP8 ICP9

1983 1986 1989 1992 1995 1998 2001 2004 2007

Zurich, Switzerland Woods Hole, USA Cambridge, UK Kiel, Germany Halifax, Canada Lisbon, Portugal Sendai, Japan Biarritz, France Shanghai, China

margins and from exposures of ancient marine sediments on the continents. An understanding of past oceans requires the employment of a wide range of approaches including sedimentology; stable isotope geology and other areas of geochemistry; paleontology; seismic stratigraphy; physical, chemical, and biological oceanography; and many others. The scope of this journal is regional and global, rather than local, and includes studies of any geologic age (Precambrian to Quaternary, including modern analogs). Within this framework, papers on the following topics are to be included: chronology, stratigraphy (where relevant to correlation of paleoceanographic events), paleoreconstructions, paleoceanographic modeling, paleocirculation (deep, intermediate, and shallow), paleoclimatology (e.g., paleowinds and cryosphere history), global sediment and geochemical cycles, anoxia, sea level changes and effects, relations between biotic evolution and paleoceanography, biotic crises, paleobiology (e.g., ecology of ‘microfossils’ used in paleoceanography), techniques and approaches in paleoceanographic inferences, and modern paleoceanographic analogs.

Perusal of the volumes of the journal published since demonstrates that it indeed covers the full range of topics indicated above. What is the shared property of all these papers? They deal with various aspects of data generation or modeling using information generated from the sedimentary record deposited in the oceans (see Authigenic Deposits, Calcium Carbonates, Clay Mineralogy, Ocean Margin Sediments, Pore Water Chemistry, and Sediment Chronologies), including such materials as carbonates, clays, and authigenic minerals, as well as carbonate, phosphate, and opaline silica secreted by marine organisms, including for instance the calcium carbonate secreted by corals (see Past Climate from Corals). Sediments recovered from now-vanished oceans as well as sediments recovered from the present oceans are included, but paleoceanography as a distinct field of study is tightly linked to recovery of sediment cores from the ocean floor, deposited onto oceanic basement. Such sediments represent times from which oceanic crust is still in existence in the oceans (i.e., has not been subducted), which is about the last

200 My of Earth history (see Mid-Ocean Ridge Geochemistry and Petrology, Propagating Rifts and Microplates, Seamounts and Off-Ridge Volcanism, and Sediment Chronologies). Paleoceanography thus is a young field of scientific endeavor because the recovery of oceanic sediments in cores started in earnest only in the 1950s, long after the Challenger Expedition (1872–76), usually seen as the initiation of modern oceanography. Little research was conducted on ocean sediments until the late 1940s and 1950s, although some short cores were recovered by the German South Polar Expedition (1901–03), the German Meteor Expedition (1925–27), the English Discovery Expedition (1925–27), and Dutch Snellius Expedition (1929–30). This lack of information on oceanic sediments is documented by the statement in the book The Oceans by H. Sverdrup, N. Johnson, and R. Fleming published in 1942: From the oceanographic point of view, the chief interest in the topography of the seafloor is that it forms the lower and lateral boundaries of the water.

This situation quickly changed in the years after the World War II, when funding for geology and geophysics increased, and new technology became available to recover material from the seafloor. Sediment cores collected by the Challenger in 1873 reached a maximum length of 2 ft (Figure 1), and little progress in core collection was made in more than 50 years, so that cores collected by the Meteor (1925–27) reached only 3 ft, as described by H. Petterson in the Proceedings of the Royal Society of London in 1947. In 1936, the American geophysicist C.S. Piggott obtained a 10-ft core using an exploding charge to drive a coring tube into the sediment, and in the early 1940s H. Petterson and B. Kullenberg developed the prototype of the piston corer. Piston corers were first used extensively on the Swedish Deep-Sea Expedition (1946–47) using the Albatross, the first expedition to focus ‘‘on the bottom deposits, their chemistry, stratigraphy, etc.,’’ and thus arguably the first paleoceanographic expedition. A major expansion occurred in US oceanographic institutions in the postwar years, and piston corers were used extensively by Woods Hole Oceanographic Institution (Woods Hole, Massachusetts) with its research vessel Atlantis, the first American ship built for sea research (1931–66), Scripps Institution of Oceanography (La Jolla, California) with the E.W. Scripps (1937–55), and the Lamont Geological Observatory of Columbia University (Palisades, New York), now the Lamont-Doherty Earth Observatory, with the Vema (1953–81). In 1948, only about 100 deep-sea cores existed, and by 1956 the Vema alone had collected 1195 cores.


PALEOCEANOGRAPHY

1873 1925 1936 1942 1945 65′ 10′ 45′ 2′ 3′ 85 ky 135 ky 450 ky

2 My

3 My Figure 1 Lengths of sediments cores (in ft) obtained by the Challenger (1873), the Meteor (1925), Piggott’s explosive-driven coring device (1936), and piston cores obtained by corers constructed by H. Petterson and B. Kullenberg, Institute of Oceanography, Go¨teborg, Sweden (1942–45). Adapted from Petterson H (1947) A Swedish deep-sea expedition. Proceedings of the Royal Society of London, Series B: Biological Sciences 134(876): 399–407, figure 5.

Before the Swedish Deep-Sea Expedition, only limited estimates of the sedimentation rates in the deep ocean were available, described by H. Petterson in 1947 as ‘‘the almost unknown chronology of the oceans’’ (Figure 1). The development of radiocarbon dating of carbonates (see Sediment Chronologies) in combination with geochemical determination of the titanium content of sediments as a tracer for clay content led to estimates of sedimentation rates by G. Arrhenius, G. Kjellberg, and W.L. Libby in 1951, and of differences in sedimentation rates in glacial and interglacial times by W.S. Broecker, K.K. Turekian, and B.C. Heezen in 1958. Cores collected by the Albatross during the Swedish Deep-Sea Expedition as well as the many later expeditions of the Atlantis and Vema were used extensively in seminal paleoceanographic studies of the Pleistocene ice ages, including those of the variation of calcium carbonate accumulation by G. Arrhenius and E. Olausson, and of the variations in populations of pelagic microorganisms (foraminifera; see Protozoa,

297

Planktonic Foraminifera) by F. Phleger, F. Parker, and J.F. Peirson published in 1953, as well as D.B. Erickson and G. Wollin published in 1956. The oxygen isotopic method to reconstruct past oceanic temperatures using carbonate sediments, as outlined by H. Urey in a paper in Science in 1948, was first applied to oceanic microfossils by C. Emiliani, in order to reconstruct ocean surface temperatures during the Pleistocene ice ages. It turned out that the oxygen isotope record combined signals of temperature change as well as ice volume (see Cenozoic Climate – Oxygen Isotope Evidence and Determination of Past Sea Surface Temperatures), but the method has become widely established since those early days. Oxygen isotope analysis was the major tool in the first attempt to look at paleoclimate globally, in the CLIMAP (Climate/Long Range Investigation Mappings and Predictions) project in the early 1970s, which led to the documentation that major, long-term changes in past climate are associated with variations in the geometry of the Earth’s orbit, as described by J. Hays, J. Imbrie, and N.J. Shackleton in 1976 in Science (see Plio-Pleistocene Glacial Cycles and Milankovitch Variability). Oxygen isotope analysis is still one of the ‘workhorses’ of paleoceanographic research, as documented in the review of global climate of the last 65 My by J.C. Zachos and others in Science (2001). Most material older than the last few hundred thousands of years became available in much longer cores which could be recovered only after the start of drilling by the Deep Sea Drilling Project (DSDP) in 1968, an offshoot of a 1957 suggestion by W. Munk (Scripps Institution of Oceanography) and H. Hess (Princeton University) to drill deeply into the Earth and penetrate the crust–mantle boundary. In 1975, various countries joined the United States of America in the International Phase of Ocean Drilling, and DSDP became a multinational enterprise. Between 1968 and 1983, the drill ship Glomar Challenger recovered more than 97 km of core at 624 drill sites. Cores recovered by DSDP provided the material for the first paper using oxygen isotope data to outline Cenozoic climate history and the initiation of ice sheets on Antarctica, published in the Initial Reports of the Deep Sea Drilling Project in 1975 by N.J. Shackleton and J.P. Kennett. DSDP’s successor was the Ocean Drilling Program (1983–2003), which recovered more than 222 km sediment at 652 sites with its vessel Joides Resolution. These two programs were succeeded by the present Integrated Ocean Drilling Program (IODP), which plans to greatly expand the reach of the previous programs by using multiple drilling platforms,


298

PALEOCEANOGRAPHY

including riser drilling by the Japanese-built vessel Chikyu, riserless drilling by a refitted Joides Resolution, and mission-specific drilling using various vessels (see Deep-Sea Drilling Methodology).

Paleoceanographic Techniques Over the last 30 years there has been an explosive development of techniques for obtaining information from oceanic sediments (Figure 2). One of the limitations of paleoceanographic research on samples from ocean cores is the limited size of each sample. However, a positive result of this limited availability is that researchers from different disciplines are forced to work closely together, leading to the generation of independent proxy records on the same sample set, thus integrating various aspects of chemical and biotic change over time. Proxies are commonly measured on carbonate shells of pelagic and benthic microorganisms (see Benthic Foraminifera, Coccolithophores, and Protozoa, Planktonic Foraminifera), thus providing records from benthic and several planktonic environments (surface, deep thermocline). Methods used since the first paleoceanographic core studies include micropaleontology, with the

Chemical and isotopic properties:

Atmosphere Influences from land: river inflow, nutrient fluxes, weathering rates

Dust transport (nutrients) DMS

δ13C, δ18O, δ15N, Uk37, TEX86, Mg/Ca, Sr/Ca, δ11B, Ba/Ca, 86/87Sr

δ13C,

δ18O,

most commonly studied fossil groups including pelagic calcareous (see Protozoa, Planktonic Foraminifera) and siliceous-walled (see Protozoa, Radiolarians) heterotroph protists, organic-walled cysts of heterotroph and autotroph dinoflagellates, siliceous-walled (Diatoms) and calcareous-walled autotroph protists (see Coccolithophores), benthic protists (see Benthic Foraminifera), and microscopic metazoa, Ostracods (Crustacea). Micropaleontology is useful in biostratigraphic correlation as well as in its own right, providing information on evolutionary processes and their linkage (or lack thereof) to climate change, as well as on changes in oceanic productivity (see Sedimentary Record, Reconstruction of Productivity from the). Classic stable isotopes used widely and commonly in paleoceanographic studies include those of oxygen (see Cenozoic Climate – Oxygen Isotope Evidence) and carbon (see Cenozoic Oceans – Carbon Cycle Models). Carbon isotope records are of prime interest in investigations of deep oceanic circulation (see Ocean Circulation: Meridional Overturning Circulation) and of oceanic productivity (see Sedimentary Record, Reconstruction of Productivity from the, and Tracers of Ocean Productivity).

δ34S,

Mg/Ca, Sr/Ca, δ11B, Ba/Ca, Cd/Ca

O2

CO2

Photosynthesis

Pelagic world O2

CO2

Nutrients N, P, Si

Stratification Upwelling Seasonality Community Structure

Biological pump:

Nutrient regeneration

SIO2, CaCO3, Corg, phosphate, barite Respiration, decomposition (CO2)

Benthic world

Epifauna Infauna

Burial Sediment

Microbial loop

Controls:

Bioturbation

Oxic/anoxic boundary

Reflux from bacterial ‘deep biosphere’: methane, sulfide, barium Figure 2 Linkages between the marine biosphere and global biogeochemical cycles, as well as to various proxies used in paleoceanographic studies. The proxies include isotope measurements (indicated by lower case deltas, followed by the elements and the heavier isotope), elemental ratios, and organic geochemical temperature proxies (e.g., Uk37’, an alkenone ratio in which the 37 refers to the carbon number of the alkenones; TEX86, a proxy based on the number of cyclopentane rings in sedimentary membrane lipids derived from marine crenarchaeota). DMS is dimethylsulfide, a sulfur compound produced by oceanic phytoplankton. The various proxies can be used to trace changes in ocean chemistry (including alkalinity), temperature, and productivity, as well as changes in the reservoirs of the carbon cycle. Adapted from Pisias NG and Delaney ML (eds.) (1999) COMPLEX: Conference of Multiple Platform Exploration of the Ocean. Washington, DC: Joint Oceanographic Institutions.


PALEOCEANOGRAPHY

In addition to these classic paleoceanographic proxies, many new methods of investigation have been and are being developed using different geochemical (isotope, trace element, organic geochemical) proxies (see Determination of Past Sea Surface Temperatures). Many more proxies are in development, including proxies on different organic compounds, to investigate aspects of global biogeochemical cycles, biotic evolution and productivity, and thermohaline circulation patterns (see Ocean Circulation: Meridional Overturning Circulation) (Figure 2). Techniques to correlate the age of features in sediment records recovered at different locations and to assign numerical ages to sediment samples are of the utmost importance to be able to estimate rates of deposition of sediments and their components. Correlation between sediment sections is commonly achieved by biostratigraphic techniques, which cannot directly provide numerical age estimates, and are commonly limited to a resolution of hundreds of thousands of years. Techniques used in numerical dating include the use of various radionuclides (see Sediment Chronologies), the correlation of sediment records to the geomagnetic polarity timescale (see Geomagnetic Polarity Timescale), and the more recently developed techniques of linking high-resolution records of variability in sediment character (e.g., color, magnetic susceptibility, density, and sediment composition) to variability in climate caused by changes in the Earth’s orbit and thus energy supplied by the sun to the Earth’s surface at specific latitudes (see Paleoceanography: Orbitally Tuned Timescales). Such an orbitally tuned timescale has been fully developed for the last 23 My of Earth history, with work in progress for the period of 65–23 Ma. Remote sensing techniques are being used increasingly in order to characterize sediment in situ in drill holes, even if these sediments have not been recovered (see Deep-Sea Drilling Methodology), and to establish an orbital chronology even under conditions of poor sediment recovery.

Contributions of Paleoceanographic Studies Paleoceanographic studies have contributed to a very large extent to our present understanding that the Earth’s past environments were vastly different from today’s, and that changes have occurred on many different timescales. Paleoceanographic studies have been instrumental in establishing the fact that climate change occurred rapidly and stepwise rather than gradually, whether in the establishment of the Antarctic ice cap on timescales of tens to hundreds of

299

thousands of years rather than millions of years, or in the ending of a Pleistocene ice age on a timescale of decades rather than tens of thousands of years (see Abrupt Climate Change and Millennial-Scale Climate Variability). Unexpected paleoceanographic discoveries over the last few years include the presence of large amounts of methane hydrates (clathrates), in which methane is trapped in ice in sediments along the continental margins (see Ocean Margin Sediments), in quantities potentially larger than the total global amount of other fossil fuels. The methane in gas hydrates may become a source of energy, with exploratory drilling advanced furthest in the waters off Japan and India. Such exploration and use of gas hydrates might, however, lead to rapid global warming if drilling and use of methane hydrates inadvertently lead to uncontrolled destabilization. Destabilization of methane hydrates may have occurred in the past (see Methane Hydrate and Submarine Slides) as a result of changes in sea level (see Sea Level Change and Sea Level Variations Over Geologic Time) and/or changes in thermohaline circulation and subsequent changes in deep-ocean temperature (see Methane Hydrates and Climatic Effects). Dissociation of gas hydrates and subsequent oxidation of methane in the atmosphere or oceans have been speculated to have played a role in the ending of ice ages, and in a major upheaval in the global carbon cycle and global warming (see Methane Hydrates and Climatic Effects). The influence on global climate and the global carbon cycle of methane hydrate reservoirs (with their inherent capacity to dissociate on timescales of a few thousand years at most) is as yet not well understood or documented (see Carbon Cycle and Ocean Carbon System, Modeling of ). Most methane hydrates are formed by bacterial action upon organic matter, and another unexpected discovery was that of the huge and previously unknown microbial biomass in seawater, also found buried deep in the sediments (see Bacterioplankton and Microbial Loops). Fundamental issues such as the conditions that support and limit this biomass are still not understood and neither are their linkage to the remainder of the oceanic biosphere and the role of chemosynthesis and chemosymbiosis in the deep oceanic food supply and the global carbon cycle. In contrast to these unexpected discoveries, paleoceanographers of a few decades ago could have predicted at least in part our increased knowledge of aspects of climate change such as the patterns of change in sea level at various timescales (see Sea Level Variations Over Geologic Time). The


300

PALEOCEANOGRAPHY

suddenness and common occurrence of rapid climate change events in Earth history, however, was unpredicted. On timescales of millions of years, the Earth’s climate was warm globally during most of the Cretaceous and the early part of the Cenozoic (65–35 Ma), and the Earth had no large polar ice caps reaching sea level (see Paleoceanography: the Greenhouse World). Drilling in the Arctic Ocean, for instance, established that average summer surface water temperatures might have reached up to 18 1C. The use of climate models has assisted in understanding such a warm world (see Paleoceanography, Climate Models in), but the models still cannot fully reproduce the necessary efficient heat transport to high latitudes at extremely low latitudinal temperature gradients. Paleoceanographic research has provided considerable information on the major biogeochemical cycles over time (see Cenozoic Oceans – Carbon Cycle Models). Geochemical models of the carbon cycle rely on carbon isotope data on bulk carbonates and on planktonic and benthic foraminifera in order to evaluate transfer of carbon from one reservoir (e.g., organic matter, including fossil fuel; methane hydrates) to another (e.g., limestone, the atmosphere, and dissolved carbon in the oceans). Information on pelagic carbonates and their microfossil content as well as stable isotope composition has assisted in delineating the rapidity and extent of the extinction at the end of the Cretaceous in the marine realm. Sedimentological data at many locations have documented the large-scale failure of the western margin of the Atlantic Ocean, with large slumps covering up to half of the basin floor in the North Atlantic as the result of the asteroid impact on the Yucatan Peninsula. One of the major successes of paleoceanographic research has been the establishment of the nature and timing of polar glaciation during the Cenozoic cooling (see Cenozoic Climate – Oxygen Isotope Evidence and Paleoceanography: the Greenhouse World). After a prolonged period of polar cooling in the middle to late Eocene, the East Antarctic Ice Sheet became established during a period of rapid ice volume growth (o100 ky) in the earliest Oligocene, c. 33.5 Ma. This establishment was followed by times of expansion and contraction of the ice sheet, and the West Antarctic Ice Sheet may have started to grow at c. 14 Ma. Until recently it was argued that the Northern Hemispheric Ice Sheets formed much later: these ice sheets increased in size around 3 Ma (in the Pliocene), and ever since have contracted and expanded on orbital timescales (see Plio-Pleistocene Glacial Cycles and Milankovitch Variability). There is now considerable evidence that the polar ice sheets

in the Northern Hemisphere also became established, at least in part, in the earliest Oligocene, that is, at a similar time as the Southern Hemisphere ice sheets. The cause(s) of the long-term Cenozoic cooling are not fully known. Possible long-term drivers of climate include the opening and closing of oceanic gateways, which direct oceanic heat transport (see Heat Transport and Climate). Such changes in gateway configuration include the opening of the Tasman Gateway and Drake Passage which made the Antarctic Circumpolar Current possible (see Antarctic Circumpolar Current), and closing of the Isthmus of Panama which ended the flow of equatorial currents from the Atlantic into the Pacific (see Atlantic Ocean Equatorial Currents and Pacific Ocean Equatorial Currents). Evidence has accumulated, however, that changes in atmospheric CO2 levels may have been more important than changes in gateway configuration. For instance, long-term episodes of global warmth (in timescales of millions of years) may have been sustained by CO2 emissions from large igneous provinces (see Igneous Provinces), and short-term global warming (on timescales of ten to one or two hundred thousands of years) could have been triggered by the release of greenhouse gases from methane hydrate dissociation or burning/ oxidation or organic material (including peat). Decreasing atmospheric CO2 levels due to decreasing volcanic activity and/or increased weathering intensity have been implicated in the long-term Cenozoic cooling, and high-resolution paleoceanographic records show that climate change driven by changes in atmospheric CO2 levels may have been modulated by changes in insolation caused by changes in orbital configuration. Recognition of variability in climatic signals in sediments at orbital frequencies has led to major progress in the establishment of orbitally tuned timescales throughout the Cenozoic (see Paleoceanography: Orbitally Tuned Timescales). Paleoceanographic research has led to greatly increased understanding of the Plio-Pleistocene Ice Ages as being driven by changes in the Earth’s orbital parameters (see Paleoceanography: Orbitally Tuned Timescales and Plio-Pleistocene Glacial Cycles and Milankovitch Variability), and the correlation of data from oceanic sediments to records from ice cores on land. Orbital forcing is the ‘pacemaker of the ice ages’, but it is not yet fully understood how feedback processes magnify the effects of small changes in insolation into the major climate swings of the Plio-Pleistocene. Ice-core data and carbon isotope data for marine sediments show that changes in atmospheric CO2 levels play a role. We also do not yet understand why the amplitude of these orbitally


PALEOCEANOGRAPHY

driven climate swings increased at c. 0.9 Ma, and why the dominant periodicity of glaciation switched from 40 000 (obliquity) to 100 000 (eccentricity) years at that time, the ‘mid-Pleistocene revolution’. Effects of glaciation at low latitudes, including changes in upwelling, productivity, and monsoonal activity, are only beginning to be documented (see Monsoons, History of). Great interest has been generated by the information on climate change at shorter timescales than 20 ky, the duration of the shortest orbital cycle, precession. Such climate variability includes the millennial-scale changes that occurred during the glacial periods (see Millennial-Scale Climate Variability) and the climate variations of lesser amplitudes that have occurred since the last deglaciation (see Holocene Climate Variability). These research efforts are beginning to provide information on timescales that are close to the human timescale, on such topics as abrupt climate change (see Abrupt Climate Change), and the fluctuations in intensity and occurrence of the El Ninõ–Southern Oscillation (see El Ninõ Southern Oscillation (ENSO)) and the North Atlantic Oscillation (NAO) (see North Atlantic Oscillation (NAO)), during overall colder and warmer periods of the Earth history.

301

those recovered in deep-sea cores, and new proxies continue to be developed, including proxies (e.g., on levels of oxygenation and temperature) using organic geochemical methods. The IODP started in 2003 and is scheduled to start drilling in 2008 with three different platforms. Drilling activity has become integrated with that of the French research vessel Marion Dufresne, which has recovered many long piston cores (several tens of meters) for studies covering the last few hundred thousand years of the Earth history, in the International Marine Global Change Study (IMAGES) program. Examples of drilling by alternative platforms include the drilling in the Arctic Ocean, one of the frontiers in ocean science, and drilling in shallower regions than accessible to the drilling vessel Joides Resolution, such as coral drilling in Tahiti for studies of Holocene climate and rates of sea level rise. If these ambitious programs are carried out, we can expect to learn much about the working of the Earth system of lithosphere–ocean–atmosphere– biosphere, specifically about the sensitivity of the climate system, about the controls on the long-term evolution of this sensitivity, and about the complex interaction of the biospheric, lithospheric, oceanic, and atmospheric components of the Earth system at various timescales.

The Future of Paleoceanography Past progress in paleoceanography has been linked to advances in technology since the invention of the piston corer in the 1940s. In the early to mid-1980s, paleoceanographic studies appeared to reach a plateau, with several review volumes published on, for example, Plio-Pleistocene ice ages (CLIMAP), the oceanic lithosphere, and the global carbon cycle, as well as a textbook Marine Geology by J.P. Kennett. Paleoceanographic research, however, has benefited from research programs in the present oceans (such as the Joint Global Ocean Flux Program, JGOFS), and new technology has spurred major research activity. The extensive use and improvement of the hydraulic piston corer by DSDP/ODP/IODP led to recovery of minimally disturbed soft sediment going back in age through the Cenozoic, making high-resolution studies possible, including paleoceanographic and stratigraphic studies of sediment composition using the X-ray fluorescence (XRF) scanner. Progress in computers led to strongly increased use of paleoceanographic data in climate modeling, and to increased possibilities of remote sensing in drill holes. Developments in mass spectrometry led to the possibility to measure isotopes and trace elements in very small samples, such as

See also Abrupt Climate Change. Atlantic Ocean Equatorial Currents. Authigenic Deposits. Bacterioplankton. Benthic Foraminifera. Calcium Carbonates. Carbon Cycle. Cenozoic Climate – Oxygen Isotope Evidence. Cenozoic Oceans – Carbon Cycle Models. Clay Mineralogy. Coccolithophores. Deep-Sea Drilling Methodology. Determination of Past Sea Surface Temperatures. El Ninõ Southern Oscillation (ENSO). Geomagnetic Polarity Timescale. Heat Transport and Climate. Holocene Climate Variability. Igneous Provinces. Methane Hydrates and Climatic Effects. Methane Hydrate and Submarine Slides. Microbial Loops. Mid-Ocean Ridge Geochemistry and Petrology. MillennialScale Climate Variability. Monsoons, History of. North Atlantic Oscillation (NAO). Ocean Carbon System, Modeling of. Ocean Circulation: Meridional Overturning Circulation. Ocean Margin Sediments. Pacific Ocean Equatorial Currents. Paleoceanography, Climate Models in. Paleoceanography: Orbitally Tuned Timescales. Paleoceanography: the Greenhouse World. Past Climate from Corals. Pelagic Biogeography. PlioPleistocene Glacial Cycles and Milankovitch Variability. Pore Water Chemistry. Propagating Rifts and Microplates. Protozoa, Planktonic


302

PALEOCEANOGRAPHY

Foraminifera. Protozoa, Radiolarians. Sea Level Change. Sea Level Variations Over Geologic Time. Seamounts and Off-Ridge Volcanism. Sediment Chronologies. Sedimentary Record, Reconstruction of Productivity from the. Tracers of Ocean Productivity.

Pisias NG and Delaney ML (eds.) (1999) COMPLEX: Conference of Multiple Platform Exploration of the Ocean. Washington, DC: Joint Oceanographic Institutions. Sarmiento JL and Gruber N (2006) Ocean Biogeochemical Dynamics. Princeton, NJ: Princeton University Press.

Further Reading Elderfield H (2004) The Oceans and Marine Geochemistry. In: Holland HD and Turekian KK (eds.) Treatise on Geochemistry, vol. 6. Amsterdam: Elsevier. Gradstein F, Ogg J, and Smith AG (2004) A Geologic Time Scale. Cambridge, UK: Cambridge University Press. Hillaire-Marcel C and de Vernal A (2007) Proxies in Late Cenozoic Paleoceanography. Amsterdam: Elsevier. National Research Council, Ocean Sciences Board (2000) 50 years of Ocean Discovery. Washington, DC: National Academies Press. Oceanography (2006) A Special Issue on The Impact of the Ocean Drilling Program. Oceanography 19(4). Olausson E (1996) The Swedish Deep-Sea Expedition with the ‘Albatross’ 1947–1948: A Summary of Sediment Core Studies. Go¨teborg, Sweden: Novum Grafiska AB. Petterson H (1947) A Swedish deep-sea expedition. Proceedings of the Royal Society of London, Series B: Biological Sciences 134(876): 399--407.

Relevant Websites http://www.andrill.org – ANDRILL: Antarctic drilling. http://www.ecord.org – European Consortium for Ocean Research Drilling (ECORD). http://www.iodp.org – Integrated Ocean Drilling Program (IODP). http://www.ipcc-wg1.ucar.edu – Intergovernmental Panel on Climate Change, Working Group 1: The Physical Basis of Climate Change. http://www.ngdc.noaa.gov – National Geophysical Data Center (Marine Geology and Geophysics). http://www.images-pages.org – The International Marine Past Global Change Study (IMAGES).


PALEOCEANOGRAPHY, CLIMATE MODELS IN W. W. Hay, Christian-Albrechts University, Kiel, Germany Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2082–2089, & 2001, Elsevier Ltd.

Introduction Climate models have been applied to the investigation of ancient climates for about 25 years. The early investigations used simple energy balance models, but since the 1970s increasingly more elaborate atmospheric general circulation models (AGCMs) have been applied to paleoclimate problems. It is only within the last decade that climate model results have been used to drive ocean circulation models to simulate the behavior of ancient oceans. Similarly, increasingly more complex ocean models have been used since the 1970s to explore the effects of changing boundary conditions on ocean circulation and climate. The early paleo-ocean models were driven by modern winds and hydrologic cycle data; the more recent simulations use paleoclimate-model-generated data. In some of the early models sea surface temperatures were specified, but this locks the model to a particular solution. Runoff from land has only been included in the most recent models. Experiments with coupled atmosphere– ocean models are just beginning. Paleo-ocean modeling has been directed largely towards trying to understand two major problems related to possible changes in ocean heat transport with global climatic implications: (1) the origin of the Late Cenozoic glacial ages, with special attention to the initiation and cessation or slowdown of the ‘Global Conveyor System,’ and (2) the warm climates of the Late Mesozoic and Eocene, with low meridional temperature gradients. One group of modeling exercises has concentrated on the effects of interocean connections, particularly on the effects of opening and closing of gateways between the major ocean basins on the global thermohaline circulation system. Another major set of investigations has focused on ocean surface and thermohaline circulation on a warm Earth, with higher concentrations of atmospheric CO2 and the very different paleogeography of the Late Mesozoic.

Ocean Gateways and Climate The effects of interocean gateways on ocean circulation and the global climate system have been a matter of speculation since the late 1960s and of experimentation via box models and numerical simulation since the 1980s. Figure 1 shows the major gateways that have opened or closed during the Cenozoic. In most instances the timing of opening or closing of gateways has been deduced from the regional plate tectonic context, but has an uncertainty of a few million years. In some cases, such as the opening of passages between Australia, South America and Antarctica the plate tectonic record is ambiguous and the timing is inferred from climate change in the surrounding regions. For a few passages, such as Panama and Gibraltar, the timing is known to within a few tens of thousands of years because the resulting oceanographic changes are reflected in ocean floor sediments on either side of the gateway. Until now it has been possible to explore the importance of only a few of the most critical gateways for the global climate system. Most of these investigations have concentrated on exploring the role of salinity in forcing the thermohaline circulation to develop an understanding of the behavior of the ‘Great Conveyor.’ The term ‘Great Conveyor’ is used to describe the global thermohaline circulation that starts with sinking of cold saline water in the Norwegian–Greenland Sea and continues southward as the North Atlantic Deep Water (NADW). On reaching the Southern Ocean, it is supplemented by water sinking in the region of the Weddell Sea. The deep water then flows into the Indian and Pacific Oceans where it returns to the surface. The return surface flow is then from the Pacific through the Indonesian Passages and across the Indian Ocean, then around South Africa into the South Atlantic and finally northward to the Norwegian–Greenland Sea. Paleoceanographic data show that the production of NADW has slowed or even stopped from time to time. This has major implications for the paleoclimate of the continents surrounding the North Atlantic because it is the sinking of water in the Norwegian–Greenland Sea that draws warm surface waters northward along the western margin of Europe and Scandinavia. It has been argued on the basis of simple model calculations that the Atlantic acts as a salt oscillator. The southward flow of NADW is thought to export salt from the North Atlantic. When so much salt has


303

304

PALEOCEANOGRAPHY, CLIMATE MODELS IN

Bering Strait ? Depth history uncertain ?

Nares Strait ? Depth history uncertain ?

Fram Strait ? 5 Ma ?

Tsushima Straits ? Depth history uncertain ?

Tethyan Passages Many openings and closings during Neogene

'Gibraltar' Closed from 5 to 6 Ma

Isthmus of Tehuantepec ? Closed ca. 2.5 Ma ?

Denmark Strait _ Iceland-Faeroes Passages Opening and (mostly) deepening since 26 Ma

Lesser Antillean Arc ? Depth history uncertain ?

Panamanian Isthmus Closing from 13 to ca. 2.5 Ma

Luzon Straits ? Depth history uncertain ? Bab-el Mendab Opened ca. 8 Ma Indonesian Passages Restriction since 20 Ma

Drake Passage Opened 30 Ma ??

African_Antarctic Passage Opened ca. 140 Ma but circulation is restricted by the Subtropical Front

Tasman_Antarctic Passage Opening from 50 to 26 Ma

Figure 1 Interocean gateways which have opened or closed during the Cenozoic (since 65 Ma).

been exported that the surface waters are no longer so saline, deep water formation in the Norwegian– Greenland Sea will stop, shutting down the Global Conveyor. The Global Conveyor is revived by the loss of water from the Atlantic to the Pacific through the atmosphere across Central America. This increases the salinity of the North Atlantic to levels where the deep water formation is initiated again. It has also been suggested that the saline Mediterranean outflow is an important forcing factor controlling the generation of NADW. Some of the Mediterranean outflow water becomes entrained in the surface waters crossing the Iceland–Scotland Ridge, raising the salinity to a level that can promote sinking as the water cools. The salinity control hypothesis has been questioned by authors who argue that the global thermohaline circulation system is driven by Southern Hemisphere winds, and that the salinity difference between the Atlantic and Pacific plays only a minor role. This conclusion is based on an idealized ocean general circulation model (OGCM) using the simple geometry of two polar continents connected by a long thin meridional land mass. The model has no salinity forcing or sea ice formation, and no separate Atlantic Basin. The ocean model is coupled to a simple energy balance atmosphere with imposed zonal winds. Removing a piece of the land mass at the latitude of the Drake Passage induces an interhemispheric conveyor, similar to the modern Great Conveyor, which warms the entire Northern Hemisphere by several degrees. In this case, the

conveyor is driven by the Southern Hemisphere winds and salinity plays no role at all. Both of these models are great oversimplifications of the real world. They raise the question of whether the same results will be obtained with simulations having more realistic geographic boundary conditions, wind forcing, and radiative balance.

Drake Passage

Paleoceanographers have assumed that the opening of the Tasman–Antarctic Passage and subsequent opening of the Drake Passage led to isolation of the Antarctic continent. It has been argued that this resulted in a sharp meridional thermal gradient across the Southern Ocean and promoted the glaciation of the Antarctic continent. The importance of the Drake Passage for the modern thermohaline circulation has been recognized by physical oceanographers since 1971 when scientists at the Geophysical Fluid Dynamics Laboratory in Princeton, New Jersey used an early OGCM to experiment with a closed Drake Passage. They found that closure of the passage would increase the outflow of Antarctic Bottom Water (AABW) to the world ocean. The same result has been obtained for the effect of closure of the Drake Passage using other ocean models. It has been concluded that the opening of the Drake Passage was responsible for creating the circulation system we observe today.



Isthmus of Panama

Another important event affecting Earth’s climate history was the closure of the connection between the Atlantic and Pacific across Central America. It has long been suspected that this paleographic change set the stage for the Northern Hemisphere glaciation. Workers at the Max Planck Institute for Meteorology in Hamburg, explored the effects of an open Central American Passage using a model with a 3.51 3.51 grid with 11 vertical levels, realistic bottom topography, and a full seasonal cycle. Although it does not simulate mesoscale eddies, the Hamburg OGCM does well at simulating the present surface and deep circulation of the ocean. A Central American Passage was opened between Yucatan and South America, and a sill depth of 2711 m was specified. The simulation was forced by the modern observed wind stress field and air temperatures. The large-scale low-latitude interchanges of waters between the Atlantic and Pacific had dramatic effects. The regional slope on the ocean surface from the western Pacific to the Norwegian–Greenland Sea was reduced. Flow through the Central American Passage was 1 Sv (1 Sv ¼ 106 m3 s1) of wind-driven surface water from Atlantic to Pacific and 10 Sv of subsurface water driven by the hydrostatic head from Pacific to Atlantic. Flow through the Bering Strait, presently from the Pacific to the Arctic, was reversed, with a low salinity outflow from the Arctic into the North Pacific. There was little effect on flow through the Drake Passage. The salinity difference between the Atlantic and Pacific was greatly reduced. The production of NADW ceased, and flow of the Gulf Stream was reduced whereas flow of the Brazil Current was enhanced. The outflow of AABW into the world ocean increased about 25% with the closed Central American Passage. Southward oceanic heat transport also increased by about 25%, as more warm water from the South Atlantic was drawn southward to replace the surface water sinking in the Weddell Sea to form AABW. These changes would have major implications for global climates, making Europe colder and providing a larger snow source for the Antarctic. The results of this experiment did much to stimulate further studies. The extreme sensitivity of the Atlantic thermohaline circulation to an open or closed Panama Strait has not been borne out by more realistic coupled atmosphere–ocean models (AOGCM). A subsequent experiment found a critical threshold for water transport through the passage affecting formation of NADW. Flows up to 5 Sv from the Pacific to the Atlantic through the passage still allow NADW to

305

form, but at a rate only about 50% that of the present day. Greater flows shut down the NADW production. In contrast, a recent AOGCM simulation using the recently developed Fanning and Weaver atmospheric energy-moisture balance model indicated that NADW formation could take place even with a fully open passage through Panama. Combined Opening and Closure of the Drake and Panama Passages

One set of experiments has combined the effect of closing the Drake Passage and opening Panama. Two scenarios were examined, a closed Drake Passage and closed Central American Passage, and a closed Drake Passage with an open Central American Passage. Atmospheric forcing of the ocean model was with present day winds. The results are summarized in Figure 2. During the past 60 million years the Drake and Central American Passages have not been closed at the same time. Nevertheless, it is an interesting experiment to examine the effect of a pole-to-pole meridional seaway connected to the world ocean through a high latitude passage on the east (between Africa and Antarctica). The major effect of this configuration was to increase the outflow of AABW to the world ocean by a factor of four and suppress NADW production. The net effect is a doubling of the global rate of thermohaline overturning. Enhanced upwelling of AABW in the northern Atlantic reduced the salinity of the surface waters there, preventing formation of deep water. The scenario with a closed Drake Passage and open Central American Passage corresponds to the paleogeography of the Atlantic from mid-Cretaceous through most of the Oligocene. The result of the experiment was similar to that closing both passages, except that the increase in AABW was less, about 3.2 times that at present. Production of NADW was suppressed, but the overall thermohaline circulation increased. In the South Atlantic, the deep waters were cooler and the thermocline weaker, promoting overturning and cooling of the surface waters. The flow of the Brazil Current was increased and the flow of the North Equatorial Current along the northern margin of South America was reduced. There was reduced equatorward transport in the eastern South Pacific and reduced flow of the Pacific western boundary current. The most controversial result of this experiment was that analysis of the surface heat flux suggested that the opening of the Drake Passage did not result in temperature changes large enough to have triggered Antarctic glaciation.


306


Central American Passage closed

Drake Passage open

0

Central American Passage open

80 Sv

1 Sv ⊗

0

Depth (km)

Depth (km)

80 Sv

NADW 17 Sv

AABW 38 Sv

_4

N

Drake Passage closed

10 Sv

AABW 48 Sv

_4

S

(A)

Drake Passage open

(B)

Drake Passage Central American Passage open closed

Central American Passage closed

0

⊗

Depth (km)

Depth (km)

0

AABW

AABW

150 Sv

115 Sv

_4 (C)

N

S

⊗

_4 S

N

(D)

S

N

Figure 2 Pacific to Atlantic flow with open and closed Panama and Drake Passages. (After Mikalojewicz et al. 1993.) } ¼ flow from Pacific to Atlantic. # ¼ flow from Atlantic to Pacific. Volume fluxes in Sverdrups (Sv).

The effects of different widths of the Central American Passage and different sill depths in the Drake and Central American Passages were also investigated. It was found that widening the low latitude passage by removing all of Central America has little effect. A shallow sill (700 m) in the Drake Passage produced effects intermediate between experiments with deep sill (2700 m) used for the control run and the closed Drake Passage run. An experiment was also conducted with a deep (2700 m) Drake Passage and a very deep (4100 m) Central American Passage. The very deep Central American Passage allowed AABW flowing northward from the strong source in the Weddell Sea to enter the Pacific Basins, resulting in a reversal of the deep circulation in the Pacific. The flow of the western boundary current in the South Pacific reversed from northward (present) to southward. Another experiment with the idealized OGCM explored the effect of making a gap in the northern hemisphere in the latitude band of the easterly trade winds. This corresponds roughly to a closed Drake Passage and an open Panama Passage. It was found that this produced an overturning circulation that cooled the tropics and warmed the high latitudes.

conditions in OGCMs. The opening and closing of the Straits of Gibraltar have been modeled by turning a ‘salt source’ at Gibraltar on and off. One simulation indicates that open Straits of Gibraltar have only a minor effect on NADW production. Compared with closed Straits, the open Straits scenario increased the export of NADW from the North Atlantic by about 1 Sv. Indonesian–Malaysian Passages

The connection between the Pacific and Indian Oceans through the Indonesian region is a critical part of the Great Conveyor. The passages through this region have become increasingly restricted as the Australian Block moves northward and collides with south-east Asia. Unfortunately, few model experiments have explored the effect on global ocean circulation. OGCM sensitivity experiments suggest that closing these passages up to the level of the thermocline would cool the Indian Ocean about 1.51C and warm the Pacific Ocean about 0.51C. The model results also indicate that the influence of these passages on the production of NADW is very small. Denmark Strait

Gibraltar

Although it has been proposed that the saline outflow from the Mediterranean may play a critical role in the formation of NADW, ocean models do not support this idea. The influence of the Mediterranean outflow on the thermohaline circulation in the Atlantic has been investigated using modern boundary

The Denmark Strait, between Greenland and Iceland, is thought to play a critical role in the formation of NADW. The main body of cold saline intermediate water from the Norwegian–Greenland Sea flows into the North Atlantic through the Denmark Strait. In the North Atlantic it mixes with other waters to form NADW.



Experiments with a medium resolution OGCM indicated that the North Atlantic circulation is very sensitive to relatively small changes in the depth of the Denmark Strait. Implications for the Global Climate Systems

Changes in the magnitude of the thermohaline circulation of the scale suggested by the circulation experiments exploring changes in interocean gateways have profound implications for the global climate system. On a planetary scale, the ocean and atmosphere carry roughly equal amounts of energy poleward, but their relative importance varies with latitude. The ocean dominates by a factor of two at low latitudes, and the atmosphere dominates by a similar amount at high latitudes. The present poleward ocean heat transport is estimated to reach a maximum of about 3.5 1015 W at 251N and 2.7 1015 W at 251S. The subtropical convergences act as barriers to poleward heat transport by the ocean. Only where deep water forms at high latitudes are warm subtropical waters drawn across the frontal systems to higher latitudes to replace the sinking waters. If the entire thermohaline circulation were involved in the heat exchange between the polar and tropical regions, there would be an equatorward flow of 55 Sv warmed by about 151C as it returns to the surface through diffuse upwelling. This is equal to an energy transport of 3.5 1015 W, about half the total surface ocean transport at low latitudes. Clearly, the transport of heat by the thermohaline system is an important component of the ocean transport system. The great increases in production of polar bottom waters indicated by the closed Drake Passage models suggest that in earlier times ocean heat transport may have dominated the earth’s energy redistribution system.

Ocean Circulation on a Warm Earth During the Late Cretaceous and Eocene, the Earth was characterized by warm polar regions with mean annual temperatures well above freezing and perhaps as warm as 101C. A major controversy has arisen as to whether increased ocean heat transport played a role in creating these conditions. It has been suggested that there may have been a reversal of the thermohaline circulation, with sinking of warm saline waters in low latitudes and upwelling of warm waters in the polar regions. A secondary question has concerned the effect of the very different paleogeography, with the low latitude circumglobal

307

Tethys seaway connecting the Atlantic and Pacific Oceans, on ocean circulation. The Ocean Heat Transport Problem

The role of ocean heat transport in producing global warmth has been assessed by using the US National Center for Atmospheric Research (NCAR) Community Climate Model 1 (CCM1) AGCM with present-day geography, but with a swamp ocean. Swamp oceans were used in many early climate models; they have no heat storage capacity and do not transport heat, but serve as a source of moisture. Simulating a world without ocean currents, it was found that polar temperatures were 5–151C warmer than observed today. Tropical temperatures were about 21C less. An experiment with the Princeton model used idealized geography, a single ocean and a wedge-shaped continent extending from pole to pole for two AGCM experiments, one with and one without ocean heat transport. The experiment with ocean heat transport produced surface air temperatures poleward of 501 that were 201C warmer than in the experiment without ocean heat transport. Subsequently, the meridional ocean heat transport required to maintain polar ocean surface temperatures at 101C was determined. If the ocean were solely responsible, such warm polar temperatures would require implausible ocean heat transports of 2.3 1015 W across the Arctic Circle, implying water volume fluxes in excess of 100 Sv, or over twice the flow of the Gulf Stream today. In a discussion of the role of ocean heat transport in paleoclimatology it has been argued that theoretical arguments as well as simple models suggest that the total poleward transport of energy is not dependent on the structure of either the atmosphere or ocean. Changes in one transport mechanism tend to be compensated by a change in the opposite sense in the other, so that the total transport remains constant. Apparently only extreme changes in the geography of the polar regions can cause uncompensated changes in poleward heat transport. By using the NCAR CCM1 AGCM linked with a slab mixed-layer ocean model, three experiments reflecting different amounts of ocean heat transport have been conducted. Doubling the present flux and reducing the flux to zero resulted in global average temperatures of 14.4 and 15.91C, respectively compared with the present 151C. The major effect of changing the ocean heat flux through this range was to lower tropical temperatures 51C while making only a small increase in polar temperatures. By using NASA’s Goddard Institute of Space Science (GISS) AGCM with slab ocean to calculate the


308


ocean heat transport required to achieve specific sea surface temperatures it was concluded that a 46% increase in ocean heat transport was required to produce a planetary warming of 61C, which was assumed to be typical for the Jurassic, and that a 68% increase could produce a warming of 6.51C, assumed to be typical for the Cretaceous. The reduced sea ice formation and ice–albedo feedback in response to the increased poleward heat transport were considered critical factors.

(A)

Sites of deep water formation

Surface and Thermohaline Circulation

In 1972 the first attempt was made to simulate Late Cretaceous ocean circulation using an analog model. An experiment was carried out using a large waterfilled rotating dish set up with continental blocks in a Cretaceous configuration, with wind stress applied to the surface by fans, assuming a poleward displacement of the winds by 10–151. This analog model produced large anticyclonic gyres in the Pacific and east to west flow through the Tethys, a pattern widely reproduced in paleo-oceanographic studies (Figure 3). In the late 1980s an OGCM was used to perform experiments on ocean circulation with mid-Cretaceous paleogeography. Two of the models contrasted the effects of the different temperatures and winds generated by AGCMs assuming present day and four times present day atmospheric CO2 concentrations. The present-day CO2 experiment showed eastward flow across the Gulf of Mexico, westward flow through the Caribbean, a broad west to east current through the western Atlantic, and a clockwise anticyclonic gyre between Eurasia and Africa. Strong western boundary currents developed along the margins of Asia, Africa, and India. The current off north-east Asia continued into the Arctic to a high-latitude site of deep water formation. The thermohaline circulation indicated a strong polar source north of east Asia, and a weak source in the eastern Tethys. For the four-times present CO2 experiment surface currents in the tropics and subtropics were similar to those in the first experiment, but the flow of the western boundary current off north-east Asia no longer continued into the Arctic. The thermohaline circulation was completely different, lacking a polar source of deep water, but having a strong source in the eastern Tethys which produced a deep flow south and eastward into the Pacific basin. A simulation of Late Cretaceous ocean circulation using higher spatial resolution and driven by a later version of the GENESIS Earth System Model suggested that low latitude deep water formation may

(B)

(C)

Site of deep water formation

Figure 3 Three models of Late Cretaceous Ocean circulation. (A) After the analog model of Luyendyk et al. (1972) (B) After the OGCM of Barron et al. (1993). (C) After the OGCM of Brady et al. (1998). Note the different directions of flow through the Tethyan circumglobal seaway.

be more complex than had been thought. The model suggested four potential sites of warm saline dense water formation along the borders of the Tethys and off western South America. However, at three of these sites the water sank only to upper intermediate levels. However, at a site north of the equator on the west African margin the water sank first to an intermediate level, then spread south to the subtropical front where it sank further into the deep ocean interior and flowed out through the Indian Ocean into the Pacific.

Models of Seaways There have been some model simulations of circulation in shallow continental-scale ocean passages, such as the Cretaceous Western Interior Seaway of North America and the Miocene Tethys–Paratethys connection. These simulations are not based on general ocean models, but use circulation models



developed for study of estuaries and tides. They may include tidal currents and simulate flows resulting from salinity differences induced by runoff from land.

Verification of Ocean Models It is not easy to test the results of a paleoceanographic simulation against observations. The simplest tests have been to use palaeontologic and sedimentologic data to confirm or reject the direction of major current systems. Much geologic evidence suggests that flow through the Tethys was east to west, but some model simulations indicate that it was west to east, others that it was east to west. Several recent attempts have been made to compare oxygen isotopic data with the temperatures and salinities predicted by the paleo-oceanographic simulations, but there is uncertainty about how to correct for the isotopic differences due to salinity because of the complexity of the relation to the hydrologic cycle. The solution will ultimately lie in incorporating isotopic fractionation into the ocean models.

Conclusions The use of models to investigate the relations between ancient climates and ocean circulation is still in its infancy. There have been few comparisons of different models using the same boundary conditions. The results produced by different OGCMs for similar boundary conditions are often significantly different. In spite of the different model results it has become clear that the opening and closing of passages between the major ocean basins have had a major effect on the Earth’s climate. Attempts to simulate ocean circulation on a warm Earth indicate that sites of deep water formation may shift to low latitudes when atmospheric CO2 concentrations are high.

See also Abrupt Climate Change. Cenozoic Climate – Oxygen Isotope Evidence. Elemental Distribution: Overview. Heat Transport and Climate.

Further Reading Barron EJ, Peterson W, Thompson S, and Pollard D (1993) Past climate and the role of ocean heat transport: model simulations for the Cretaceous. Paleoceanography 8: 785--798.

309

Bice KL, Sloan LC, and Barron EJ (2000) Comparison of early Eocene isotopic paleotemperatures and threedimensional OGCM temperature field: the potential for use of model-derived surface water d18O. In: Huber BT, MacLeod KG, and Wing SL (eds.) Warm Climates in Earth History, pp. 79--131. Cambridge: Cambridge University Press. Brady EC, DeConto RM, and Thompson SL (1998) Deep water formation and poleward ocean heat transport in the warm climate extreme of the Cretaceous (80 Ma). Geophysical Research Letters 25: 4205--4208. Broecker WS, Bond G, and Klas M (1990) A salt oscillator in the glacial Atlantic? 1. The concept. Paleoceanography 5: 469--477. Bush ABG (1997) Numerical simulation of the Cretaceous Tethys circum global current. Science 275: 807--810. Covey C and Barron E (1988) The role of ocean heat transport in climatic change. Earth-Science Reviews 24: 429--445. Ericksen MC and Slingerland R (1990) Numerical simulation of tidal and wind-driven circulation in the Cretaceous interior seaway of North America. Geological Society of America Bulletin 102: 1499--1516. Fanning AF and Weaver AJ (1996) An atmospheric energymoisture balance model: climatology, interpentadal climate change and coupling to an ocean general circulation model. Journal of Geophysical Research D101: 15111--15128. Hay WW (1996) Tectonics and climate. Geologische Rundschau 85: 409--437. Kump LR and Slingerland RL (1999) Circulation and stratification of the early Turonian western interior seaway: sensitivity to a variety of forcings. In: Barrera E and Johnson CC (eds.) Evolution of the Cretaceous Ocean-Climate System, pp. 181--190. Boulder, co: Geological Society of America. Luyendyk B, Forsyth D, and Phillips J (1972) An experimental approach to the paleocirculation of ocean surface waters. Geological Society of America Bulletin 83: 2649--2664. Maier-Reimer E, Mikolajewicz U, and Crowley TJ (1990) Ocean general circulation model sensitivity experiment with an open Central American Isthmus. Paleoceanography 5: 349--366. Martel AT, Allen PA, and Slingerland R (1994) Use of tidal-circulation modeling in paleogeographical studies: an example from the Tertiary of the alpine perimeter. Geology 22: 925--928. Mikolajewicz U and Crowley TJ (1997) Response of a coupled ocean/energy balance model to restricted flow through the central American isthmus. Paleoceanography 12: 429--441. Mikolajewicz U, Maier-Reimer E, Crowley TJ, and Kim KY (1993) Effect of Drake and Panamanian gateways on the circulation of an ocean model. Paleoceanography 8: 409--426. Murdock TQ, Weaver AJ, and Fanning AF (1997) Paleoclimatic response of the closing of the Isthmus of Panama in a coupled ocean-atmosphere model. Geophysical Research Letters 24: 253--256.


310


Poulsen CJ, Barron EJ, Peterson WH, and Wilson PA (1999) A reinterpetation of mid-Cretaceous shallow marine temperatures through model-data comparison. Paleoceanography 14: 679--697. Rahmstorf S (1998) Influence of Mediterranean outflow on climate. EOS Transactions of the American Geophysical Union 79: 281--282. Rind D and Chandler M (1991) Increased ocean heat transports and warmer climate. Journal of Geophysical Research 96: 7437--7461. Roberts MJ and Wood RA (1997) Topographic sensitivity studies with a Bryan-Cox-type ocean model. Journal of Physical Oceanography 27: 823--836. Schmidt GA (1999) Forward modeling of carbonate proxy data from planktonic foraminifera using oxygen isotope

tracers in a global ocean model. Paleoceanography 14: 482--497. Toggweiler JR and Samuels B (1993) Is the magnitude of the deep outflow from the Atlantic Ocean actually governed by Southern Hemisphere winds? In: Heimann M (ed.) The Global Carbon Cycle, vol. 15, pp. 303-331. Berlin: Springer–Verlag. Toggweiler R and Samuels B (1998) Energizing the ocean’s large-scale circulation for climate change. Lisbon: 6th International Conference on Paleoceanography. Washington WM and Meehl GA (1983) General circulation model experiments on the climatic effects due to a doubling and quadrupling of carbon dioxide concentrations. Journal of Geophysical Research 88: 6600--6610.


PALEOCEANOGRAPHY: ORBITALLY TUNED TIMESCALES T. D. Herbert, Brown University, Providence, RI, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Geologists rely on a variety of ‘clocks’ built into sediments to place paleoenvironmental events into a time frame. These include radiometric decay systems, annual banding in trees, corals, and some marine and lake sediments, and, increasingly, the correlation of isotopic, geochemical, and paleontological variations to pacing supplied by changes in the Earth’s orbit. Variations in three parameters of the orbital system – eccentricity, obliquity, and precession – cause solar insolation to vary over the Earth as a function of latitude, season, and time, and hence cause global changes in climate. Because the timing of orbital changes can be calculated very precisely over the past 30 My, and because their general character can be deduced for much longer intervals of geological time, orbital variations provide a template by which paleoceanographers can fix paleoclimatic variations to geological time. Paleoceanographers now commonly assign either numerical ages or elapsed time to sediment records by optimizing the fit of sedimentary variations to a model of orbital forcing, a process referred to as ‘orbital tuning’. Although orbital tuning was first developed to create better timescales for studying the late Pleistocene (approximately the last 400 ky) Ice Ages, it now finds applications to dating sediments and estimating sedimentary fluxes at least into the Mesozoic Era (65–210 Ma). Orbital tuning came about because of the difficulty in assigning ages to long, continuous records of climate change that became available with ocean sediment coring. The simplest assumption for scaling time to stratigraphic in sediments – that sediment accumulates at a constant rate over long spans of time – is not likely to be true. Sediment compacts with burial, so that a layer 1-cm thick at 50-m burial depth represents significantly more time than the same centimeter of highly porous material at the top of the sediment column. Furthermore, we suspect that climate itself influences the rate of marine sediment accumulation over time, either by varying the production and preservation of biogenic components, or the supply of detrital materials from

land. Random factors such as small-scale erosion or excess deposition surely occur as well. Some of these variations in deposition rate can be documented with radiometric systems such as 14C and uranium disequilibrium series. Unfortunately, the radiocarbon clock cannot extend much past 4 104 years, and uranium series methods, which have more restrictive conditions to work well, extend to perhaps 3 105 years. Strata that contain minerals with radiometric systems that date the age of the sediment (e.g., 40Ar/39Ar and other systems in appropriate minerals) are few and far between. Other techniques such as paleomagnetic stratigraphy provide valuable age constraints, but at resolution of 105–106 years (the frequency of magnetic reversals). The end product of orbital tuning is a mapping function of stratigraphic position (depth scale in the sediment column) to time, using criteria discussed at more length below (Figure 1). Paleobiological and paleoceanographic patterns can be placed into a time frame perhaps accurate to a few thousand years (ky). Time-series studies benefit enormously, since removing the distorting effects of variations in sedimentation rate results in much ‘cleaner’ frequency spectra. Equally important, the large improvement in time resolution offered by orbital tuning in comparison to other stratigraphic techniques allows paleoceanographers to measure the dynamics of sedimentary records with far more precision than the constant sedimentation rate assumption. Did events recorded in different sedimentary locations occur simultaneously or with an age progression? How long did fundamental biological and geochemical turnovers in Earth history take to occur? What are the precise durations of magnetic polarity intervals? How do marine sediment fluxes vary with climatic state? Orbital tuning methods have addressed these and other questions over the last two decades.

Orbital Parameters and Potential Age Resolution by Orbital Tuning Gravitational interactions with the other planets and the Earth’s moon perturb the orbit of our planet in three modes that affect the solar radiation received by the Earth as a function of latitude and season. Variations in ‘eccentricity’ describe the quasiperiodic evolution of the Earth’s orbit around the Sun from more circular to more elliptical. Orbital


311

312

PALEOCEANOGRAPHY: ORBITALLY TUNED TIMESCALES

ODP 926B

Magnetic susceptibility

41-ky sine wave

10

5

Sediment data

0 375

380 385 390 Core depth (meters below seafloor)

395

Figure 1 Example of how a sedimentary signal (magnetic susceptibility, a carbonate proxy) recorded as a function of depth below seafloor can be ‘tuned’ to an orbital chronometer. In this case, Shackleton and colleagues tuned the record of early Miocene sedimentation (lower curve) to a 41-ky obliquity cycle (upper curve). The triangles indicate position of consecutive obliquity cycles identified at ODP Site 926B. Note that the spacing of cycles is not constant, an indication of variation in deposition rate. Note also the presence of smaller higher-frequency features caused by precession (mean period of about 21 ky). Absolute ages can be determined by fitting the amplitude envelopes of the obliquity and precessional signals measured in the sediments to similar features in long-term numerical calculations of the orbital terms. Data from Shackleton NJ, Crowhurst SJ, Weedon GP, and Laskar J (1999) Astronomical calibration of Oligocene–Miocene time. Philosophical Transactions of the Royal Society London, Series A 357: 1907–1929, courtesy of N. J. Shackleton.

eccentricity evolves in a complex manner, but has significant terms at about 95, 125, 405, 2000, and 2800 ky. The two shorter periods are close enough together to produce a beat pattern with a mean period of 109 ky, and the two longer terms produce a beat pattern at 2425 ky. Eccentricity is the only orbital cycle to alter the mean annual insolation, but its direct effect is very small. Axial ‘tilt’ (‘obliquity’) varies around its mean value of 23.51 by about 1.51 in a nearly sinusoidal pattern with a modern period of 41 ky. The obliquity cycle causes insolation to be distributed poleward during intervals of higher than average tilt, and equatorward during intervals of lower obliquity. Climatic ‘precession’ affects the summer–winter insolation contrast, and unlike the other two orbital parameters, acts in opposite sign between the hemispheres. The precessional index is modulated by eccentricity, so that its amplitude bears the imprint of the c. 109-, 405-, and 2425-ky eccentricity cycles mentioned above. Climatic precession has a mean modern period of 21.2 ky, but, because of the eccentricity amplitude modulation, it can range in repeat time from 14 to 28 ky. Its Fourier series representation has concentrations of variance at 1/23 and 1/19 ky 1. Long-term secular changes in the Earth–Moon and solar system influence the Earth’s orbital parameters

on timescales 41Ma. The observed slowing of Earth rotation rate due to tidal friction requires that the periods of obliquity and precessional cycles have gradually decreased as we move back in time. The predicted shortenings (relative to the Present) are less than 1% through the Cenozoic, but amount to several percent in the Mesozoic and more into the Paleozoic. One should also note that because the rate of tidal dissipation has probably not been constant over time, we do not have a good model for how obliquity and precessional periods have evolved over the entire course of the Earth–Moon system. A second significant uncertainty in using orbital variations as a time template comes from the weakly chaotic motion of the solar system. Any numerical solution of the orbital system will show an exponential dependence on initial conditions. There is therefore a limit in time beyond which one may not calculate the phases of the orbital cycles with confidence. This does not mean that one has no idea of the behavior of the orbital system, but rather that one has to rely on average statistical properties, or unusually stable elements of the orbit (the 405-ky eccentricity cycle is believed to be one such case) to perform orbital tuning in sediments older than about 30 Ma. The fundamental limit on the accuracy of orbital calculations thus imposes a twofold division of Earth



time amenable to orbital tuning. Sediments younger than 30 My can in theory be tuned to a precision and accuracy which depends solely on the match to an orbital template. In theory, the error of such an approach depends largely on the unknown time lag between orbital change and the climate change recorded by sedimentary proxies. Such response times could vary from nearly instantaneous (adjustments of ocean surface temperature and hydrology) to ky scale (growth and decay of large ice sheets, changes in the deep-ocean circulation, variations in greenhouse gas content). Orbital tuning in older sediments functions to measure ‘elapsed time’ in a sediment record, since one can no longer unequivocally associate a sedimentary feature with a precisely known time value of the Earth’s orbit. The latter approach resembles using a yardstick to measure distance from an independently agreed-upon datum. Datums in Earth history come from magnetic reversals, biostratigraphic, or paleochemical events. Elapsed time can be measured to the errors in the mean periods of the orbital series. Reasonable estimates of the uncertainties in the latter approach lie in the range of 5–20 ky. Examples of both numerical dating and the ‘yardstick’ modes of orbital tuning will follow below. One should note, however, that the presumed stability of the 405-ky eccentricity cycle over great stretches of geological time may allow geologists to develop a numerical time frame based on the Earth’s orbit, and therefore accurate to a fraction of the 405-ky period, for perhaps the past 100 My.

Methods of Orbital Tuning No single approach to orbital tuning is appropriate to all stratigraphic problems. The ideal case, producing an age–depth model tied precisely to an ‘absolute’ timescale, can only be achieved in cases where the stratigrapher has a global orbitally forced signal that can be correlated to a precise astronomical reference frame. This situation obtains from late Pleistocene (c. 4 Ma) onward, where significant changes in global ice volume altered the entire ocean inventory of oxygen isotopes on orbital timescales. Despite the general decline in the oxygen isotope signal in older marine strata, orbital tuners can find the imprint of orbital forcing in other aspects of sedimentation, such as variables like calcium carbonate content, redox state, and microfossil content. Such variations, because they originate from regional changes in the input of dust, biological productivity, and bottom water circulation, cannot provide a globally synchronous signal equal to that of isotopic

313

measures. They can, however, be correlated with confidence to the numerical timescale up to limit of the accuracy of orbital solutions, or about 30 Ma. In older strata, orbital tuning generally produces ‘floating’ timescales. One can choose among several approaches to matching sedimentary signals to an orbital forcing template. One can maximize the match of a sediment series to an orbital model by constructing an age– depth model that gives the largest correlation coefficient between presumed forcing and climatic response. An alternative is to work more selectively with the sediment record to extract the orbital components from the natural record, and to maximize the coherence of these to the corresponding orbital components. Such methods generally involve working in the frequency domain, using techniques such as band-pass filtering and complex demodulation. The most straightforward tactic relies on visual correlation of successive peaks in a sedimentary time series to a presumed orbital signal. A new timescale emerges as peaks and troughs of a stratigraphic signal are aligned to the orbital template. The alignment may be performed visually, or by designing an objective mapping function that optimizes the fit of stratigraphic signal to orbital template. Such timedomain tuning, while it offers the highest possible age resolution, is quite sensitive to the orbital model chosen and to noise in the geological series. The possibility clearly exists to produce a tuned sedimentary series that has been forced to resemble an orbital template by overenthusiastic correlation. Several strategies exist to lessen the subjectivity and heighten the reliability of time (depth) domain tuning. The first recognizes that higher-frequency orbital cycles such as precession and obliquity have lowfrequency ‘envelopes’ that cause their amplitude to vary systematically. The technique of complex demodulation detects such features in time series data, and the envelopes of ‘tuned’ stratigraphic cycles can be compared to those of the orbital cycles themselves. Erroneous tuning of individual 21- or 41-ky peaks and valleys in a stratigraphic series will be revealed by the misalignment of the envelopes of the geological cycles relative to orbital forcing. Moving-window (‘evolutive’) spectral analysis also works effectively to produce smooth tunings of stratigraphic signals to time. Here the analyst divides the data into segments of a specified stratigraphic length, and searches in the frequency domain for strong signals associated with orbital components. The average sedimentation rate is deduced by optimizing the match of scaling of stratigraphic frequency (cm 1) to orbital frequencies. By subdividing the data into relatively short (usually 300–500-ky


314


length) windows, the technique reduces spectral distortions associated with changes in sedimentation rate along section, and with the nonstationarity of climate response to orbital forcing. A conceptual example of how evolutive spectral analyses can detect and tune for gradual changes in sedimentation rate is illustrated in Figure 2. Moving-window analyses have the virtue of producing smooth estimates of sedimentation rate, since they rely on the average spectral properties of stratigraphic signals over a depth or time range. It is essential to remember that any ‘cyclostratigraphy’ carries an implicit climate model, and in many cases, an implicit sedimentation model. Our knowledge of the climate system is not adequate to

predict the precise response of the Earth’s climate orbital forcing over time; our correlation of paleoclimate records to orbital forcing must allow the recipe of orbital influences, which reflect changes over time in key climatic regions of the globe and in climatic feedbacks, to vary. Study of the relatively well-dated Pleistocene marine record (see Plio-Pleistocene Glacial Cycles and Milankovitch Variability) demonstrates that one orbital recipe may not describe the evolving Ice Ages, as the relative importance of c. 21-, 41-, and 100-ky ice volume cycles changes significantly over the past 2 My due to processes internal to the climate system. Orbital tuning should therefore include a healthy amount of flexibility and regard for independent checks on the quality of the result.

Criteria for Success Sedimentation rate drops (stratigraphic period drops)

Stratigraphic depth arbitrary units

Window

Signal Figure 2 Changes in sedimentation rate act as frequency modulations of cyclic signals. In this synthetic example, sedimentation rate was made to decrease by a factor of four from the youngest to oldest interval of the record. Note that the repeat distance in the stratigraphic record is greatest where sediment accumulation is highest, and tapers toward the base of the sequence where accumulation rate is low. Such changes can be traced objectively by the moving window spectral method, which produces sliding estimates of the local stratigraphic frequency (the inverse of stratigraphic period) down-section. In this case, the cycle frequency moves (top–base) from low to high as the cycles become more condensed at lower sedimentation rates.

Orbital tuning is rarely applied to sediments without first considering independent age constraints from fossil events and paleomagnetic reversals. These provide a preliminary age scale and therefore a guide to approximate, time-averaged, sedimentation rates to be modified by orbital tuning. Independent age markers also provide important tests of the quality of orbital tuning. For example, one can compare the results of orbital timescales to ‘known’ ages of events such as magnetic reversals where high-quality radiometric ages are available. Work conducted during the last decade suggests that orbital methods yield impressively consistent estimates of the ages of biostratigraphic and magnetostratigraphic datums. For example, Wilson demonstrated that Shackleton et al.’s revision of the Plio-Pleistocene geomagnetic polarity timescale (GPTS) based on orbital tuning yielded smoother, and therefore more plausible, spreading rate histories on a number of mid-ocean ridge systems than did the previous timescale based on K/Ar dates of basalts. Continuity of sedimentation is clearly another necessary condition for success of the orbital approach. The best sections have the following characteristics: they lack observable breaks in sedimentation, either erosional or depositional (e.g., turbidites), are composed of pelagic or hemipelagic facies, and do not have strong changes in overall sediment composition (e.g., major break from calcareous to detrital sedimentation) that tend to accompany strong changes in deposition rate. In the ideal case, sections will span millions of years of deposition and accumulate at a sufficient rate (empirically at 43 cm ky 1) to resolve precessional variations if such exist. Paleoceanographers have also learned to drill multiple offset holes at sites of interest



and to splice records across coring gaps in individual holes to create ‘composite sections’ that are verifiably complete. Although orbital tuning has often been criticized as circular reasoning, it has passed a number of tests in recent years. External checks exist in the form of independent datums such as magnetic polarity reversals and biostratigraphic events. A good orbital tuning solution should produce consistent estimates of the numerical age of the datum event or the duration between datum events at a number of sites. Such concordant results have been demonstrated at least into the middle Miocene, and new studies continue to extend continuous tuning further back in time. Another powerful test of an orbital tuning model comes from time series analysis. Orbital signals have quite distinctive ‘fingerprints’ due to amplitude modulations. The modulations are most pronounced in the case of the eccentricity and precessional cycles, but also exist in the more sinusoidal obliquity cycle. Spectral analyses should therefore detect a hierarchy of frequencies correctly corresponding to modulation terms of the central orbital frequencies in a successful tuning. For example, one would expect to recover the 95-, 125-, and 404-ky modulating terms in a time series tuned to a presumed precessional signal.

Some Examples A team of Dutch stratigraphers, paleontologists, and paleomagnetists has made significant progress in dating marine sedimentary successions in the Mediterranean region in the age range 0–12 Ma by astronomical tuning. The dates are particularly valuable to stratigraphers because many of the sections studied define classical substages of the Pleistocene, Pliocene, and Miocene epochs. Furthermore, many of the studied sections have good magnetic properties, permitting the Dutch team to propose an astronomical chronology for the GPTS. As magnetic polarity reversal boundaries constitute globally synchronous events recorded in many types of sediments and frozen into the ocean crust by the process of seafloor spreading, improving the GPTS has widespread implications. Hilgen and co-workers recognized orbital forcing by a grouping of sapropels (dark, organic-rich beds) into units of B100 and 400 ky by eccentricity modulation of precessional climate changes. Their resulting calibration of the GPTS yielded significantly greater ages for magnetic reversal boundaries than the previously accepted dates based on K/Ar radiometric age dating. After initial controversy, the ages

315

proposed by Hilgen and others have largely been verified by recent advances in 40Ar/39Ar dating of volcanic ash layers at a number of magnetic reversal boundaries. The power of orbital tuning to resolve sedimentation rates at high precision for paleoceanographic studies is nicely illustrated by Shackleton’s analysis of cyclic deep sea sediments recovered from the western equatorial Atlantic by Ocean Drilling Program Leg 154. Coring was designed to aquire long records at variable water depths. Depth largely controls carbonate sedimentation rate in the region, due to a direct pressure relationship (carbonate minerals are more soluble at greater pressures) and through the indirect effect of vertical variations in water mass carbon chemistry. Carbonate sedimentation in turn largely determines overall sediment accumulation. Figure 3 displays the results of tuning variations in magnetic susceptibility, a means of estimating the noncarbonate fraction of the sediment, to an obliquity (41 ky) pacing over an interval of early Miocene age. Sediment variations at Site 926C can be matched one-for-one to variations at drill Site 929A, approximately 750 m deeper in the water column (lower two curves, Figure 3). The orbital tuning also generates a sedimentation rate function at each site, displayed as the upper two curves in Figure 3. Higher sedimentation rates at Site 926C throughout the interval agree with expectations that carbonate dissolution should be less intense at the shallower location, while changes in the difference between the sites documents changing gradients in dissolution over time. It is important to note that conventional dating methods based on biostratigraphic or magnetic polarity stratigraphy would not have the resolving power to monitor the carbonate deep-water chemistry proxy at the resolution afforded by orbital tuning. Orbital chronology has also played a role in studying events across one of the truly catastrophic passages of geological time at the Cretaceous/ Tertiary boundary. While a number of observations suggest that the paleontological transition was very abrupt, it has proved difficult to constrain the timing of events before and after the boundary, and to differentiate rates of change across the K/T boundary quantitatively from background variability. Work by Herbert and others demonstrates that precessional cycles can be recognized at many deep sea sites that contain K/T boundary sequences. These provide a yardstick for dating events to within about 10 ky (one-half precessional wavelength) across the boundary. As one example of the information recovered, Figure 4 displays a composite of sedimentation rates in South Atlantic pelagic sites


316


Ceara Rise 40

30 929A (deep) 20

10 15

Sedimentation rate (m/M yr)

926C (shallow)

Magnetic susceptibility

0 926C 10

5

929A 0 20

20.2

20.4

20.6 Age (Ma)

20.8

21

21.2

Figure 3 High-resolution orbital tuning of sediment records from different water depths allows paleoceanographers to study variations on a common accurate time base; 41-ky variations driven by obliquity forcing were identified at both shallower (ODP 926B) and deeper (ODP 929A) sites (Shackleton et al., 1999). Tuning generates mapping functions of depth to time that reflect sedimentation rates at each site. Data from Shackleton NJ, Crowhurst SJ, Weedon GP, and Laskar J (1999) Astronomical calibration of Oligocene–Miocene time. Philosophical Transactions of the Royal Society London, Series A 357: 1907–1929, courtesy of N. J. Shackleton.

across the K/T boundary, using precessional cycles as an accumulation rate gauge. Data at each site were normalized to a value of 1 for the mean late Cretaceous accumulation rate. The step-function drop in sedimentation rate coincident with the K/T boundary reveals not only a catastrophic drop in the flux of biogenic material to the deep sea following the extinction event, but also a prolonged period of low flux in the early Tertiary as the planktonic ecosystem slowly recovered.

Where the Field is Headed Paleoceanographers constantly search for new tools to generate objective sediment data for spectral and stratigraphic analysis. Over the last decade, sediment proxy measurements such as color reflectance, magnetic susceptibility, gamma ray attenuation porosity evaluator, X-ray fluorescence elemental scanners, p-wave velocity, and downhole geophysical and geochemical logging have produced long, densely sampled time series from many sediment coring locations. Variations in the measured parameters all

reflect in some way changes in the sediment composition (generally the proportion of carbonate to noncarbonate sediment). The variables measured yield less insight into the mechanisms of climate change than data such as stable isotopic measurements and analyses of microfossil populations, but they have the virtue of being inexpensive, rapid, and nondestructive. At least three important conceptual targets remain to be solved by improvements in orbital tuning. The first relates directly to questions of the climate system itself. Paleoceanographers are becoming more aware of the climatic insights to be gained by comparing leads and lags in responses of various measured parameters. These can be ascertained by measuring multiple proxies in the same core, so that phase relations determined by cross-spectral analyses are independent of the absolute timescale, and between coring sites if one has a common, synchronous, tuning variable. Benthic d18O is generally assumed to provide such a chronological tie in late Pleistocene sediments. By measuring different components in the same cores, Clemens and colleagues demonstrate



317

South Atlantic composite K/ T boundary

Normalized sedimentation rate

1.4 1.2 1 0.8 0.6 0.4 0.2

64

64.5

65 Age (Ma)

65.5

66

Figure 4 Composite record of pelagic sedimentation rates in the South Atlantic across the Cretaceous/Tertiary (K/T) boundary, deduced from tuning variations in sediment carbonate content and color to c. 21-ky precessional cycles. Note the absence of a decline in sedimentation rate prior to the K/T boundary (lack of ‘precursor’ events) and the prolonged interval of dramatically reduced sedimentation rate following the extinction event. Adapted from Herbert TD and D’Hondt SL (1990) Precessional climate cyclicity in late Cretaceous–early Tertiary marine sediments: A high resolution chronometer of Cretaceous–Tertiary boundary events. Earth and Planetary Science Letters 99: 263–275.

that the relative timing of different aspects of the monsoon system in the Indian Ocean clearly evolves over the past 6 Ma. While each component senses orbital forcing, the linkages between winds, ice volume, and sea surface temperature change have varied over time. As long, orbitally tuned records accumulate, we can anticipate that more instances of significant and unusual biotic, climatic, and geochemical events in Earth history coincide with unusual combinations of orbital forcing. We think of orbital pacing of climate acting on characteristic timescales of 104–105 years, but much longer-scale effects may arise from lowfrequency modulations of individual cycles, or from complex patterns of constructive and destructive interference that may operate on a timescale of 106–107 years. As one example, the marine extinctions that define the Miocene/Oligocene boundary, and the pronounced d18O excursion that defines a major long-lived glacial period at that time, have been convincingly tied by Zachos and colleagues to an unusual coincidence of very low obliquity variations and low orbital eccentricity. There is also a persistent indication that anomalies in the carbon cycle, as exhibited by isotopes of carbon locked into marine carbonates, may follow (and amplify) longwavelength components of orbital forcing. Finally, one can anticipate that the improved chronology afforded by orbital tuning will offer large refinements in the geological timescale for the Cenozoic and Mesozoic. Improvements in estimating the duration of magnetic polarity chrons will allow

geophysicists to study variations in ocean spreading rate at least into the Cretaceous without the constraint of having to specify a constant-spreading-rate ridge system. Members of the radiometric dating community have proposed that an intercalibration of radiometric and orbital dating systems may be the most accurate way to choose age standards for the 40 Ar/39Ar dating system. The possibility that orbital signals preserved in marine sediments will supply celestial mechanicists with constraints on the longterm evolution of the Earth–Sun–Moon system also appears on the horizon.

See also Calcium Carbonates. Deep-Sea Drilling Methodology. Paleoceanography, Climate Models in. PlioPleistocene Glacial Cycles and Milankovitch Variability.

Further Reading Billups K, Palike H, Channell JET, Zachos JC, and Shackleton NJ (2004) Astronomic calibration of the late Oligocene through early Miocene geomagnetic polarity time scale. Earth and Planetary Science Letters 224: 33--44. Clemens SC, Murray DW, and Prell WL (1996) Nonstationary phase of the Plio-Pleistocene Asian monsoon. Science 274: 943--948.


318


Herbert TD and D’Hondt SL (1990) Precessional climate cyclicity in late Cretaceous–early Tertiary marine sediments: A high resolution chronometer of Cretaceous– Tertiary boundary events. Earth and Planetary Science Letters 99: 263--275. Hilgen FJ (1991) Astronomical calibration of Gauss to Matuyama sapropels in the Mediterranean and implication for the geomagnetic polarity time scale. Earth and Planetary Science Letters 104: 226--244. Husing SK, Hilgen FJ, Aziz HA, et al. (2007) Completing the Neogene geological time scale between 8.5 and 12.5 Ma. Earth and Planetary Science Letters 253: 340--358. Huybers P and Wunsch C (2004) A depth-derived Pleistocene age model: Uncertainty estimates, sedimentation variability, and nonlinear climate change. Paleoceanography 19(1): PA1028. Laskar J, Robutel P, Joutel F, Gastineau M, Correia ACM, and Levrard B (2004) A long-term numerical solution for the insolation quantities of the Earth. Astronomy and Astrophysics 428: 261--285. Lisiecki LE and Raymo ME (2005) A Pliocene–Pleistocene stack of 57 globally distributed benthic delta O-18 records. Paleoceanography 20: PA1003. Lourens LJ, Sluijs A, Kroon D, et al. (2005) Astronomical pacing of late Palaeocene to early Eocene global warming events. Nature 435(7045): 1083--1087.

Martinson DG, Pisias NG, Hays JD, et al. (1987) Age dating and the orbital theory of the ice ages: Development of a high-resolution 0 to 300 000-year stratigraphy. Quaternary Research 27: 1--29. Raffi I, Backman J, Fornaciari E, et al. (2006) A review of calcareous nannofossil astrobiochronology encompassing the past 25 million years. Quaternary Science Reviews 25: 3113--3137. Renne PR, Deino AL, Walter RC, et al. (1994) Intercalibration of astronomical and radioisotopic time. Geology 22: 783--786. Shackleton NJ, Crowhurst SJ, Weedon GP, and Laskar J (1999) Astronomical calibration of Oligocene–Miocene time. Philosophical Transactions of the Royal Society London, Series A 357: 1907--1929. Wade BS and Pa¨like H (2004) Oligocene climate dynamics. Paleoceanography 19: PA4019 (doi:10.1029/ 2004PA001042). Wilson DS (1993) Confirmation of the astronomical calibration of the magnetic polarity timescale from seafloor spreading rates. Nature 364: 788--790. Zachos JC, Shackleton NJ, Revenaugh JS, et al. (2001) Climate response to orbital forcing across the Oligocene– Miocene boundary. Science 292: 274--278.


PALEOCEANOGRAPHY: THE GREENHOUSE WORLD M. Huber, Purdue University, West Lafayette, IN, USA E. Thomas, Yale University, New Haven, CT, USA

are the focus of most research. These three key features are:

& 2009 Elsevier Ltd. All rights reserved.

• •

Introduction Understanding and modeling a world very different from that of today, for example, a much warmer world with a so-called ‘greenhouse’ climate, requires a thorough grasp of such processes as the carbon cycle, ocean circulation, and heat transport in the oceans – indeed it pushes the limits of our current conceptual and numerical models. A review of what we know of past greenhouse worlds reveals that our capacity to predict and understand key features of such intervals is still limited. Greenhouse climates represent most of the past 540 My, called the Phanerozoic, the part of Earth history for which an extensive fossil record is available. One could thus argue that a climate state with temperatures much warmer than modern, without substantial ice at sea level at one or both Poles, is the ‘normal’ mode, and the glaciated, cool state – such as has existed over the past several million years – is unusual. Since we were born into this glaciated climate state, however, our theories of near-modern climates are well developed and powerful, whereas greenhouse climates with their different boundary conditions challenge our understanding. The two best-documented periods with greenhouse climates are the Cretaceous (B145–65.5 Ma) and the Eocene (B55.8–33.9 Ma), on which this article will concentrate. But, other intervals, both earlier (e.g., Silurian, 443.7–416 Ma) and more recent (e.g., early–middle Miocene, 23.03–11.61 Ma, mid-Pliocene, 3.5 Ma) are characterized by climates clearly warmer than today, although less torrid than the Eocene or Cretaceous. Furthermore, within longterm greenhouse intervals, there may lurk short-term periods of an icier state. The alien first impression one derives from looking at the paleoclimate records of past greenhouse climates is the combination of polar temperatures too high for ice formation and warm winters within continental interiors at mid-latitudes. In fact, there are three main characteristics of past greenhouse climates that are remarkable and puzzling, and hence

•

Global warmth: a global mean surface temperatures much warmer (410 1C) than the modern global mean temperature (15 1C); Equable climates: reduced seasonality in continental interiors compared to modern, with winter temperatures above freezing; and Low temperature gradients: a significant reduction in equator-to-pole and vertical ocean temperature gradients (i.e., to o20 1C). Annual mean surface air temperatures today in the Arctic Ocean are about 151C, in the Antarctic about 501C, whereas tropical temperatures are on average 26 1C, hence the modern gradient is 40–75 1C.

Not all greenhouse climates are the same: some have one or two of these key features, and only a few have all the three. For example, there is ample and unequivocal evidence for global warmth and equable climates throughout large portions of the Mesozoic (251–65.5 Ma) and the Paleogene (65.5–23.03 Ma) (Figure 1). But, evidence for reduced meridional and vertical ocean temperature gradients is much more controversial, however, and (largely due to the nature of the rock and fossil records) only reasonably well established for the mid-Cretaceous and the early Eocene (see Cenozoic Climate – Oxygen Isotope Evidence and Paleoceanography). Superimposed on these fundamental aspects of greenhouse climate modes is a pervasive variability apparently driven by orbital variations (see PlioPleistocene Glacial Cycles and Milankovitch Variability), and abrupt climate shifts (see Abrupt Climate Change), indicative either of crossing of thresholds or of sudden changes in climatic forcing factors, and notable changes in climate due to shifting paleogeographies, paleotopography, and ocean circulation changes (see Heat Transport and Climate). A variety of paleoclimate proxies, including oxygen isotopic paleotemperature estimates (see Cenozoic Climate – Oxygen Isotope Evidence), as well as newer proxies such as Mg/Ca (see Determination of Past Sea Surface Temperatures), and ones that might still be considered experimental, such as TEX86 (see Paleoceanography), have been used to reconstruct greenhouse climates. New proxies and long proxy records of atmospheric carbon dioxide concentrations (pCO2)


319

320

Atmospheric CO2 (ppm)

(a)

PALEOCEANOGRAPHY: THE GREENHOUSE WORLD

8000 GEOCARB III Proxies

6000

4000

2000

0

Continental glaciation (deg paleolatitude)

(b)

20 30 40 50 60 70 80 90 600

500

400 300 Time (Ma)

200

100

0

Figure 1 (a) Atmospheric carbon dioxide concentrations produced by geochemical models and proxies for the Phanerozoic. (b) Major intervals of continental glaciation and the latitude to which they extend. The major periods in Earth’s history with little continental ice correspond to those periods with high greenhouse gas concentrations at this gross level of comparison. From Royer DL, Berner RA, Montanez IP, Tabor NJ, and Beerling DJ (2004) CO2 as a primary driver of Phanerozoic climate. GSA Today 14: 4–12.

are becoming available and allow us to constrain the potential sensitivity of climate change to greenhouse gas forcing (Figure 1). The investigation of past greenhouse climates is making exciting and unprecedented progress, driven by massive innovations in multiproxy paleoclimate and paleoenvironmental reconstruction techniques in both the marine and terrestrial realms, and by significant developments in paleoclimate modeling (see Paleoceanography and Paleoceanography, Climate Models in).

Cretaceous The Cretaceous as a whole was a greenhouse world (Figure 1): most temperature records indicate high latitude and deep ocean warmth, and pCO2 was high. Nevertheless, substantial multiproxy evidence indicates the presence of, perhaps short-lived, below-freezing conditions at high latitudes and in continental interiors, and the apparent buildup of moderate terrestrial ice sheets (potentially in the

early Cretaceous Aptian, 125–112 Ma, and in the late Cretaceous Maastrichtian, 70.6–65.5 Ma). Peak Cretaceous warmth and the best example of a greenhouse climate lies in the mid-Cretaceous (B120–80 Ma), when tropical temperatures were between 30 and 35 1C, mean annual polar temperatures above 14 1C, and deep ocean temperatures were c. 12 1C. Continental interior temperatures were probably above freezing year-round, and terrestrial ice at sea level was probably absent. Global mean surface temperatures were much more than 10 1C above modern, and the equator-to-pole temperature gradient was approximately 20 1C. It is not clear exactly when the thermal maximum occurred during this overall warm long period (e.g., in the Cenomanian, 99.6–93.5 Ma, or in the Turonian, 93.5–89.3 Ma), because regional factors such as paleogeography and ocean heat transport might substantially alter the expression of global warmth (Figure 2), and a global thermal maximum does not necessarily denote a global maximum everywhere at the same time. A feature of particular interest in the Cretaceous (and to a lesser extent in the Jurassic, 199.6–145.5 Ma) oceans are the large global carbon cycle perturbations called ‘oceanic anoxic events’ (OAEs). These were periods of high carbon burial that led to drawdown of atmospheric carbon dioxide, lowering of bottom-water oxygen concentrations, and, in many cases, significant biological extinction. Most OAEs may have been caused by high productivity and export of organic carbon from surface waters, which has then preserved in the organic-rich sediments known as black shales. At least two Cretaceous OAEs are probably global, and indicative of oceanwide anoxia at least at intermediate water depths, the Selli (late early Aptian, B120 Ma) and Bonarelli events (Cenomanian–Turonian, B93.5 Ma). During these events, global sea surface temperatures (SSTs) were extremely high, with equatorial temperatures of 32–36 1C, polar temperatures in excess of 20 1C. Multiple hypotheses exist for explaining OAEs, and different events may bear the imprint of one mechanism more than another, but one factor may have been particularly important. Increased volcanic and hydrothermal activity, perhaps associated with large igneous provinces (LIPs) (see Igneous Provinces), changed increased atmospheric pCO2, and altered ocean chemistry in ways to promote upper ocean export productivity leading to oxygen depletion in the deeper ocean. These changes may have been exacerbated by changes in ocean circulation and increased thermohaline stratification which might have affected benthic oxygenation, but the direction and magnitude of this



(a) Temperature difference

4 3 2 1 0 C −1 −2

321

continued to evolve, because the Cretaceous ended with a bang – the bolide which hit the Yucatan Peninsula at 65.5 Ma profoundly perturbed the ocean and terrestrial ecosystems (see Paleoceanography). Much of the evolution of greenhouse climates is related to the carbon cycle and hence to biology and ecology, so it is difficult to know whether the processes that maintained warmth in the Cretaceous with such different biota are the same as in the Cenozoic.

−3 −4

(b) Salinity difference

3 2 1 0 psu −1 −2 −3

Figure 2 In these coupled ocean–atmospheric climate model results from Poulsen et al. (2003), major changes in protoAtlantic Ocean temperature and salinity occur by deepening of the gateway between the North and South Atlantic. This study concluded that part of the pattern of warming, anoxia, and collapse of reefs in the mid-Cretaceous organic-rich sediments known as black might have been driven by ocean circulation changes engendered by rifting of the Atlantic basin.

potential feedback are poorly constrained. The emplacement of LIPs was sporadic and rapid on geological timescales, but in toto almost 3 times as much oceanic crust was produced in the Cretaceous (in LIPs and spreading centers) as in any comparable period. The resulting fluxes of greenhouse gases and other chemical constituents, including nutrients, were unusually large compared to the remainder of the fossil record. The direct impact of this volcanism on the ocean circulation through changes in the geothermal heat flux was probably small as compared to the ocean’s large-scale circulation, but perturbations to the circulation within isolated abyssal basins might have been important. As the Cretaceous came to a close, climate fluctuated substantially (Figure 3), but it is impossible to know how Cretaceous climate would have

Early Paleogene After the asteroid impact and subsequent mass extinction of many groups of surface-dwelling oceanic life-forms, the world may have cooled for a few millennia, but in the Paleocene (65.5–55.8 Ma) conditions were generally much warmer than modern, although evidence for at least intervals of deep sea and polar temperatures cooler than peak Cretaceous or Eocene warmth exists. Overall, in the first 10 My after the asteroid impact, ecosystems recovered while the world followed a warming trend, reaching maximum temperatures between B56 and B50 Ma (latest Paleocene to early Eocene). Within the overall greenhouse climate of the Paleocene to early Eocene lies a profoundly important, but still perplexing, abrupt climatic maximum event, the Paleocene–Eocene Thermal Maximum (PETM). The record of this time period is characterized by global negative anomalies in oxygen and carbon isotope values in surface and bottomdwelling foraminifera and bulk carbonate. The PETM may have started in fewer than 500 years, with a recovery over B170 ky. The negative carbon isotope excursion (CIE) was at least 2.5–3.5% in deep oceanic records and 5–6% in terrestrial and shallow marine records (Figure 4). These joint isotope anomalies, backed by independent temperature proxy records, indicate that rapid emission of isotopically light carbon caused severe greenhouse warming, analogous to modern anthropogenic fossilfuel burning. During the PETM, temperatures increased by 5–81C in southern high-latitude sea surface waters; c. 4–51C in the deep sea, equatorial surface waters, and the Arctic Ocean; and c. 51C on land at mid-latitudes in continental interiors. There is some indication of an increase in the vigor of the hydrologic cycle (Figure 4), but the true timing and spatial dependence of the change in hydrology are not clear. Diversity and distribution of surface marine and terrestrial biota shifted, with migration of thermophilic biota to high latitudes and evolutionary


322


(a)

(b)

(c)

65.4 65.5

P K

65.6

Ma

65.7 65.8

C29r

65.9

C30n

66.0 66.1 66.2 66.3 P. elegans datum Benthics A. australia G. prairiehillensis H. globulosa

P. elegans acme

66.4

Bass River

66.5

Site 525

66.6

Site 1050

Leaves, best bins Leaves, rangethrough

66.7 8 10 12 14 16 18 20 22 24 26 28 6

8 10 12 14 Estimated temperature (°C)

6 8 10 12 14 16 18 20 22 24 26 28

Figure 3 A summary of climate change records near the end of the Cretaceous (Maastrichtian). Paleotemperatures estimated from oxygen isotope data from benthic (filled symbols) and planktonic (open symbols) foraminiferea from middle (a) and high (b) latitudes. (c) Combined representative data (with leaf data) from (a) and (b) with terrestrial paleotemperature estimates derived from macrofloral records for North Dakota. The end of Cretaceous was clearly a time of very variable climate. From Wilf P, Johnson KR, and Huber BT (2003) Correlated terrestrial and marine evidence for global climate changes before mass extinction at the Cretaceous–Paleogene boundary. Proceedings of the National Academy of Sciences of the United States of America 100: 599–604.

378

Apectodinium (% of dinocysts)

, TEX86 (°C)

Terrestrial palynornorphs (%)

Isorenieratane (μg g−1)

28X

Hole 4A core

A. aug

382

30X

380

31X

386

A. aug

390

32X

Depth (m.c.d)

384

−32 −30 −28 −26 −240 δ13CTOC (%% vs. PDB)

10

20

30 0 20 40 60 80 15 17 19 21 230 20 40 60 80 0 20 40 60 80 0.0 0.2 0.4 0.6 0 BIT Index Anglosperms (% of Low-salinity-tolerant terrestrial palynomorphs) dinocysts (% of dinocysts)

1

2

Figure 4 IODP Site 302 enabled the recovery of PETM palynological and geochemical records from the Arctic Ocean. The carbon isotope records the PETM negative excursion (far left). The dinocyst Apectodinium (second from left) is a warm-water species, and a nearly global indicator of the PETM. TEX86 is a new paleotemperature proxy (middle) which reveals extreme polar warmth even previous to the PETM, and a 5 1C warming with the event. Other indicators show an apparent decrease in salinity in the Arctic ocean. From Sluijs A, Schouten S, Pagani M, et al. (2006) Subtropical Arctic Ocean conditions during the Palaeocene Eocene thermal maximum. Nature 441: 610–613 (doi:10.1038/nature/04668).



include a large range of options: organic matter heated by igneous intrusions in the North Atlantic, by subduction in Alaska, or by continental collision in the Himalayas; peat burning; oxidation of organic matter after desiccation of inland seas and methane release from extensive marshlands; and mantle plume-induced lithospheric gas explosions. The interval of B52–49 Ma is known as the early Eocene Climatic Optimum (EECO; Figure 5), during which temperatures reached levels unparalleled in the Cenozoic (the last 65.5 My) with the brief exception of the PETM. Crocodiles, tapir-like mammals, and palm trees flourished around an Arctic Ocean with warm, sometimes brackish surface waters. Temperatures did not reach freezing even in continental interiors at midto high latitudes, polar surface temperatures were 430 1C warmer than modern, global deep water temperatures were B10–12 1C warmer than today, and polar ice sheets probably did not reach sea level – if they existed at all. Interestingly, EECO tropical ocean temperatures, once thought to be at modern or even cooler values, are now considered to have been B8 1C warmer than today, still much less than the extreme polar warming. We do not know how the high latitudes were kept as warm as 15–23 1C with tropical temperatures only at B35 1C (Figure 6). The associated small latitudinal temperature gradients make it difficult to explain either high heat transport through the atmosphere or

turnover, while deep-sea benthic foraminifera suffered extinction (30–50% of species). There was widespread oceanic carbonate dissolution: the calcium carbonate compensation depth (CCD) rose by more than 2 km in the southeastern Atlantic. One explanation for the CIE is the release of B2000–2500 Gt of isotopically light (B 60%) carbon from methane clathrates in oceanic reservoirs (see Methane Hydrates and Climatic Effects). Oxidation of methane in the oceans would have stripped oxygen from the deep waters, leading to hypoxia, and the shallowing of the CCD, leading to a widespread dissolution of carbonates. Proposed triggers of gas hydrate dissociation include warming of the oceans by a change in oceanic circulation, continental slope failure, sea level lowering, explosive Caribbean volcanism, or North Atlantic basaltic volcanism. Arguments against gas hydrate dissociation as the cause of the PETM include low estimates (500–3000 Gt C) for the size of the oceanic gas hydrate reservoir in the recent oceans, implying even smaller ones in the warm Paleocene oceans. The observed 42 km rise in the CCD is more than that estimated for a release of 2000–2500 Gt C. In addition, pre-PETM atmospheric pCO2 levels of greater than 1000 ppm require much larger amounts than 2500 Gt C to raise global temperatures by 5 1C at estimated climate sensitivities of 1.5–4.5 1C for a doubling of CO2. Alternate sources of carbon

14

4

3

18O (% %) 2 1 0

−1

−2

Age (Ma)

Oligocene

30 40

Eocene

50 60

Palaeocene Cretaceous Ice sheets 0 4 ‘Ice free’ w = −1.2 %%

2

Small ice sheets?

Miocene

20

Growth−decay; large ice sheets

Plio-Pleistocene

10

70

1.0

Ice-free?

0

0

Mg temperature (°C) 2 4 6 8 10 12

323

Possibly ice free 8 “Modern” w = −0 .28 %% 6

10

14

Temperature (°C)

Calculated w (%%) 0.5 0 −0.5 −1.0 −1.5

Major ice sheet expansion Major ice sheet expansion Ice sheet decay Major ice sheet expansion Small ice sheets 48−49 Ma assumed ice-free for Mg calibratlon

75%

50%

25%

0%

Percent full pleistocene ice volume

Figure 5 Lear et al. (2000) utilized Mg/Ca ratios from benthic foraminifera to create a paleotemperature proxy record (left). When compared against the benthic oxygen isotope record (middle) there is a general congruence in pattern which is expected, given that a large part of the oxygen isotope record also reflects temperature change. When taken in combination, the two records can be used to infer changes in the oxygen isotopic composition (right) of seawater, a proxy for terrestrial ice volume. The major ice buildup at the beginning of the Oligocene (Oi1) is apparent in the substantial change in oxygen isotopic composition in the absence of major Mg/Ca change (solid horizontal line).


324


35

Late

Olig.

50 55 60

Eocene Middle Early

45

Paleocene E. Late

Age (Ma)

40

Modern SST range 5

10

20 15 Temperature (°C)

25

30

35

Figure 6 Pearson et al. (2007) paleotemperature estimates from Tanzania (red) derived from oxygen isotope ratios of planktonic foraminiferal tests and from TEX86 (yellow). The other data points show benthic and polar surface paleotemperature estimates. The Tanzanian data reveal tropical temperatures were significantly warmer than modern but that they were stable compared to benthic and polar temperatures through the Cenozoic climate deterioration. The authors’ interpretation is that previous tropical temperature estimates (filled blue and green circles) were spuriously tracking deep ocean temperatures (open blue and green circles) and temperature trends because of diagenesis. If correct, this implies that tropical temperatures were at least 5 1C warmer than previously thought in greenhouse intervals.

40

through the ocean given the well-proven conceptual and numerical models of climate (Figure 7). Despite more than two decades of work on this paradox of high polar temperatures and low heat transport, climate models consistently compute temperatures for high latitudes and mid-latitude continental winters that are lower than those indicated by biotic and chemical temperature proxies. Even this paradigm of greenhouse climate displayed profound variability. Within the EECO are alternating warm and very warm (hyperthermal) periods. These hypothermals are characterized by dissolution horizons associated with isotope anomalies and benthic foraminiferal assemblage changes and have been identified in upper Paleocene–lower Eocene sediments worldwide, reflecting events similar in nature to, but less severe than, the PETM. It remains an open question whether these hyperthermals directly reflect greenhouse gas inputs, or cumulative effects of changing ocean chemistry and circulation, perhaps driven by orbital forcing. While pCO2 levels may have been high (1000–4000 ppm) to maintain the rise of the global mean temperatures of this greenhouse world, it does not simultaneously explain the small temperature gradients (as low as 15 1C), warm poles, and warm continental interiors. Furthermore, we are still far from explaining the amplitude of orbital, suborbital, and nonorbital variability exhibited by these climates. The mystery is compounded by the rapidity by which the Eocene greenhouse world came to an end with the

30

20

10

0

−10

−20 90S

60S

30S

0

30N

60N

90N

Figure 7 As described in Shellito et al. (2003), model-predicted zonally averaged mean annual temperatures compared with paleotemperature proxy records for the Eocene. These simulations were carried out with Eocene boundary conditions and with pCO2 specified at three levels: 500 (solid), 1000 (dotted), and 2000 (dashed lines) ppm. The other symbols represent terrestrial and marine surface temperature proxy records. With increasing greenhouse gas concentrations, temperatures gradually come into better agreement with proxies at high latitudes, but the mismatch persists until extremely high values are reached, at which point tropical temperatures are too warm.



sudden emplacement of a massive Antarctic ice sheet near the epoch’s end, but the transition provides critical clues to understanding the functioning of a greenhouse climate.

The End of the Greenhouse We do not fully understand the greenhouse world, and neither do we understand the transition into the present cool world (‘icehouse world’) and the cause of global cooling. This much we think we do know. Beginning in the middle Eocene (at B48.6 Ma), the Poles and deep oceans began to cool, and pCO2 may have declined. The cause of this cooling and apparent carbon draw-down are unknown. As this cooling continued, the diversity of planktic and benthic oceanic life-forms declined in the late Eocene to the earliest Oligocene (B37.2–33.9 Ma). Antarctic ice sheets achieved significant volume and reached sea level by about 33.9 Ma, while sea ice might have covered parts of the Arctic Ocean by that time. Small, wet-based ice sheets may have existed through the late Eocene, but a rapid (B100 ky) increase in 18 benthic foraminiferal d O values in the earliest Oligocene (called the ‘Oi1 event’) has been interpreted as reflecting the establishment of the Antarctic ice sheet. Oxygen isotope records, however, reflect a combination of changes in ice volume and temperature (see Cenozoic Climate – Oxygen Isotope Evidence), and it is still debated whether the Oi1 event was primarily due to an increase in ice volume or to cooling. Paleotemperature proxy data (Mg/Ca) suggest that the event was due mainly to ice volume increase, although this record might have been affected by changes in position of the CCD in the oceans at the time. Ice sheet modelers argue that the event contains both an ice volume and a cooling component, and several planktonic microfossil records suggest surface water cooling at the time. Whatever the Oi1 event was in detail, it is generally seen as the time of change from a largely unglaciated to a largely glaciated world, and the end of the last greenhouse world. There are several proposed causes of this final transition to a world with substantial glaciation, including Southern Ocean gateway opening, decreases in greenhouse gas concentrations, orbital configuration changes, and ice albedo feedbacks. The Antarctic continent had been in a polar position for tens of millions of years before glaciation started, so why did a continental ice sheet form in B100 Ky during the Oi1? A popular theory is ‘thermal isolation’ of the Antarctic continent: opening of the Tasman Gateway and Drake Passage

325

triggered the initiation of the Antarctic Circumpolar Current (ACC), reducing meridional heat transport to Antarctica by isolation of the continent within a ring of cold water. We cannot constrain the validity of this hypothesis by data on the opening of Drake Passage, because even recent estimates range from middle Eocene to early Miocene (a range of 20 My). The Tasman Gateway opened several million years before Oi1 which seems to make a close connection between Southern Ocean gateway changes and glaciation unsupportable. Data on microfossil distribution indicate that there probably was no warm current flowing southward along eastern Australia, because a counterclockwise gyre in the southern Pacific prevented warm waters from reaching Antarctica. Furthermore, ocean–atmosphere climate modeling indicates that the change in meridional heat transport associated with ACC onset was insignificant at high latitudes. That does not mean that Southern Ocean gateway changes had no impact on climate. The oceanographic changes associated with opening of Drake and Tasman gateways (the socalled ‘Drake Passage effect’) has been reproduced by climate model investigations of this time interval, and the magnitude of the change anticipated from idealized model results has been reproduced. Opening of Drake and Tasman gateways produces a 0.6 PW decrease in southward heat transport in the southern subtropical gyre (out of a peak of B1.5 PW), and shifts that heat into the Northern Hemisphere. The climatic effect of this shift is felt primarily as a small temperature change in the subtropical oceans and a nearly negligible (o3 1C) change in the Antarctic polar ocean region. Climate and glaciological modeling implicates a leading role of this transition to decreases in atmospheric greenhouse gas concentrations, because Antarctic climate and ice sheets have been shown to be more sensitive to pCO2 than to other parameters. Consequently, Cenozoic cooling of Antarctica is no longer generally accepted as having been primarily caused by changes in oceanic circulation only. Instead, decreasing CO2 levels with subsequent processes such as ice albedo and weathering feedbacks (possibly modulated by orbital variations) are seen as significant long-term climate-forcing factors. The difficulty with this explanation is that existing pCO2 reconstructions do not show a decrease near the Oi1 event; instead, the primary decreases were much earlier or much later. This suggests that our understanding remains incomplete. Either pCO2 proxies or climate models could be incorrect, or, more interestingly, they may not contain key feedbacks and processes that might have transformed the long-term greenhouse gas decline into a sudden ice


326


buildup. It is likely, both from a proxy and modeling perspective, that changes in Earth’s orbital configuration played a role in setting the exact timing of major Antarctic ice accumulation, but this area remains a subject of active research.

Problems Posed by the Greenhouse Climate A strong line of geologic evidence links higher greenhouse gas concentrations with low temperature gradients and warm global mean and deep ocean temperatures through the Cretaceous and early Paleogene. This correlation is fairly strong, but not as strong, for equable continental climates. The concentrations necessary to sufficiently warm high latitudes and continental interiors, however, are so high that keeping tropical SSTs relatively cool becomes a problem. Less obvious and more successful mechanisms have been put forward to solve this problem, such as forcing by polar stratospheric clouds and tropical cyclone-induced ocean mixing. So far, all attempts to accurately reproduce warm Eocene continental interior temperatures lead to unrealistically hot tropical temperatures in view of biotic and geochemical proxy estimates (see Determination of Past Sea Surface Temperatures). As summarized in Table 1 and discussed below, many opportunities exist for continued progress to solve these problems. Cretaceous and early Paleogene tropical SST reconstructions might be biased toward cool values, and other variables need to be constrained to more precisely define the data–model mismatch in the Tropics. For instance, if planktonic foraminifera record tropical temperatures from below the mixedlayer, that is greater than c. 40 m depth, rather than 6 m, the low-gradient problem is ameliorated, but such a habitat may be difficult to reconcile with evidence for the presence of photosymbionts. Alternatively, a sampling bias toward warm season values at high latitudes and cold (upwelling) season temperatures in low latitudes could partially resolve the low-gradient problem. Warm terrestrial extratropical winter temperatures tend to rule out a large high-latitude bias in planktonic foraminiferal SST estimates. A commonly advocated conjecture might appear to resolve the mysteries of greenhouse climates, and it relies on oceanic heat transport in conjunction with high greenhouse gas concentrations. Attention has focused on the challenging problem of modeling paleo-ocean circulations and, especially, on placing

bounds on the magnitude of ocean heat transport in such climates. Climate models have been coupled to a ‘slab’ mixed-layer ocean model, that is, assuming that the important ocean thermal inertia is in the upper 50 m and that ocean poleward heat transport is at specified levels, to predict equator-to-pole surface temperature gradients. With appropriate (Eocene or Cretaceous) boundary conditions, near-modern pCO2, and ocean heat transport specified at near-modern values, an equator-to-pole temperature gradient (and continental winter temperatures) is very close to the modern result (Figure 7). With higher pCO2, tropical SSTs increase (to B32 1C for 1000 ppm) without changing meridional gradients substantially, and continental interiors remain well below freezing in winter. Substantial high-latitude amplification of temperature response to increases in pCO2 or other forcings is generally obtained because of the nonlinearity introduced by crossing a threshold from extensive sea ice cover to little or no sea ice cover. Proxy data imply that it was unlikely – but not impossible – that sea ice was present during the warmest of the greenhouse climates; therefore, this sensitivity is probably not representative for the true past greenhouse intervals. Nevertheless, evidence is growing of the possibility of the presence of sea ice and potentially ice sheets during some of the cooler parts of greenhouse climates. With heat transport approximately 3 times to modern in a slab ocean model, and pCO2 much higher than modern (42000 ppm), some of the main characteristics of greenhouse climates are reproduced, but we do not know how to reach such a high heat transport level without invoking feedbacks in the climate system that have not traditionally been considered, such as those due to tropical cyclones. There is a range, however, of smaller-than-modern temperature gradients that may not imply an increase in ocean heat transport over modern values. Near-modern values of ocean heat transport could support an equator-to-pole surface temperature gradient of 24 1C, whereas a 15 1C gradient requires ocean heat transport roughly 2–3 times as large. In other words, warm polar temperatures of 10 1C and tropical temperatures of 34 1C can exist in equilibrium with near-modern ocean transport values, but climates with smaller gradients are paradoxical. Thus even small errors or biases in tropical or polar temperatures affect our interpretation of the magnitude of the climate mystery posed by past greenhouse climates. In conclusion, two main hypotheses potentially operating in conjunction might explain greenhouse climates. The first is increased greenhouse gas



Table 1

327

Greenhouse climate research: key areas to resolve the low gradient and equable climate problems

Tropical proxies may be biased to cool values

What are the true depth habits of planktonic foraminifera? Are ‘mixed-layer’ dwellers calcifying below the mixed layer?

Could proxies reflect seasonal biases due to upwelling? Are productivity and calcification limited to relatively cool conditions?

Polar temperature proxies might be biased to warm values

Is there sea ice? Sea ice introduces a fundamental nonlinearity into climate. When did it first form? Is resolution in the atmosphere and ocean models too coarse to get correct dynamics or capture the ‘proxy scale’? What resolution is enough? How accurate are existing pCO2 proxies? How can they be improved? No proxy for methane or other greenhouse gases. How did other radiatively important constituents vary?

Are proxies recording only the warm events or warm season? Is polar warmth a fiction of aliasing?

Models missing key features

Greenhouse gases

Circulation proxies

Although ocean circulation tracers (e.g., Nd) provide information on the sources and directions of flow, there are currently no tracers that directly reflect ocean flow rates. Can quantitative proxies for rates of deep water formation and flow be developed? Was the ocean circulation weaker or stronger?

Is there missing physics: e.g., polar stratospheric clouds, enhanced ocean mixing; or incorrect physics: e.g., cloud parametrizations? The sensitivity of climate models to greenhouse gas forcing varies from model to model and the ‘correct’ values are an unknown, what can we learn from forcing models with greenhouse gases? Density-driven flows in the ocean are function of temperature and salinity distributions. Most proxies focus on temperature. Can existing techniques for estimating paleosalinity (especially surface) be refined and extended? Can new, quantitative proxies be developed? Was the ocean circulation of past greenhouse climates driven primarily by salinity gradients?


How much do preservation and diagenesis alter the proxy signals? Can we only trust foraminifera with ‘glassy’ preservation? Under what conditions do Mg/Ca produce valid temperatures? Are compound-specific isotope analyses also subject to alteration? What is the oxygen isotopic composition of polar seawater? Do hydrological cycle and ocean circulation changes alter this? Are paleogeographic/ paleobathymetric/ soil/vegetation/ozone boundary conditions correct enough?

Carbon cycle feedbacks. What caused long-term and short-term carbon perturbations? How did climate and the carbon cycle feed back onto each other? Oceanic carbon storage and oxygenation are largely controlled by ocean circulation. Can integrated carbon–oxygenisotope-dynamical models reveal unique and important linkages between proxies and ocean dynamics?

328


concentration, but we need better understanding of the coupled carbon cycle and climate dynamics on short and long timescales. The second is a physical mechanism that increases poleward heat transport strongly as the conditions warm.

See also Abrupt Climate Change. Cenozoic Climate – Oxygen Isotope Evidence. Determination of Past Sea Surface Temperatures. Heat Transport and Climate. Methane Hydrates and Climatic Effects. Paleoceanography. Paleoceanography, Climate Models in. Plio-Pleistocene Glacial Cycles and Milankovitch Variability.

Further Reading Barker F and Thomas E (2004) Origin, signature and palaeoclimate influence of the Antarctic Circumpolar Current. Earth Science Reviews 66: 143--162. Beckmann B, Flo¨gel S, Hofmann P, Schulz M, and Wagner T (2005) Orbital forcing of Cretaceous river discharge in tropical Africa and ocean response. Nature 437: 241--244 (doi:10.1038/nature03976). Billups K and Schrag D (2003) Application of benthic foraminiferal Mg/Ca ratios to questions of Cenozoic climate change. Earth and Planetary Science Letters 209: 181--195. Brinkhuis H, Schouten S, Collinson ME, et al. (2006) Episodic fresh surface waters in the early Eocene Arctic Ocean. Nature 441: 606--609 (doi:10.1038/ nature04692). Coxall HK, Wilson A, Palike H, Lear CH, and Backman J (2005) Rapid stepwise onset of Antarctic glaciation and deeper calcite compensation in the Pacific Ocean. Nature 433: 53--57. DeConto RM and Pollard D (2003) Rapid Cenozoic glaciation of Antarctica induced by declining atmospheric CO2. Nature 421: 245--249. Donnadieu Y, Pierrehumbert R, Jacob R, and Fluteau F (2006) Modelling the primary control of paleogeography on Cretaceous climate. Earth and Planetary Science Letters 248: 426--437. Greenwood DR and Wing SL (1995) Eocene continental climates and latitudinal temperature gradients. Geology 23: 1044--1048. Huber BT, Macleod KG, and Wing S (1999) Warm Climates in Earth History. Cambridge, UK: Cambridge University Press. Huber BT, Norris RD, and MacLeod KG (2002) Deep sea paleotemperature record of extreme warmth during the Cretaceous Period. Geology 30: 123--126. Huber M, Brinkhuis H, Stickley CE, et al. (2004) Eocene circulation of the Southern Ocean: Was Antarctica kept warm by subtropical waters. Paleoceanography PA4026 (doi:10.1029/2004PA001014).

Jenkyns HC, Forster A, Schouten S, and Sinninge Damste JS (2004) Higher temperatures in the late Cretaceous Arctic Ocean. Nature 432: 888--892. Lear CH, Elderfield H, and Wilson A (2000) Cenozoic deep-sea temperature sand global ice volumes from Mg/Ca in benthic foraminiferal calcite. Science 287: 269--272. MacLeod KG, Huber BT, and Isaza-Londonõ C (2005) North Atlantic warming during ‘global’ cooling at the end of the Cretaceous. Geology 33: 437--440. MacLeod KG, Huber BT, Pletsch T, Ro¨hl U, and Kucera M (2001) Maastrichtian foraminiferal and paleoceanographic changes on Milankovitch time scales. Paleoceanography 16: 133--154. Miller KG, Wright JD, and Browning JV (2005) Visions of ice sheets in a greenhouse world. In: Paytan A and De La Rocha C (eds.) Marine Geology, Special Issue, No. 217: Ocean Chemistry throughout the Phanerozoic, pp. 215--231. New York: Elsevier. Pagani M, Zachos J, Freeman KH, Bohaty S, and Tipple B (2005) Marked change in atmospheric carbon dioxide concentrations during the Oligocene. Science 309: 600--603. Parrish JT (1998) Interpreting Pre-Quaternary Climate from the Geologic Record, 348pp. New York: Columbia University Press. Pearson PN and Palmer MR (2000) Atmospheric carbon dioxide concentrations over the past 60 million years. Nature 406: 695--699. Pearson PN, van Dongen BE, Nicholas CJ, et al. (2007) Stable tropical climate through the Eocene epoch. Geology 35: 211--214 (doi: 10.1130/G23175A.1). Poulsen CJ, Barron EJ, Arthur MA, and Peterson WH (2001) Response of the mid-Cretaceous global oceanic circu-lation to tectonic and CO2 forcings. Paleoceanography 16: 576--592. Poulsen CJ, Gendaszek AS, and Jacob RL (2003) Did the rifting of the Atlantic Ocean cause the Cretaceous thermal maximum? Geology 31: 1115--1118. Prahl FG, Herbert TD, Brassell SC, et al. (2000) Status of alkenone paleothermometer calibration: Report from Working Group 3: Geochemistry. Geophysics, Geosystems 1 (doi:10.1029/2000GC000058). Royer DL, Berner RA, Montanez IP, Tabor NJ, and Beerling DJ (2004) CO2 as a primary driver of Phanerozoic climate. GSA Today 14: 4--12. Schouten S, Hopmans E, Forster A, van Breugel Y, Kuypers MMM, and Sinninghe Damste´ JS (2003) Extremely high sea-surface temperatures at low latitudes during the middle Cretaceous as revealed by archaeal membrane lipids. Geology 31: 1069--1072. Shellito C, Sloan LC, and Huber M (2003) Climate model constraints on atmospheric CO2 levels in the early– middle Palaeogene. Palaeogeography, Palaeoclimatology, Palaeoecology 193: 113--123. Sluijs A, Schouten S, Pagani M, et al. (2006) Subtropical Arctic Ocean conditions during the Palaeocene Eocene Thermal Maximum. Nature 441: 610--613 (doi:10. 1038/nature/04668).



Tarduno JA, Brinkman DB, Renne PR, Cottrell RD, Scher H, and Castillo P (1998) Evidence for extreme climatic warmth from late Cretaceous Arctic vertebrates. Science 282: 2241--2244. Thomas E, Brinkhuis H, Huber M, and Ro¨hl U (2006) An ocean view of the early Cenozoic greenhouse world. Oceanography 19: 63--72. von der Heydt A and Dijkstra HA (2006) Effect of ocean gateways on the global ocean circulation in the late Oligocene and early Miocene. Paleoceanography 21: 1--18 (doi:/10.1029/2005PA001149). Wilf P, Johnson KR, and Huber BT (2003) Correlated terrestrial and marine evidence for global climate changes

329

before mass extinction at the Cretaceous–Paleogene boundary. Proceedings of the National Academy of Sciences of the United States of America 100: 599--604. Zachos JC, Ro¨hl U, Schellenberg SA, et al. (2005) Rapid acidification of the ocean during the Paleocene–Eocene Thermal Maximum. Science 308: 1611--1615. Zachos J, Pagani M, Sloan L, Thomas E, and Billups K (2001) Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292: 686--694.


PARTICLE AGGREGATION DYNAMICS A. Alldredge, University of California, Santa Barbara, CA, USA

dissolved pools of matter in the ocean making it also relevant for marine chemistry.


Significance of Particle Dynamics Until the late 1970s it was thought that particulate organic matter sedimented to the deep sea via the slow downward sinking of tiny, suspended plants, animals, and debris. These particles supposedly produced a steady rain of fine detritus onto the deep seafloor. However, more recent studies using sediment traps (suspended cones or cylinders that intercept sedimenting material) demonstrate that large, rare, rapidly sinking particles are actually responsible for the majority of the downward vertical transport of matter from the surface waters of the ocean. The flux of these large sedimenting particles far exceeds that of the fine suspended particles, even though the large forms are far less numerous. Where do these large, rapidly sinking particles come from? This article examines the processes that produce and destroy sedimenting particles in the ocean. The processes by which particles are transformed in size and composition are collectively known as ‘particle dynamics.’ An understanding of particle dynamics is essential for predicting marine sedimentation rates and ultimately, for estimating the quantity of carbon that can be sequestered in the deep sea. Most of the organic carbon in the ocean originates through the fixation of carbon dioxide by photosynthetic organisms near the surface. Sequestration of large quantities of this organic carbon in the sediments of the deep ocean can effectively remove carbon from the global carbon cycle and ultimately reduce concentrations of atmospheric CO2, a major greenhouse gas, especially on long timescales of thousands of years. Thus oceanic sedimentation rates and particle dynamics are relevant to global climate change and the greenhouse effect. Particle dynamics also alters the sizes and types of particles available to animal grazers and microbial decomposers in the water column, making it important to the biology of many marine organisms. Moreover, processes that shift the particle size distributions in the ocean affect the optical properties of the water altering water clarity and signal transduction. Finally, particle dynamics involves transformation between the particulate and

330

Processes Forming Large Particles The average depth of the ocean is about 4000 m. Sedimenting particles must sink at speeds of about 50 to several hundred meters per day in order to reach such great depths before they completely dissolve or decompose. Even then, a particle generated in surface waters and sinking at 100 m per day would require 40 days to strike the ocean bottom. Only particles bigger than about 0.5 mm in diameter or very dense particles, such as mollusk shells or sand grains, can sink at rates fast enough to reach the seafloor before destructive processes, including decomposition, dissolution, disaggregation, or consumption by grazers, completely destroy or transform them. Thus sedimentation rates in the ocean are governed by the mechanisms responsible for repackaging smaller particles into larger units capable of rapid settlement. The most abundant types of large particles sedimenting to the seafloor fall into two general categories: fecal pellets and marine snow. Fecal pellets large enough to sink rapidly are produced by larger zooplankton such as euphausiids (krill), chaetognaths (arrow worms), amphipods, large copepods, pelagic crabs, salps (pelagic tunicates), pteropods (pelagic mollusks) and many other minor taxa. The generic term ‘marine snow’ describes aggregates of highly diverse origins, structure, and characteristics that are larger than 0.5 mm in diameter (see Marine Snow). Marine snow includes fragile, porous, loose associations of smaller particles, highly cohesive, robust gelatinous webs produced by zooplankton, and some flocculent, porous fecal pellets. The two classes of rapidly sinking particles are produced from smaller particles via two major pathways in the ocean. First, biologically mediated aggregation results when small particles are consumed and transformed into fecal pellets or feeding webs by the feeding activities of marine animals. Second, the coagulation of small discrete particles into larger heterogeneous aggregates occurs largely through physically mediated processes of collision and sticking. Origins of Primary Particles

Both the coagulation and consumption pathways depend on the types and abundances of smaller


PARTICLE AGGREGATION DYNAMICS

particles available for physically or biologically mediated aggregation. These smaller particles are referred to as ‘primary particles’ because they are the basic units involved in aggregation. Figure 1 illustrates the most important types of primary particles. Most primary particles consist of organic matter whose origins can be traced back to the fixation of inorganic carbon dioxide by photosynthetic

331

phytoplankton. These microscopic algae are also at the base of the food web, supporting the growth of other living primary particles such as bacteria and protozoans. Certain phytoplankton, including diatoms, coccolithophorids (single celled, calcareous algae), and prymnesiophytes (golden-brown algae) contribute directly to marine snow formation. Chainforming diatoms become entangled and aggregate to

Dust CO2

Runoff

Dissolved organic matter (including colloids)

Photosynthesis Uptake

Surface coagulation adsorption

Primary particles

Phytoplankton, algal detritus

Microorganisms and bacteria

Gels, microaggregates (TEP, CSP etc.)

Flakes from bubble dissolution

Inorganic particles

Collided together by: _ shear _ differential settlement _ Brownian motion _ organism motility _ diffusive capture _ filtration

Animal consumption

Coagulation Stuck together by: _ exocellular polymers _ products of cell lysis _ cell surface properties Aggregated particles

Shells, tests, carcasses, molts

Mucus from zooplanktons

Fecal pellets

Marine snow and microaggregates

Lateral advection Settlement

Sediment resuspension

Figure 1 Major sources of particles in sea water. Primary particles are formed from inorganic or dissolved organic sources via processes such as photosynthesis, microbial uptake, and bubble dissolution. The large particles responsible for the sedimentation of particulate organic matter in the ocean are formed from primary particles via animal consumption (feeding webs, fecal pellets) and coagulation. TEP, transport exopolymer particles; CSP, Coomassie stained particles. (Modified with permission from Hurd DC, Spencer DW (eds) (1991) Marine Particles: Analysis and Characterization. Washington DC: American Geophysical Union.


332


form large sinking composite particles and prymnesiophytes of the genus Phaeocystis exist within a mucus matrix and are often responsible for the large quantities of mucus reported to cause damage to fishing nets and beaches in the North Atlantic. Phytoplankton also excrete copious polysaccharides and other organic molecules which increase the pool of dissolved organic matter in the ocean and support the growth of bacteria and the formation of primary particles from dissolved precursors. Finally, as they age, die, and decompose the broken spines and tests, cell walls, and cellular components of phytoplankton contribute to the pool of nonliving detritus. Dissolved organic matter (DOM) in the ocean is also a major source of primary particles. There is about 10 times more DOM in the ocean than particulate organic matter (POM) and about 50% of this DOM consists of colloids. These tiny particles, between 1 nm and 1 mm in size, reach abundances as high as 107–109 per ml in the ocean and can readily coagulate into larger particles. Three major types of primary particles important in the formation of sedimenting particles are generated from DOM: bacteria, flakes, and gels (Figure 1). First, new bacteria cells are produced when bacteria transport readily usable dissolved molecules such as amino acids, carbohydrates, and lipids across their cell membranes and use them to fuel cellular metabolism, growth, and cell division. Thus uptake of DOM by bacteria generates particulate matter in the form of new bacteria cells

Streamlines

Rising bubble

which are then available for further aggregation or consumption up the food web. Second, flakes are produced by the dissolution of rising bubbles of air generated by white caps and waves breaking at the sea surface. Turbulent mixing forces these bubbles underwater, sometimes to depths of many meters or even tens of meters. As the bubbles rise surface active dissolved matter, including fatty acids, sterols, fatty alcohols, and proteins, adsorb to the bubble surface and many colloidal particles collide with and stick to the bubble surface through surface coagulation (Figure 2). The air within the bubble gradually dissolves causing the material adhering to the bubble surface to collapse into a small flake of particulate organic matter. The size of these flakes is dependent on the initial diameter of the bubble. Most bubbles generated by breaking waves in the ocean are o100 mm in diameter and generate flakes of POM on the order of 2–20 mm in size. Anyone who has ever seen breaking ocean waves during high winds or violent storms can readily appreciate the potential magnitude of this process for producing particulate matter in the ocean. Finally, transparent gel particles form from DOM, especially polysaccharides and proteins. Phytoplankton can excrete as much as 50% of their photosynthetically fixed carbon as long-chain polysaccharides. These molecules align in sea water, often via attractions between positively and negatively charged portions of the molecules, into colloidalsized mucilaginous units called fibrils. These

Streamlines Pore

Diffusion limited region

(A) Brownian motion

(B) Shear

(C) Differential settlement

(D) Surface coagulation

(E) Diffusive capture

(F) Filtration

(G) Motility

Figure 2 Physical collision mechanisms of particles in the ocean. (A) Brownian motion – particles smaller than a few micrometers collide while moving randomly. This is the major mechanism of collision for colloidal-sized particles. (B) Shear – particles carried by parallel streamlines of different velocities collide when one particle overtakes the other. This mechanism is most important for particles larger than a few micrometers in diameter. (C) Differential settlement – rapidly sinking particles overtake and collide with smaller, more slowly sinking ones. Collision depends on the smaller particle following streamlines close enough to the larger particle to produce contact. This process often results in collisions in the wake of the larger particle, generating comet-shaped aggregates. (D) Surface coagulation – the aggregation of colloids on rising bubbles is orders of magnitude more efficient than it would be if the bubbles were solid spheres because streamlines carrying the colloids are drawn to the bubble by internal gas circulation that moves the bubble surface. Colloids are also caught in the wake of the rising bubble. (E) Diffusive capture – rapidly sinking particles have boundary layers around them. Small particles may be carried across this boundary layer by diffusion and collide. (F) Filtration – large aggregates are porous with many channels through their interior. Smaller particles may be captured as they flow into these channels. (G) Organism motility – microorganisms swim and tumble leading to collision with other particles in the water.



aggregate and swell to form gel-like material in water. These gel particles, known as transparent exopolymer particles (TEP), are discrete particles up to hundreds of micrometers long that appear as amorphous spheres, films, sheets, or strings on filters when stained with Alcian Blue, a dye specific for acidic polysaccharides (Figure 3A). In marine environments TEP occur at concentrations up to thousands per millimeter. Similar transparent gel-like particles rich in proteins, rather than carbohydrates, known as Coomassie stained particles (CSP) after Coomassie, a stain specific for proteins, exist at even higher abundances than TEP. Cell lysis which releases protein-rich cell components, adsorption of proteins onto mucilage surfaces, and microbial exoenzymes may be sources of CSP. The existence of both TEP and CSP was only discovered in the mid 1990s because unstained particles are totally transparent to the eye and the microscope. Since they exist as gels, they are mostly water. But their large size, high abundance, and sticky surfaces make them particularly important primary particles for the formation of large sinking aggregates. Gel particles appear to be the ‘glue’ that helps bind many of the other types of primary particles together to form larger aggregates (Figure 3B). The last major class of primary particles is inorganic and includes primarily clay-minerals derived from sediment resuspension, terrestrial runoff, and wind-blown dust. Inorganic particles are relatively rare in the open ocean but can be a significant component of sedimenting aggregates near shore or near the water–sediment interface. The surfaces of inorganic particles entering the ocean become coated with absorbed organic molecules and develop a slight negative charge that affects their ability to aggregate.

333

Formation of Aggregated Particles

Despite the highly variable appearance and composition of large sedimenting particles (Figure 4) they are formed from combinations of primary particles by only two major pathways which are discussed here using Figure 1 as a framework. Biologically mediated aggregation: Particles produced as by-products of animal consumption Herbivorous zooplankton consume phytoplankton and microorganisms in the ocean and they, in turn, are consumed by carnivorous animals up the food chain. Flakes, gels, and detrital particles are also consumed. Feeding at all trophic levels brings together primary particles and transforms them into new animal growth, respired carbon dioxide, various dissolved excretory products, and fecal pellets. Several types of large particles result from feeding (Figure 1) including fecal pellets, feeding webs, and molts. Assimilation efficiencies of zooplankton vary from about 60% to about 90%; thus 10–40% of the food consumed by zooplankton is repackaged into fecal pellets. Fecal pellets contain not only partially digested particles but also bacteria and some phytoplankton that escape digestion. They may be encased in a membrane (crustaceans, chaetognaths, larvaceans, etc.) or be loose, fluffy conglomerations of digestive waste (fish, doliolids). Larger pellets produced by big crustaceans (krill, amphipods, pelagic crabs), salps, chaetognaths, pteropods and many others can sink at rates of several hundred to over 2000 m per day. At these rates many large pellets reach the seafloor before decomposing completely. Smaller pellets produced by copepods, protozoans,

(A) Figure 3 (A) Gel-like transparent exopolymer particles (TEP), here dispersed around a diatom chain, occur as stings, spheres, and sheets when stained with Alcian Blue, a stain specific for acidic polysaccharides. (B) These gels form the matrix of many marine aggregates of phytoplankton and detrital material. Scale bar, 100 mm. (B reprinted from Alldredge AL, Passow U and Logan BE (1993) The abundance and significance of a class of large, transparent organic particles in the ocean. Deep-Sea Research 40: 1131–1140, with permission of Elsevier Publishing Co, London.)


334


(A)

(B)

(C)

(D)

Figure 4 Examples of large aggregated particles. (A) Aggregate formed by the coagulation of living, chain-forming diatoms; scale bar, 1 cm. (B) Scanning electron micrograph of the interstices of a diatom floc composed predominantly of individuals of the genus Chaetoceros; scale bar, 40 mm. (C) Abandoned house of a larvacean with filtering structures evident. This particle is formed via the animal consumption pathway; scale bar, 5 mm. (D) Light micrograph of a typical aggregate of detritus and debris formed by the coagulation pathway; scale bar, 1 mm.

and other small zooplankton may sink only a few meters per day. Small pellets rapidly decompose in the water column and rarely contribute significantly to the downward flux of particulate matter. Feeding webs are a second type of aggregated particle generated by zooplankton feeding. Planktonic tunicates known as larvaceans and planktonic snails, or pteropods, produce large mucous particles as part of their feeding biology. Larvaceans secrete a gelatinous ‘house’ around themselves with which they filter bacteria, small algae, and other particles from the surrounding seawater (Figure 4C). These houses are aggregates of gelatinous mucus and a variety of phytoplankton, bacteria, and smaller particles filtered from the water during feeding. The house filters clog quickly and an average larvacean may produce four to six houses per day, discarding each house sequentially. Most larvacean houses range from about 1 to 10 mm in diameter although

species in the deep sea produce houses almost 100 cm across. Abundances up to 80 houses l 1 and particle fluxes as high as 9 104 houses m2 d 1 have been reported from tropical bays. Larvaceans are also common in the open ocean and can be an important source of sinking aggregates when abundant. Mucus feeding webs produced by planktonic snails and foraminifera (spined calcareous amoeboid protozoa) are also a source of marine snow, especially in oceanic areas where these animals are abundant. However, no quantitative data presently exist as to the abundance of these webs or their contribution to particle flux in the ocean. Finally, animals generate shells, tests, carcasses, and molts as they grow and die. These are included with aggregated particles because they are generated by animals through the consumption and transformation of primary particles. Shells, in particular, sink rapidly and can reach the bottom in many parts of



the ocean before they dissolve, accumulating to form hydrocarbon oozes, such as pteropod and foraminiferal oozes, and calcareous and siliceous deposits. Physically mediated aggregation: coagulation Although fecal pellets are important in particle flux, most POM sediments to the seafloor as marine snow. Marine snow is formed by the coagulation of the many types of smaller particles available in the water column including phytoplankton, bacteria, flakes, gels particles, detritus, fecal pellets, and inorganic particles. Coagulation is a two step process. First, primary particles must collide together. Second, they must stick together on collision to form an aggregate. This is an iterative process. Two primary particles collide and stick. This aggregate than collides with a third particle, then a fourth, and so on forming a larger and larger aggregate over time. All sizes of particles from colloids to marine snow can coagulate up the size spectrum to generate the large sedimenting particles important for particle flux. Coagulating particles are brought together in the ocean largely by physical processes. The mechanisms that contribute to the collision of suspended particles in the water column are illustrated in Figure 2. The mechanisms most important for the aggregation of colloidal particles include Brownian motion, surface coagulation, diffusive capture, and possibly filtration. These mechanisms efficiently aggregate colloids up the size spectrum until they become large enough to contribute to the formation of even larger aggregates of marine snow. Aggregates in the marine snow size range (>0.5 mm in diameter) are produced primarily by collisions of larger primary particles and aggregates via shear and differential settlement (Figure 2). Swimming also brings motile bacteria, phytoplankton, and protists into contact with aggregates and facilitates colonization and the growth of complex detrital communities on particles. Once collided together, particles must stick together. Most particles in the ocean are relatively unsticky. When marine sediments, healthy phytoplankton, and other small marine particles collide generally less than 1 in 10 collisions actually results in the two colliding particles sticking together. Often rates are even lower with only 1 in 100 inorganic particles or phytoplankton joining on collision. Such low sticking efficiencies imply that coagulation rates should be very low in the ocean. However, gel particles are quite sticky and facilitate aggregation. Sedimenting marine snow is formed primarily from the collisions of particles on the order of tens to hundreds of micrometers in size. The polymers

335

associated with marine snow increase stickiness to the point where 70–80% of collisions between larger aggregates result in sticking. This is about 10 times higher than the sticking probabilities of smaller particles of marine sediment which lack the extensive biological communities and polysaccharide glues associated with marine snow. Our theoretical understanding of coagulation in marine systems has been guided largely by the work of colloidal chemists, atmospheric scientists and waste water engineers. The coagulation theory that marine scientists inherited from these disciplines was developed to describe simple systems containing identical, spherical, nonliving particles. In theory, any aggregates formed were larger but otherwise identical in shape and density to the original particles. These assumed particle properties are contradicted in almost every way in natural marine systems that contain heterogeneous mixtures of particles from many sources. The physical and chemical properties of natural particles vary; their shapes are seldom spherical; and they often contain spines. Furthermore, living particles grow and divide, even within aggregates. When particles do coagulate, they form porous aggregates with properties described by fractal rather than Euclidean geometry. Fractal aggregates have smaller mass and slower sinking rates than predicted for Euclidean particles of similar size. Finally, quantifying and modeling the impact of each individual collision mechanism given the diversity of both particle characteristics and physical conditions in various regions of the ocean is a daunting task. These complexities have made it very difficult to accurately predict coagulation rates in the ocean and quantitative estimates of coagulation rates in nature and the development of theory to accommodate complex oceanic systems are both in their infancy. However, basic theoretical considerations can still identify significant factors governing aggregation. Coagulation rates for the formation of larger, sinking particles are primarily a function of three major factors: the abundance and size distribution of primary particles; sticking efficiency between particles; and the intensity of the particular physical process colliding particles, especially shear and differential settlement. Early application of coagulation theory to particle dynamics in the water column indicated that the abundance of primary particles and the intensity of physical processes, especially shear, were simply too low to explain the high abundance of marine snow and other aggregates observed in nature. However, recent applications of coagulation theory that take into account increased particle abundances and sizes resulting from the presence of TEP and CSP, increased sticking efficiencies of larger


336


particles, and nonspherical particle sizes and shapes indicate that coagulation is an important process controlling maximum phytoplankton concentrations and the vertical flux of organic matter, even in the nonbloom conditions of the oligotrophic ocean. Although biological aggregation through animal consumption is certainly the most significant processes regulating the aggregation of particles smaller than about 5–10 mm in the ocean, coagulation appears to be as or more important than grazing for aggregating particles larger than about 10 mm. However, extensive quantitative comparisons of these two pathways have not been made. One system where coagulation is particularly important and greatly increases the flux of large particles is phytoplankton blooms. Aided by increasing abundances of TEP from phytoplankton exudation, blooms of chain-forming diatoms, particularly the chain-forming genera Nitzschia and Chaetoceros, aggregate into centimeter-sized particles of marine snow, often coinciding with nutrient depletion (Figure 4A, B). These flocs settle at velocities of 100– 200 m day 1 and quickly transport photosynthetically derived carbon directly to the seafloor, often before it can be consumed by zooplankton. The rapid sinking of large flocs formed by diatoms helps to explain why diatoms and the shells of other small organisms are often distributed in deep-sea sediments directly below the surface populations originally forming them. Previously these distributions had been surprising given the relatively slow settling rates of individual shells and the possibility for horizontal displacement by underlying ocean currents. We now know that these organisms reach the seafloor in large, rapidly settling flocs of marine snow.

Processes that Destroy and Transform Large Particles Particle dynamics also includes processes that either directly consume and destroy aggregates or that transform them into nonsinking or more slowly sinking forms. These destructive processes are very effective. Generally less than 10% of photosynthetically fixed carbon sinks to the seafloor in shallow coastal seas and less than 1% reaches the seafloor over most of the deeper ocean. Five major processes are responsible for the transformation of large sinking particles in the ocean. Quantitative estimates of the significance of these various destructive and transformative processes are very rare and considerable future research is needed to fully understand their role in particle dynamics.

First, many types of zooplankton including copepods, euphausiids, salps, pteropods, and some fish, such as midwater myctophids, eat marine snow and others are known to consume fecal pellets. Animal feeding transforms marine snow into smaller (and usually more slowly sinking) fecal pellets, new animal growth and reproduction, and respired carbon dioxide. Although the consumption of rapidly sinking particles by zooplankton and fish has not been quantified, it is most likely a major loss process given the abundance of zooplankton. In the deep sea, in particular, where food is relatively scarce, rapidly sinking particles may be particularly attractive for grazers. A second major process is microbial decomposition, mostly by bacteria inhabiting sinking particles. These organisms utilize the particulate matter on the sinking particles to generate new bacteria cells but much of the organic carbon is lost to respiration. A third related process is solubilization. All bacteria must first convert particulate matter into dissolved form in order to transport it across their cell membranes. Bacteria attached to particles produce copious exoenzymes which solubilize the particulate matter and convert it into dissolved organic molecules that diffuse into the surrounding sea water. Simultaneously, the attached bacteria divide and release many of their offspring into the surrounding sea water, as well. Thus, the attached forms seed the sea water with both offspring and the dissolved food to help sustain them by exploiting particulate matter on rapidly sinking particles. Solubilization of marine snow adds to the pool of dissolved organic matter in the ocean at the expense of the particulate pool. Decomposition and solubilization processes degrade aggregates over long time spans of weeks to months. These rates are slow enough to allow larger aggregates time to reach the seafloor. Physical processes such as fluid shear from currents, upwelling and mixing can physically break sinking particles into smaller more slowly sinking forms. These smaller particles remain longer in the water column making them even more susceptible to animal grazing and decomposition by bacteria. Finally, turbulence generated by the swimming and vertical migration patterns of larger animals, especially krill, salps, medusa, and other large zooplankton also disrupt marine snow and break it into smaller, more slowly sinking particles. This latter process may be more important than shear generated by mixing since average shear rates in most of the ocean have been shown to be far too low to disrupt most types of marine snow. These five destructive processes and the various aggregation processes all act on marine particles simultaneously. The ultimate fate of any particular



particle depends greatly on which processes dominate its transformation and its interactions with other particles and organisms in the water column.

Conclusions Particle dynamics affects not only the rate of sedimentation in the ocean but also the ecology of the organisms living in the water column, the sizes of the pools of important chemical constituents such as carbon and nitrogen, and the optical properties of the water itself. The complex array of processes that contribute to particle dynamics have been identified. However, the relative importance of each of these processes at different times and places in the ocean has yet to be fully investigated. A quantitative understanding of particle dynamics will be necessary in order to accurately predict sedimentation rates and the role of the ocean in global scale problems such as climate change.

Summary Most of the material sedimenting to the seafloor sinks as large fecal pellets or as rapidly settling aggregates of phytoplankton and organic debris known as marine snow. Particle dynamics encompasses the processes by which these particles are produced, transformed, and consumed in the ocean. Large, rapidly settling particles are formed from the aggregation of small, slowly sinking particles including phytoplankton, microorganisms, gels, flakes, detritus, and clay-minerals by two pathways. First, primary particles can be transformed biologically into fecal pellets, molts, and mucus feeding webs through the feeding activities of zooplankton. Second, they can collide with each other and stick together to form progressively larger aggregates through the physical process of coagulation. Although it may take large particles many days or weeks to sediment to the seafloor, some reach the ocean bottom before they are destroyed by animal consumption, bacterial decomposition, or disaggregation. The balance between the processes of production and destruction determines the quantity and size distribution of particulate organic matter in the water column and the quantity of material deposited on the seafloor.

Glossary Assimilation efficiency The percentage of consumed food that is absorbed through the gut wall and becomes available for cellular metabolism. Colloids Particles from 1 nm to about 1 mm in size. Colloids do not sink and they are considered to be

337

part of the pool of dissolved organic matter in the ocean. Detritus Non-living particulate organic matter including algal debris. Dissolved organic matter Organic matter that passes through a filter with a pore size of 0.7– 0.8 mm. Fractal geometry A nonlinear, noninteger mathematics that has evolved to describe rugose, wrinkled objects such as clouds, coastlines, and aggregates. Marine snow Aggregated marine particles larger than 0.5 mm in diameter. Oligotrophic Regions of the ocean characterized by low nutrient concentrations, low phytoplankton biomass, and high water clarity. Particle Flux The quantity of particulate matter sinking through a given area of the water column, usually a square meter, at a given depth per unit time. Phytoplankton Plant plankton. Zooplankton Animal plankton, including protozoans.

See also Carbon Cycle. Floc Layers. Gelatinous Zooplankton. Marine Snow.

Further Reading Alldredge AL (1976) The Appendicularia. Scientific American 235: 94--102. Alldredge AL and Silver MW (1988) Characteristics, dynamics and significance of marine snow. Progress in Oceanography 20: 41--82. Fowler SW and Knauer GA (1986) Role of large particles in the transport of elements and organic compounds through the oceanic water column. Progress in Oceanography 16: 147--194. Kepkay PE (1994) Particle aggregation and the biological reactivity of colloids. Marine Ecology Progress Series 190: 293--304. Lalli CM and Parsons TR (1997) Biological Oceanography: an Introduction. Oxford: Butterworth-Heinemann. Logan BE (1999) Environmental Transport Processes. New York: John Wiley. McCave IN (1984) Size spectra and aggregation of suspended particles in the deep ocean. Deep-Sea Research 31: 329--352. Stumm W (1992) Chemistry of the Solid–Water Interface. New York: John Wiley. Stumm W and Morgan JP (1981) Aquatic Chemistry. New York: Wiley Interscience. Summerhayes CP and Thorpe SA (1996) Oceanography, An Illustrated Guide. London: Manson.


PAST CLIMATE FROM CORALS A. G. Grottoli, University of Pennsylvania, Philadelphia, PA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2098–2107, & 2001, Elsevier Ltd.

Introduction The influence of the tropics on global climate is well recognized. Our ability to understand, model and predict the interannual, decadal and long-term variability in tropical climate depends on our knowledge of past climate. However, our understanding of the natural variability in tropical climate is limited because long-term instrumental records prior to 1950 are sparse or nonexistent in many tropical regions. Continuous satellite monitoring did not begin until the 1970s and in situ equatorial Pacific Ocean monitoring has only existed since the 1980s. Therefore, we depend on proxy records to provide information about past climate. Proxy records are indirect measurements of the physical and chemical structure of past environmental conditions chronicled in natural archives such as ice cores, sediment cores, coral cores and tree rings. In the tropical oceans, the isotopic, trace and minor elemental signatures of coral skeletons can vary as a result of environmental conditions such as temperature, salinity, cloud cover and upwelling. As such, coral cores offer a suite of proxy records with potential for reconstructing tropical paleoclimate on intraannual-to-centennial timescales. Massive, symbiotic stony corals are good tropical climate proxy recorders because: (1) they are widely distributed throughout the tropics; (2) their unperturbed annual skeletal banding pattern offers excellent chronological control; (3) they incorporate a variety of climate tracers from which paleo sea surface temperature (SST), sea surface salinity (SSS), cloud cover, upwelling, ocean circulation, ocean mixing patterns, and other climatic and oceanic features can be reconstructed; (4) their proxy records can be almost as good as instrumental records; (5) their records can span several centuries; and (6) their high skeletal growth rate (usually in the range of 5–25 mm y 1) permits subseasonal sampling resolution. Thus proxy records in corals provide the best means of obtaining long seasonal-to-centennial timescale paleoclimate information in the tropics.

338

Coral-based paleoclimate research has grown tremendously in the last few decades. Most of the published records come from living corals and report the reconstructed SST, SSS, rainfall, water circulation pattern, or some combination of these variables. Records from fossil and deep-sea corals are also increasing. The goal of this chapter is to provide an overview of the current state of coral-based paleoclimate research and a list of further readings for those interested in more detailed information.

Coral Biology General

Corals are animals in the cnidarian family, order Scleractinia (class Anthozoa). Their basic body plan consists of a polyp containing unicellular endosymbiotic algae known as zooxanthellae overlaying a calcium carbonate exoskeleton (Figure 1). Most coral species are colonial and include many polyps interconnected by a lateral layer of tissue. The polyp, consisting of tentacles (used to capture prey), an oral opening and a gastrovascular cavity, has three tissue layers: the epidermis, mesoglea, and gastrodermis. The symbiotic zooxanthellae are located in the gastrodermis. Corals deposit a calcium carbonate (CaCO3) aragonite skeleton below the basal epidermis. Corals reproduce sexually during mass-spawning events by releasing egg and/or sperm, or egg–sperm bundles into the water column. Mass spawning events typically occur a few times a year for each species, and are triggered by the lunar cycle. Corals can also reproduce asexually by fragmentation. Animal–Zooxanthellae Symbiosis

Corals acquire the greater part of their food energy by two mechanisms: photosynthesis and heterotrophy (direct ingestion of zooplankton and other Tissue layers Tentacles

Epidermis Mesoglea Gastrodermis

Oral opening Gastrovascular cavity

Skeleton Figure 1 Cross-section of coral polyp and skeleton.


PAST CLIMATE FROM CORALS

339

organic particles in the water column by the coral animal). Photosynthesis is carried out by the endosymbiotic zooxanthellae. The bulk of photosynthetically fixed carbon is translocated directly to the coral host. In some cases, the coral animal can obtain all of its daily energy requirements via photosynthesis alone. In general, as light intensity increases, photosynthesis increases.

Table 1 Environmental variable(s) that can be reconstructed from coral skeletal isotopes, trace and minor elements, and growth records

Skeleton

D14C

Corals deposit skeleton below the basal epidermis. Typically, corals deposit one high-density and one low-density band of skeleton each year. The highdensity band has thicker skeletal elements than the low-density band. Each band is often composed of several finer bands called dissepiments, deposited directly at the base of the coral tissue. At discrete intervals, the polyp presumably detaches from the dissepiment, and begins to lay down a new skeletal dissepiment. Some evidence suggests that dissepiments may form on a lunar cycle. High- and low-density bands are deposited seasonally. Overall, high-density bands form during suboptimal temperature conditions and low-density bands form during optimal temperature conditions. At higher latitudes (i.e., Hawaii, Florida) optimal growth temperature occurs in summer. At lower latitudes (i.e., Gala´pagos, equatorial Pacific regions, Australian Great Barrier Reef), optimal growth temperatures occur in the cooler months. The width and density of growth bands also vary with environmental variables such as light, sedimentation, season length, and salinity. In general, as light levels decrease due to increased cloud cover, increased sedimentation or due to increasing depth, maximum linear skeletal extension decreases, calcification decreases, and skeletal density increases.

The Interpretation of Isotopes, Trace Elements, and Minor Elements in Corals Several environmental variables can be reconstructed by measuring changes in the skeletal isotope ratios, trace and minor elemental composition, and growth rate records in coral cores (Table 1). The width, density, and chemical composition of each band are generally thought to reflect the average environmental conditions that prevailed during the time over which that portion of the skeleton was calcified. Reconstructions of seawater temperature, salinity, light levels (cloud cover), upwelling, nutrient composition and other environmental parameters have been obtained from coral records (Table 1).

Proxy Isotopes d13C d18O

Trace and minor elements Sr/Ca Mg/Ca U/Ca Mn/Ca Cd/Ca d11B F Ba/Ca

Skeleton Skeletal growth bands

Fluorescence

Environmental variable

Light (seasonal cloud cover), nutrients/zooplankton levels Sea surface temperature, sea surface salinity Ocean ventilation, water mass circulation

Sea surface temperature Sea surface temperature Sea surface temperature Wind anomalies, upwelling Upwelling pH Sea surface temperature Upwelling, river outflow, sea surface temperature

Light (seasonal changes), stress, water motion, sedimentation, sea surface temperature River outflow

Method

Continuous records of past tropical climate conditions can be obtained by extracting a core from an individual massive coral head along its major axis of growth. Typically, this involves placing a coring device on the top and center of the coral head (Figure 2A). The extracted core is cut longitudinally into slabs ranging in thickness from 0.7 to 1 cm that are then X-rayed. X-ray-positive prints reveal the banding pattern of the slab and are used: (1) as a guide for sample drilling and (2) to establish a chronology for the entire coral record when the banding pattern is clear (Figure 2B). Samples are drilled out along the major axis of growth by grinding the skeletal material with a diamond-tipped dental drill. For high-resolution climate reconstructions, samples are extracted every millimeter or less down the entire length of the core. Since corals grow about 5–15 mm per year, this sampling method can yield approximately bimonthly-tomonthly resolution. Much higher resolution sampling is possible, yielding approximately weekly samples, but this is not commonly performed. In most cases, the d13C (the per mil deviation of the ratio of 13C/12C relative to the Peedee Belemnite (PDB) Limestone Standard) and d18O (ratio of 18 O/16O relative to PDB) values of each sample are


340


Porites sp.

1996 1995

1994

1993

1992

1991

1990 (A)

(B)

Figure 2 Collecting coral cores. (A) Coral core being extracted from top and center of an individual massive coral head using a pneumatic coring device. (Photo courtesy of M. Kazmers/Shark Song Tax ID #374-50-5314.) (B) X-ray positive print reveals the banding pattern of the slab and is used to help establish a chronology for the entire coral record.

measured. Since the d13C and/or d18O compositions of corals usually have a strong seasonal component, they are often used to establish an accurate chronology and/or to confirm the chronology established from the X-rays. Temperature and Salinity Reconstructions

Coral skeletal d18O reflects a combination of the local SST and SSS. In the many regions of the tropical ocean where the natural variation in salinity is small, changes in coral skeletal d18O primarily reflect changes in SST. The d18O of coral skeleton responds to changes in temperature usually according to the standard paleotemperature relationship for carbonates. Based on empirical studies, a 11C increase in water temperature corresponds to a decrease of about 0.22% (parts per thousand) in d18O. Precipitation has a low d18O value relative to that of sea water. Therefore, in regions with pronounced variability in rainfall and/or river runoff, coral d18O values reflect changes in SSS. Thus, depending on the nature of the coral collection site, the d18O record is

used to reconstruct the SST and/or SSS. Additional studies show that other proxy indicators of temperature include the ratios of strontium/calcium (Sr/ Ca), magnesium/calcium (Mg/Ca), and uranium/ calcium (U/Ca), fluorine levels (F) and skeletal band thickness (Table 1). The ratios of Sr/Ca, Mg/Ca, and U/Ca incorporated into the skeleton is largely determined by the temperature-dependent distribution coefficient of Sr/Ca, Mg/Ca, and U/Ca between aragonite and sea water. As temperatures increase, the Sr/Ca and U/Ca ratios decrease and the Mg/Ca ratio increases. Cloud Cover and Upwelling

d13C seems to indicate seasonal changes in cloud cover and upwelling. Thus far, only a small number of studies have used d13C records to confirm seasonal rainfall patterns established using the d18O signature. Only one study has directly linked a d13C record with seasonal upwelling. d13C in coral skeletons has been difficult to use as a paleoclimate tracer because it is heavily influenced by metabolic processes,



namely photosynthesis and heterotrophy. Firstly, as light levels decrease due to cloud cover, the rate of photosynthesis by the coral’s symbiotic zooxanthellae decreases, and skeletal d13C decreases. The reverse occurs when light levels increase. Secondly, zooplankton have a low d13C value relative to coral. During upwelling events in the Red Sea, nutrient and zooplankton level increases have been linked to decreases in coral skeletal d13C values. Other upwelling tracers include cadmium (Cd) and barium (Ba) concentrations, and D14C. Cadmium and barium are trace elements whose concentrations are greater in deep water than in surface water. During upwelling events, deep water is driven to the surface and cadmium/calcium (Cd/Ca) and barium/ calcium (Ba/Ca) ratios in the surface water, and consequently in the coral skeleton, increase (Table 1). Although SST also influences Ba/Ca ratios, most of the variation in Ba/ Ca ratios in corals is due to nutrient fluxes and upwelling. D14C is also an excellent tracer for detecting upwelling and changes in seawater circulation (D14C is the per mil deviation of the ratio of 14C/12C relative to a nineteenth century wood standard). For example, in the eastern equatorial Pacific Ocean, the D14C value of deep water tends to be very low relative to the D14C of surface water. Here, increased upwelling or increases in the proportion of deep water contributing to surface water results in a decrease in the D14C of the coral skeleton. Manganese (Mn) is a trace element whose concentration is highest in surface waters and decreases with depth. Therefore, during upwelling events, Mn/Ca ratios decrease. The ratio of Mn/Ca can also record prolonged and sustained changes in winds. In at least one case, Mn/Ca ratios from a Tarawa Atoll coral increased during El Ninõ events as a result of strong and prolonged wind reversals that had remobilized manganese from the lagoon sediments. Other Proxy Indicators

Other environmental parameters that can be inferred from coral skeleton structure and composition are river outflow (fluorescence bands) and pH (boron isotope levels) (Table 1). Large pulses in river outflow can result in an ultraviolet-sensitive fluorescent band in the coral record. New evidence strongly suggests that the fluorescent patterns in coral skeletal records are due to changes in skeletal density, not terrestrially derived humics as previously thought. Variations in salinity associated with fresh water discharge pulses from rivers appear to cause changes in coral skeletal growth density which can be observed in the skeletal fluorescence pattern. In the case of boron, d11B levels in sea water increase as pH

341

increases. Changes in the pH at the site of coral calcification seem to reflect changes in productivity of the symbiotic zooxanthellae. As photosynthesis increases, pH increases, and d11B levels in the coral skeleton increase.

Coral Records: What has been Learned About Climate From Corals? To date, there are over 100 sites where coral cores have been recovered and analyzed (Figure 3). Of these, at least 22 have records that exceed 120 years in length. In most cases, d13C and d18O have been measured at annual-to-subannual resolution. d18O as a SST and/or SSS proxy is the best understood and most widely reported of all the coral proxy measurements. In a few cores, other isotopic, trace, and minor elements, or skeletal density and growth measurements have also been made. Typically, coral-derived paleoclimate records are studied on three timescales: seasonal, interannual-todecadal, and long-term trends. The seasonal variation refers to one warm and one cool phase each year. An abrupt shift in the proxy’s long-term mean often indicates a decadal modulation in the data. Long-term trends are usually associated with a gradual increase or decrease in the measured proxy over the course of several decades or centuries. The following sections explore some of the seasonal, decadal and long-term trends in coral-derived paleoclimate records and some of the limitations associated with interpreting coral proxy records. Seasonal Variation in Coral Climate Records

Seasonal variation accounts for the single largest percentage of the variance in most coral isotope, trace, and minor element records. In regions such as Japan and the Gala´pagos with distinct SST seasonality, the annual periodicity in d18O, Sr/Ca, Mg/Ca, and U/Ca is pronounced. In regions heavily affected by monsoonal rains such as Tarawa Atoll, the seasonal variation in d13C is regular and pronounced. D14C also has an annual periodicity in coral from regions with a strong seasonal upwelling regime such as is seen in the Gala´pagos (see next section). The strength and duration of the upwelling season is reflected in the length and degree of D14C decrease in the coral record. Interannual-to-decadal Variation in Climate and El Ninõ Southern Oscillation (ENSO)

The second largest component of the variance in Pacific coral isotope records is associated with the interannual-to-decadal variation in the El Ninõ


342


9

7 1 3

8

2

4 6

5

Figure 3 Map indicating the approximate locations of current paleoclimate research. The coral sites involve the work of many investigators and may be incomplete. The d18O records from the numbered sites are shown in Figure 5 and are discussed in the text. , sites with records longer than 120 years (most are published); J, sites where cores have been recovered and data collection is underway. (Reproduced from Gagan et al. (2000), with permission from Elsevier Science.)

Southern Oscillation (ENSO)1. Several of the longer Pacific d18O coral records reveal that the frequency of ENSO has changed on decadal timescales over the past few centuries. Over the last 300 years, the dominant mode of ENSO recorded by a Gala´pagos coral has been at 4.6 years. However, during that time period there have been shifts in that mode from

1 ENSO refers to the full range of variability observed in the Southern Oscillation, including both El Ninõ and La Ninã events in the Pacific. The Southern Oscillation Index (SOI) is a measure of the normalized difference in the surface air pressure between Tahiti, French Polynesia and Darwin, Australia. Most of the year, under normal seasonal Southern Oscillation cool phase conditions, easterly trade winds induce upwelling in the eastern equatorial Pacific and westward near-equatorial surface flow. The westward flowing water warms and piles up in the western Pacific creating a warm pool and elevating sea level. The wind-driven-transport of this water from the eastern Pacific leads to an upward tilt of the thermocline and increases the efficiency of the local trade-winddriven equatorial upwelling to cool the surface resulting in an SST cold tongue that extends from the coast of South America to near the international date line. Normal seasonal Southern Oscillation warm-phase conditions are marked by a relaxation of the zonal component of trade winds, reduced upwelling, and a weakening or reversal of the westward flowing current coupled with a deepening of the thermocline in the eastern equatorial Pacific Ocean, and increased SST in the central and eastern equatorial Pacific. This oscillation between cool and warm phases normally occurs annually. Exaggerated and/or prolonged warm-phase conditions are called El Ninõ events. They usually last 6–18 months, occur irregularly at intervals of 2–7 years, and average about once every 3–4 years. The SOI is low during El Nin˜ events. Exaggerated and/ or prolonged ENSO cool phase conditions are called La El Ninã events. They often follow El Nin˜ events (but not necessarily). La El Ninã events are marked by unusually low surface temperatures in the eastern and central equatorial Pacific and a high SOI. For a detailed description of ENSO, see El Ninõ Southern Oscillation (ENSO).

4.6 to 7 years during the 1600s, 3–4.6 years from 1700–1750, and 3.5 years from 1800–1850. These major shifts in ENSO frequency may indicate major reorganizations in Pacific climate at various intervals over time. A 101-year long d18O record from Clipperton Atoll reveals a pronounced period of reduced ENSO frequency from B1925 to 1940 suggesting a reduced coupling between the eastern and western Pacific. At Clipperton, decadal timescale variability represents the largest percentage of the variance in d18O and appears to be related to the processes influencing the Pacific Decadal Oscillation phenomenon (PDO)2. Another component of ENSO variability recovered from coral d18O records is the shift in rainfall patterns during El Ninõs associated with: (1) the migration of the Indonesian Low pressure cell to the region of the date line and the equator in the western Pacific, and (2) the northern migration of the intertropical convergence zone (ITCZ) in the eastern Pacific. Eastward migration of the Indonesian Low results in decreased precipitation in the Indian Ocean and increased precipitation in the western and central Pacific. These phenomena are reflected in the d18O record of Seychelles and Tarawa Atoll corals, respectively. Decadal variability in the Seychelles record suggests that regional rainfall variability may

2 The PDO appears to be a robust, recurring two-to-three decade pattern of ocean–atmosphere climate variability in the North Pacific. A positive PDO index is characterized by cooler than average SST in the central North Pacific and warmer than average SST in the Gulf of Alaska and along the Pacific Coast of North America and corresponds to warm phases of ENSO. The reverse is true with a negative PDO.



originate from the ocean. Decadal variability in a 280-year d18O record from a Panamanian coral indicates decadal periods in the strength and position of the ITCZ. Changes in the decadal variability of coral skeletal D14C reveals information about the natural variability in ocean circulation, water mass movement and ventilation rates in surface water. Biennial-to-decadal shifts in D14C between 1880 and 1955 in a Bermuda coral indicates that rapid pulses of increased mixing between surface and subsurface waters occurred in the North Atlantic Ocean during the past century and that these pulses appeared to correlate with fluctuations in the North Atlantic Oscillation. In a post-bomb Gala´pagos coral record, abrupt increases in monthly D14C values during the upwelling season after 1976 suggest a decadal timescale shift in the vertical thermal structure of the eastern tropical Pacific (Figure 4). The decadal variability in D14C in the Bermuda and Gala´pagos records are testimony to the power of coral proxy records to provide information about ocean circulation patterns. Additional D14C records from Nauru, Fanning Island, Great Barrier Reef, Florida, Belize, Guam, Brazil, Cape Verde, French Frigate Shoals, Tahiti, Fiji, Hawaii and a few other locations are either published or in progress. As the number of coral D14C records increases, our understanding of the relationship

343

between climate and ocean circulation patterns will also increase. Some decadal-to-centennial trends in climate are consistent among many of the longer d18O coral records. For example in Figure 5, all six records longer than 200 years show a cooler/dryer period from AD 1800 to 1840. Cooling may be related to enhanced volcanism during this period. Following this cooler/ dryer interval, four of the six records show shifts towards warmer/wetter conditions around 1840– 1860 and five of the six show another warming around 1925–1940. These abrupt shifts towards warmer/wetter conditions detected in corals from a variety of tropical locations suggest that corals may be responding to global climate forcing. Long-term Trends in Climate

There are three major long-term trends observed in several coral records: (1) a prolonged cool phase prior to 1900 generally consistent with the Little Ice Age; (2) a gradual warming/freshening trend over the past century; and (3) evidence of increased burning of fossil fuels. First, in three of the four longest d18O coral records the cool/dry period of the Little Ice Age is observed from the beginning of their respective records, up to the mid to late 1800s (Figure 5). However, the lack of this cool/dry period in the Gala´pagos coral indicates that the Little Ice Age

80 14

60

Δ C ± 4‰

40

14

Δ C‰

20 0 _ 20 _ 40

El Nino

_ 60 _ 80 _ 100 1955

1960

1965

1970

1975

1980

1985

Date

Figure 4 Gala´pagos coral D14CD record from 1957 to 1983. El Ninõs are indicated by the shaded bars. , upwelling maxima; J, nonupwelling season. Dashed lines indicated linear trend in the upwelling and nonupwelling seasons. The seasonal variation in D14C is pronounced with high D14C during the nonupwelling season and low D14C values during the upwelling season. A shift in D14C baselines began in 1976. (Reproduced from Guilderson TP and Schrag DP (1998) Abrupt shift in subsurface temperatures in the tropical Pacific associated with changes in El Ninõ. Science 281: 240–243; with permission from the American Association for the Advancement of Science.)


344


1. Panama 2. Galapagos

Warmer / Wetter

3. Tarawa

Coral O‰

4. Vanuatu

18

5. New Caledonia

Cooler / Drier

6. Great Barrier Reef

_1 ‰

7. Phillippines

8. Seychelles

9. Red Sea

1600

1700

1800

1900

2000

Year Figure 5 Annual mean coral d18O records in the Pacific and Indian Ocean region extending back at least for 100 years (locations of cores are indicated in Figure 3). Mean d18O values for each site indicated with a horizontal line. Abrupt shifts in d18O towards warmer/ wetter conditions indicated by black triangles. Data are from the World Data Center-A for Paleoclimatology, NOAA/NGDC Paleoclimatology Program, Boulder, Colorado, USA (http://www.ngdc.noaa.gov/paleo/corals.html) and the original references. Core details list locality, species name, record length, and original reference: 1, Gulf of Chiriqui, Panama, Porites lobata 1708–1984 Linsley BK, Dunbar RB, Wellington GM, Mucciarone DA (1994) A coral-based reconstruction of intertropical convergence zone variability over Central America since 1707. Journal of Geophysical Research 99: 9977–9994); 2, Urvina Bay, Gala´pagos, Pavona clavus and Pavona gigantea, 1607–1981 Dunbar RG, Wellington GM, Colgan MW, Glynn PW (1994) Eastern Pacific sea surface temperature since 1600 A.D.: the d18O record of climate variability in Gala´pagos corals. Paleoceanography 9: 291–315; 3, Tarawa Atoll, Republic of Kiribati, Porites spp., 1893–1989 Cole JE, Fairbanks RG, Shen GT (1993) Recent variability in the Southern Oscillation: isotopic results from Tarawa Atoll coral. Science 260: 1790–1793; 4, Espiritu Santo, Vanuatu, Platygyra lamellina, 1806–1979 Quinn TM, Taylor FW, Crowley TJ (1993) A 173 year stable isotope record from a tropical south Pacific coral. Quaternary Science Review 12: 407–418; 5, Amedee Lighthouse, New Caledonia, Porites lutea, 1657–1992 Quinn TM, Crowley TJ, Taylor FW, Henin C, Joannot P, Join Y (1998) A multicentury stable isotope record from a New Caledonia coral: Interannual and decadal sea surface temperature variability in the southwest Pacific since 1657 A.D. Paleoceanography 13: 412–426; 6, Abraham Reef, Great Barrier Reef, Australia, Porites australiensis, 1635–1957 Druffel ERM, Griffin S (1993) Large variations of surface ocean radiocarbon: evidence of circulation changes in the southwestern Pacific. Journal of Geophysical Research 98: 20 249–22 259; 7, Cebu, Philippines, Porites lobata, 1859–1980 Pa¨tzold J (1986) Temperature and CO2 changes in the tropical surface waters of the Philippines during the past 120 years: record in the stable isotopes of hermatypic corals. Berichte Reports, Gol.-Palaönt, Inst. Univ. Kiel, 12; 8, Mahe Island, Seychelles, Porites lutea, 1846–1995 Charles CD, Hunter ED, Fairbanks RG (1997) Interaction between the ENSO and the Asian monsoon in a coral record of tropical climate. Science 277: 925–928; 9, Aqaba, Red Sea, Porites sp., 1788–1992 Heiss GA (1996) Annual band width variation in Porites sp. from Aquaba, Gulf of Aquaba, Red Sea. Bulletin of Marine Science 59: 393–403; (Reproduced from Gagan (2000), with permission from Elsevier Science.)



effects may not have been uniform throughout the tropical oceans. Second, this cool phase was followed by a general warming/freshening of the global tropical ocean beginning during the nineteenth century (Figure 5). This overall warming/freshening trend is observed in seven of the nine records. The timing of the onset of this warming/freshening is consistent with the onset of industrialization and the consequent increases in greenhouse gases due to increased emissions from fossil fuel consumption. If the shift in d18O were solely due to increases in SST, it would be equivalent to an increase of 0.3–2.01C since 1800. Instrumental data indicate that the tropics only warmed by B0.51C since 1850. The influence of SSS on d18O is probably responsible for the difference and needs to be taken into account when interpreting d18O records. Although Sr/Ca ratios are thought to be unaffected by SSS, only a few shorter coral records are currently published. Until recently, Sr/Ca measurements were very time-consuming. With recently developed technology, the use of Sr/Ca as a paleothermometer proxy should increase. Two main limitations exist with the correct interpretation of decadal and long-term d18O trends: (1) an interdecadal cycle of unknown origin is commonly identified in long coral d18O records; and (2) longterm trends of increasing d18O are observed in some coral while other coral d18O records show a decreasing trend. Whether these trends are due to biological processes or are the result of gradual environmental changes (i.e., global warming) is unclear. Finally, evidence of increased fossil fuel emission into the atmosphere can be seen in the general decrease in D14C from 1850 to 1955 in shallow corals from the Atlantic and Pacific Oceans. This phenomenon, referred to as the Suess Effect, is mainly the result of 14C-free CO2 produced from combusted fossil fuel entering the atmosphere, the oceans and eventually, the coral skeleton (post-1950, coral D14C values skyrocketed as a result of 14C produced by thermonuclear bombs effectively swamping out the Suess effect).

Fossil Corals Fossil corals provide windows into past climate. Records covering several decades to centuries offer the opportunity to compare the same three components (seasonal, interannual-to-decadal, and long term) in climate in the distant past to the present. A 3.0 million-year-old south-western Florida coral reveals a seasonal d18O derived temperature pattern

345

similar to today but B3.51C cooler. A North Sulawesi, Indonesian coral indicates that 124 000 years BP the variability in ENSO was similar to modern ENSO frequency from 1856 to 1976. However, the shift in ENSO frequency observed in modern records after 1976 is not found in the fossil coral record nor in pre-1976 instrumental records. This suggests that the current state of ENSO frequency is outside of the natural range of ENSO variability. Perhaps anthropogenic effects are having an effect on ENSO frequency. Finally, long-term changes in climate can also be reconstructed from fossil coral records. A series of coral records from Vanuatu indicate that B10 300 years BP the south-western tropical Pacific was 6.51C cooler than today followed by a rapid rise in temperature over the subsequent 15 000 years. This rapid rise in temperature lags the post-Younger Dryas warming of the Atlantic by B3000 years suggesting that the mechanism for deglacial climate change may not have been globally uniform. How seasonal, decadal (ENSO) and longterm climate changes varied in the distant past throughout the tropics can be addressed using fossil coral records and can offer us a better idea of the natural variability in tropical climate over geologic time.

Deep-sea Corals Deep sea corals do deposit calcium carbonate exoskeleton but do not contain endosymbiotic zooxanthellae and are not colonial. Their isotopic and trace mineral composition reflects variation in ambient conditions on the seafloor. Although this does not directly reflect changes in climate on the surface, ocean circulation patterns are tightly coupled with atmospheric climatic conditions. Understanding the history of deep and intermediate water circulation lends itself to a better understanding of climate. For example, the origin of the Younger Dryas cooling event (13 000 to 11 700 years BP) has recently been attributed to a cessation or slowing of North Atlantic deep water formation and subsequent reduction in heat flux. Isotopic evidence from deep-sea corals suggests that profound changes in intermediatewater circulation also occurred during the Younger Dryas. Other studies of deep-sea corals show rapid changes in deep ocean circulation on decadal-tocentennial timescales at other intervals during the last deglaciation. Reconstructing intermediate and deep ocean circulation patterns and their relationship to climate using isotopic, trace element and minor element records in deep sea coral promises to be an expanding line of paleoclimate research.


346


Discussion

Acknowledgments

The geochemical composition of coral skeletons currently offers the only means of recovering multicentury records of seasonal-to-centennial timescale variation in tropical climate. d18O-derived SST and SSS records are the workhorse of coral-based paleoclimate reconstructions to date. Improved methodologies are now making high resolution, multicentury Sr/Ca records feasible. Since Sr/Ca is potentially a less ambiguous SST recorder, coupling Sr/Ca with d18O records could yield more reliable SST and SSS reconstructions. d13C as a paleorecorder of seasonal variation in cloud cover and upwelling is also gaining credibility. However, more experimental research needs to be done before d13C records can be used more widely for paleoclimate reconstructions. Coral D14C records are highly valued as an ocean circulation/ventilation proxy. Increasing numbers of high-resolution D14C coral records are being published shedding invaluable new light on links between climate and ocean circulation processes. Coral trace and minor element records are also becoming more common and can add critical information about past upwelling regimes, wind patterns, pH, river discharge patterns, and SST. The growing number of multicentury coral oxygen isotope records is yielding new information on the natural variability in tropical climate. Eastern equatorial Pacific corals track ENSO-related changes in SST and upwelling. Further west, coral records track ENSO-related changes in SST and SSS related to the displacement of rainfall associated with the Indonesian Low. Decadal timescale changes in ENSO frequency and in ocean circulation and water mass movement detected in d18O and D14C records, respectively, indicate a major reorganization in Pacific climate at various intervals over time. Long-term trends in coral oxygen isotope records point to a gradual warming/freshening of the oceans over the past century suggesting that the tropics are responding to global forcings. Although the coral-based paleoclimate records reconstructed to date are impressive, much work remains to be done. It is necessary to develop multiple tracer records from each coral record in order to establish a more comprehensive reconstruction of several concurrent climatic features. In addition, replication of long isotopic and elemental records from multiple sites is invaluable for establishing better signal precision and reproducibility. Coupled with fossil and deep-sea coral records, coral proxy records offer a comprehensive and effective means of reconstructing tropical paleoclimates.

I thank B Linsley, E Druffel, T Guilderson, J Adkins and an anonymous reviewer for their comments on the manuscript. I thank the Henry and Camille Dreyfus Foundation for financial support.

See also Coral Reefs. El Ninõ Southern Oscillation (ENSO). Pacific Ocean Equatorial Currents.

Further Reading Beck JW, Recy J, Taylor F, Edwards RL, and Cabioch G (1997) Abrupt changes in early Holocene tropical sea surface temperature derived from coral records. Nature 385: 705--707. Druffel ERM (1997) Geochemistry of corals: proxies of past ocean chemistry, ocean circulation, and climate. Proceedings of the National Academy of Sciences of the USA 94: 8354--8361. Druffel ERM, Dunbar RB, Wellington GM, and Minnis SS (1990) Reef-building corals and identification of ENSO warming episodes. In: Glynn PW (ed.) Global Ecological Consequences of the 1982–83 El Ninõ – Southern Oscillation pp. 233–253 Elsevier Oceanography, Series 52. New York: Elsevier. Dunbar RB and Cole JE (1993) Coral records of oceanatmosphere variability. NOAA Climate and Global Change Program Special Report No. 10, Boulder, CO: UCAR. Dunbar RB and Cole JE (1999) Annual Records of Tropical Systems (ARTS). Kauai ARTS Workshop, September 1996. Pages workshop report series 99–1. Fairbanks RG, Evans MN, Rubenstone JL et al. (1997) Evaluating climate indices and their geochemical proxies measured in corals. Coral Reefs 16 suppl.: s93– s100. Felis T, Pa¨tzold J, Loya Y, and Wefer G (1998) Vertical water mass mixing and plankton blooms recorded in skeletal stable carbon isotopes of a Red Sea coral. Journal of Geophysical Research 103: 30731-30739. Gagan MK, Ayliffe LK, Beck JW, et al. (2000) New views of tropical paleoclimates from corals. Quaternary Science Reviews 19: 45--64. Grottoli AG (2000) Stable carbon isotopes (d13C) in coral skeletons. Oceanography 13: 93--97. Linsley BK, Ren L, Dunbar RB, and Howe SS (2000) El Ninõ Southern Oscillation (ENSO) and decadal-scale climate variability at 101N in the eastern Pacific from 1893 to 1994: a coral-based reconstruction from Clipperton Atoll. Paleoceanography 15: 322--335. http:// pangea. stanford. edu/Oceans/ARTS/arts_report/arts_ report_ home. html



Hudson JH, Shinn EA, Halley RB, and Lidz B (1981) Sclerochronology: a tool for interpreting past environments. Geology 4: 361--364. NOAA/NGCD Paleoclimatology Program http://www.ngdc. noaa.gov/paleo/corals.thml Shen GT (1993) Reconstruction of El Ninõ history from reef corals. Bull. Inst. fr. e´tudes andines 22(1): 125--158. Smith JE, Risk MJ, Schwarcz HP, and McConnaughey TA (1997) Rapid climate change in the North Atlantic during the Younger Dryas recorded by deep-sea corals. Nature 386: 818--820.

347

Swart PK (1983) Carbon and oxygen isotope fractionation in scleractinian corals: a review. Earth-Science Reviews 19: 51--80. Weil SM, Buddemeier RW, Smith SV, and Kroopnick PM (1981) The stable isotopic composition of coral skeletons: control by environmental variables. Geochimica et Cosmochimica Acta 45: 1147--1153. Wellington GM, Dunbar RB, and Merlen G (1996) Calibration of stable oxygen isotope signatures in Gala´pagos corals. Paleoceanography 11: 467--480.


PATCH DYNAMICS K. L. Denman, University of Victoria, Victoria, BC, Canada J. F. Dower, 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 4, pp 2107–2114, & 2001, Elsevier Ltd.

Introduction Once out of sight of land, the vastness and apparently unchanging nature of the open ocean might lead one to think that the plankton would be distributed evenly in space. In fact, this is rarely the case. Rather, planktonic organisms are generally distributed unevenly, in clumps of all shapes and sizes usually referred to as patches. As oceanographic platforms and various sensors for observing plankton have improved, we have discovered that plankton patches exist on scales of less than a meter (microscales) right up to scales of hundreds of kilometers (called the mesoscale) that are characteristic of the oceanic equivalent of storms. We have also learned that the two main groups of plankton – phytoplankton (the microscopic plants in the ocean) and zooplankton (the small animals that usually feed on phytoplankton) – are not patchy in the same way. Rather, zooplankton seem to be organized into many more patches at smaller scales, such that their distributions sometimes approach being random. This characteristically smaller patch size of zooplankton is believed to reflect their greater ability for swimming, swarming, or other directed motions, perhaps in response to patchiness in their food sources or other environmental cues. Similarly, the spatial distribution of larval fish, which have an even greater capability for directed motion and are capable of swimming and orienting themselves in groups in response to cues in their environment, is usually even more patchy. What does all this mean? What causes plankton patchiness? Are the causes physical, biological, or both? Are there ecological advantages or disadvantages to plankton patchiness? Does such patchiness promote or interfere with the transfer of energy and biomass from phytoplankton to zooplankton and larval fish and on to higher trophic levels? The related issue of how small-scale physical processes influence the biology of planktonic organisms is addressed elsewhere (see Small-Scale Physical

348

Processes and Plankton Biology). There the focus is on how individual predators interact with individual prey within their physical environment. Observations of interactions between individual planktonic organisms are possible primarily in the laboratory. Here we consider patches as being comprised of groups of individual planktonic organisms: how they are actually observed by oceanographers in the ocean, how we then analyze such observations and assess their potential significance to food web dynamics. The two approaches (patch-based versus individual-based observations) are connected as follows. Better understanding and prediction of patch dynamics and their importance to food web dynamics requires understanding of the processes by which individual organisms interact with each other and with the fluid flow. Generally, we are not yet able to make detailed observations of interactions between individual organisms in the ocean, and so we must rely on laboratory observations of individual organisms. To put these individual-based observations into an appropriate ecological context, however, it is also necessary to formulate or generalize how such individual interactions contribute to the behavior and responses at the level of a patch, which may contain millions of individuals. The other point of contact between the two approaches is to consider the case of an individual predator interacting with a patch of prey. One might reasonably ask whether there is some threshold difference in size between predator and prey at which an individual predator begins to interact with a patch of prey organisms (e.g., a gray whale feeding on a swarm of mysids) rather than with individual prey. Because the observational technologies and analytic techniques used to study plankton patchiness are both complex and fascinating in their own right, it is wise during the discussion to follow to remain conscious of the underlying question being addressed: What are the implications of plankton patchiness to foodweb dynamics?

History The existence of plankton patchiness was known as far back as the 1930s from observations that net tows taken simultaneously from two sides of the ship did not capture identical or often even similar collections of zooplankton. Subsequent studies and statistical analyses into the 1960s determined that this observed


PATCH DYNAMICS

variability was not random but had structure, and that patches or clumps existed at sizes of less than a meter to sizes spanning hundreds of kilometers. In the 1970s continuous-flow fluorometers allowed highresolution measurements of chlorophyll fluorescence (an indicator of the concentration of phytoplankton) continuously and simultaneously with comparable temperature observations. Similarities in the statistical distributions of fluorescence versus temperature quickly led to the realization that plankton patchiness was in large part controlled by physical turbulence. However, the fact that fluorescence distribution patterns did not exactly mimic that of the temperature field signaled that phytoplankton are not just ‘passive tracers’ of the flow, but that they are capable of changing their spatial distribution by growing. Comparable high-resolution observations of zooplankton distributions were more difficult to obtain, but their statistical properties differed even more from those of the temperature field, suggesting that, in addition to growth, zooplankton are also not passive tracers of the flow field because of their ability to swim and aggregate. In the 1980s, the development of new observational techniques – moored fluorometers, instrumented multiple-net systems, multifrequency acoustics, optical plankton counters, and satelliteborne multichannel remote color sensors (to name but a few) – allowed a three-dimensional picture of the structure of patchiness and its relationship to the physical environment to emerge. In contrast, during the 1990s there was a shift toward the development of observational techniques that focused on individual organisms, usually under laboratory conditions. An exception to this trend was the development of the video plankton recorder (VPR), a towed oceanographic instrument that allowed in situ detection, identification, and counting of individual planktonic organisms, albeit with great manual labor in the analysis. Together with sensors of related oceanographic properties, the VPR allows construction of three-dimensional images of plankton patchiness simultaneously with a three-dimensional description of the immediate ocean environment. In concert with the continuing emergence of an observational description of plankton patchiness in the context of the physical environment, there has also been a parallel development in our conceptual understanding of (i) the factors that create and control patchiness and (ii) the significance of patchiness to planktonic food web dynamics. Taking phytoplankton patchiness as an example, the 1970s view was that patch formation and decay was a balance between phytoplankton growth (which would act to stabilize the structure of a patch) versus small-scale

349

turbulent mixing and diffusion (which combine to erode patch structure). The idea was that, in patches larger than some characteristic size, phytoplankton growth would win out over mixing and diffusion and the patch would be stable, but that, for patches smaller than this characteristic size, erosion by smallscale turbulent diffusion would be more efficient and the patch would be smoothed out. Around 1980, however, there was a realization that grazing by zooplankton could also affect the patchiness of phytoplankton. In addition, it was recognized that zooplankton can respond to variability in both phytoplankton abundance and physical structure by actively swimming to different positions to graze selectively in certain areas. Although recognized as an important process, the actual means by which zooplankton respond to cues in their environment and prey (both in terms of changing their distribution through swimming and by varying the intensity of their grazing) remains largely unknown. Since the 1980s we have come to appreciate that the action of ocean currents and turbulence is much more complex than simply smoothing out smallerscale structure through diffusive mixing. We now think more in terms of ‘geostrophic turbulence’, where variable currents and eddies (with structure from hundreds of meters down to submeter scales) create, distort, and evolve patches in various complex ways. Variable currents interacting with gradients in phytoplankton concentration can stretch, distort, and sharpen the boundaries of patches, sometimes creating very clear boundaries (often referred to as fronts) that separate patches from their background. Depending on the associated physical motions, these sharp frontal boundaries can increase or decrease the transfer of planktonic organisms across the front, or they can expose those predators that are capable of swimming to prey either of different concentration or of different community structure. We have also come to realize that perhaps the most important patches in terms of food web dynamics are those that tend either to recur or to persist at certain locations owing to features of coastal geometry or bottom topography interacting with the flow field. In fact, it is the very ‘predictability’ of these patches that makes them so important. These can be sites of enhanced tidal mixing, of increased retention of zooplankton and fish larvae due to convergent circular currents, or of increased dispersal due to the enhancement or divergence of currents. That such sites are important for the transfer of energy to higher trophic levels is evidenced by the fact that they are often sites where larger fish, as well as sea birds and marine mammals also congregate.


350

PATCH DYNAMICS

20

60 50 40 30

10

2.0

5.0 4.0 3.0

1.0

0.2

0.5 0.4 0.3

0.1

0.05 0.04 0.03 0.02

0.01

3

Chlorophyll a Concentration (mg/m )

(A)

Figure 1 (A) Derived phytoplankton pigment concentration from the SeaWiFS satellite on 8 March 2000. The image (about 800 km in the vertical dimension) shows phytoplankton patchiness (red indicates highest pigment concentration) associated with filaments of nutrient-rich upwelled waters flowing offshore from the western coast of South Africa. (B) Sea surface temperature for the same day. (Images provided by the SeaWiFS Project, NASA/Goddard Space Flight Center and ORBIMAGE, courtesy of Gene Feldman (NASA) and Scarla Weeks (OceanSpace).)


PATCH DYNAMICS

Hondeklipbaii

Groenrivier

Soulrivier

Olifants River Donkin Bay Lamberts Bay Elandsbaii

Cape Columbine

Yserfontein Dassen Island

Robben Island Hout Bay Slangkoppunt Cape Hangklip Danger Point Cape Agulhas

(B) Figure 1 Continued


351

352

PATCH DYNAMICS

Our understanding of the importance of patchiness to food web dynamics has also evolved accordingly. Typical mean concentrations of prey organisms in the ocean, when compared with feeding studies conducted under laboratory conditions, suggest that predators generally should be starving. Other observations of predators, such as determining the contents of their guts, seem to indicate that generally they somehow get enough to eat. Initially, patchiness was evoked to solve this dilemma in the following way. Predators were thought to be capable of locating patches of their prey, then feeding to excess while in patches. Thereby, on average they were assumed to receive enough nutrition by locating patches and feeding in them. Indeed, empirical studies have demonstrated that upon encountering a patch of prey, predators that normally adopt a search pattern approximating a ‘random walk’ often respond by increasing their turning frequency, a behavioral adaptation apparently designed to keep the predator in the prey patch as long as possible. Aided largely by the evolution of laboratory studies of interactions between individual predators and their prey organisms, we have further refined our understanding of how patchiness affects food web transfers by considering the role of turbulence as an active agent in feeding interactions. The basic idea is that an increase in the intensity of turbulence should increase the contact rate of predators with prey organisms (i.e., as turbulence increases, a predator should randomly encounter more prey organisms per unit time). Thus, it has been suggested that increased turbulence should allow a predator to capture more prey organisms in a given time. However, as the contact rate increases with increasing turbulence, the contact time (i.e., the length of time when the predator and prey are in close proximity to each other) decreases. Since predators need a finite time to respond to the presence of a prey item before they can capture it, we might therefore expect that as the contact time decreases the capture success of the predator will also decrease. Some model simulations suggest that, as turbulence increases, feeding success will increase, but that at some point the turbulence becomes so disruptive that feeding success starts to decrease. Whether the predator is inside or outside a patch might be expected to change the threshold at which turbulence moves from enhancing to decreasing grazing success. In addition, high turbulent intensities will also tend to spread out or disperse patches of prey. Laboratory experiments have yet to provide definitive evaluation of these theories or hypotheses, and field observations are even more ambiguous. To add further complexity, some zooplankton are

raptorial (e.g., carnivorous copepods or larval fish), seeking out and capturing individual prey organisms, while other zooplankton (notably herbivorous copepods) are filter feeders, passing water through their appendages while filtering whatever nutritional particles they can out of the water. Still others, such as jellyfish and ctenophores, are ‘contact predators’, which capture prey that literally bump into their tentacles. Whether these different types of predators are all affected the same way by turbulence remains to be seen. As we enter the new century, we are endeavoring to move the individual-based laboratory techniques into the field, and at the same time are developing the analytic and modeling tools to integrate our understanding of the behavior of individuals to the level of patches and populations.

Observations and Analysis Techniques It is not yet possible in the sea to obtain accurate ‘snapshots’ of the distribution of plankton over any large volume. The basic problem is that of trying to construct a synoptic (yet necessarily static) picture of the distribution of plankton, many of which display species-specific behaviors and which inhabit a threedimensional moving fluid (which often moves in different directions at different depths). Technologies that can count individual organisms (e.g., cell cytometers for phytoplankton and video recorders for zooplankton) have both a limited range and a narrow field of view. A sufficiently dense network of sensors is rarely possible because of resource, logistical, and intercalibration limitations. A mapping strategy, whereby a grid is followed, requires a finite time for execution, during which the distribution of organisms is itself changing, blurring the ‘snapshot’ to an unknown degree. Within these constraints, oceanographers employ two basic types of sampling to maximize the information that they can obtain about the patchy nature of plankton distributions. The first strategy is to tow sensors along horizontal or vertical straight lines, or along oblique ‘sawtooth’ paths, over a grid of transects in the horizontal plane. The second strategy involves remote sensing from satellites (Figure 1) or airplanes, which can give nearly ‘synoptic’ (snapshot) views of near surface horizontal distributions, but with little or no information on changes with depth. Such sensors include stimulated chlorophyll fluorescence, electrical conductivity, various types of acoustics, optical plankton counters, video plankton recorders, and various combinations of instrumented nets. Some of these sensors are used to count individual


PATCH DYNAMICS

353

Table 1 General properties of the most common sensors used to observe plankton patchiness. Size range of typical phytoplankton organisms is 1 mm to 20 mm; size range of typical zooplankton organisms is 1 mm to 1 cm Sensor

Organism

Range (horizontal)

Determine size and identity

Fluorometer Satellite color sensor Towed nets Optical plankton counter Acoustic sounding Video plankton recorder

Phytoplankton Phytoplankton Zooplankton Zooplankton Zooplankton Zooplankton and phytoplankton chains

0.1–100 km 1–1000 km 100 m–10 km 1 m–100 km 10 m–100 km 10 cm–100 km

No No Yes Size Approximate size Yes

organisms, while others are designed to estimate plankton biomass in a given sampling volume. Table 1 summarizes the most common sensors in terms of their spatial resolution and range, their ability to detect and count individual organisms, and the target planktonic group. Usually, oceanographers employ a combination of these sensors to obtain the best picture of the plankton distributions (Figure 2). Increasingly, we want simultaneous high-resolution observations of both phytoplankton and zooplankton as well as related environmental variables (e.g., incident light, temperature, salinity, nutrients, turbulent mixing, local flow phenomena, etc.) in an attempt to identify and quantify mechanisms relating causes and consequences of patchy distributions in the planktonic ecosystem. Of course, logistic problems aside, the problem with attempting to tie all this information together is that the plankton distribution observed at some point in time may correlate better with environmental conditions at some time in the recent past (hours or days ago) rather than with the environmental conditions at the time of capture. This delay may occur because, being living organisms, plankton require time to respond to changes in their physical environment. For planktonic organisms that reproduce quickly (e.g., phytoplankton and microbes) this ‘lag time’ might be of the order of hours to a day. However, for organisms that take longer to grow and/or reproduce (e.g., zooplankton and larval fish) the lag time between a favorable change in the environment and a measurable response in the plankton may be of the order of days to weeks, or even longer. Thus, although we now have the technology to sample both physics and biology at very high resolutions in the ocean, we must be careful how we interpret the masses of data we collect. Direct observations of patchiness, such as visual images from satellites and video plankton recorders, can be compelling, especially when they point to

previously unknown or unexpected phenomena. But how do we proceed from ‘pretty pictures’ to mechanisms and then to prediction? Usually, we try to apportion the observed variability or variance into characteristic size scales: are there larger changes over kilometers or over meters? In addition, to determine interactions between different organisms or groups of organisms, or between plankton and their environment, it is necessary to perform various types of correlative analyses: are changes in phytoplankton concentrations associated mostly with changes in populations of zooplankton or with changes in temperature or nutrients? For observations obtained in a regular fashion (at regular intervals in time or space), spectrum analysis has proved to be useful, both to apportion variance according to patch size and to identify correlative structures between different variables. More recently, fractal analysis, which requires fewer assumptions regarding the ‘wellbehavedness’ of the statistics of the observations, promises to reveal more insights into the spatial variability of plankton.

Food Web Implications From the perspective of food web dynamics, it is interesting to consider whether plankton patchiness can be caused by predator–prey interactions. Certainly, empirical observations and some field-based work (particularly in freshwater ecosystems) have demonstrated that, under certain conditions, predation can alter plankton community structure. However, although this concept is appealing, the bulk of observational evidence in marine ecosystems strongly suggests that most plankton patchiness results from environmental factors. Examples include locally enhanced phytoplankton growth in regions of favorable light and nutrient conditions; turbulent mixing eroding patches; variable circulation patterns


354

PATCH DYNAMICS

0 10

3.0

20 2.5

Depth (m)

30 2.0 40 1.5

50 60

1.0

70 0.5 80 0.0

90 10

20

30 40 Distance (km)

50

60

Figure 2 Distribution, abundance, and two representative images of the copepod Calanus finmarchicus along a transect across the Great South Channel east of Woods Hole, USA, with the video plankton recorder. Lines represent contours of constant temperature (0.51C intervals). (From Gallager SM et al. (1996) High-resolution observations of plankton spatial distributions correlated with hydrography in the Great South Channel, Georges Bank. Deep-Sea Research II 43: 1627–1663 ^ 1996, with permission of Elsevier Science.

creating and distorting spatial structure in organism distributions; zooplankton swarming in response to environmental cues; or phytoplankton altering their buoyancy in response to variability in light, gravitational, or chemical cues. To be fair, however, it should also be noted that observations in the ocean have generally been inadequate to detect without ambiguity evidence of patchiness generated through grazing interactions. In fact, such ‘top-down’ regulation of planktonic concentration patterns is expected on ecological grounds. Thus, its existence should not yet be dismissed on the basis of contemporary observational capabilities. Evidence is more clear with regard to flow patterns that can cause nutrient injection and/or the retention or even concentration of planktonic organisms for sufficient duration that ecologically significantly enhanced feeding by predators on their prey can occur. Examples of such flow patterns are fronts, rotating rings or eddies, and regular or recurring tidal patches

associated with headlands, subsurface banks, or seamounts. Persistent or regularly recurring fronts, eddies, or convergent zones may also induce behavioral adaptation in predators, especially fish populations, such that they return to sites of recurring enhanced prey. For ‘passive’ plankton without the capability for directed motion, it is not always easy to determine whether the higher concentrations or numbers of organisms associated with these flow features result from locally enhanced growth or from concentration by convergent flow. Generally, concentration can only result when there is convergent flow and the organisms have a density different from that of the local fluid. In recent years, our thinking has generally moved away from the concept that random encounters of patches of predators and prey result in enhanced food web transfers, and toward the concept that persistent or recurring patches (usually associated with certain flow or mixing regimes caused by


PATCH DYNAMICS

particular geographical features) have the most potential to affect food web dynamics. The exception to this is the concept that turbulent motions alter the ‘contact rate’ between individual predator and prey organisms, an area treated in more detail in SmallScale Physical Processes and Plankton Biology.

What Remains ‘Unknown’? To what extent has our understanding of the importance of plankton patchiness to food web dynamics been determined by sampling inadequacies? The question can only be answered in the context of how changes in our conceptual understanding have paralleled changes in our sampling capabilities over the last several decades. As sampling capabilities have improved both under laboratory conditions and at sea, we have started to move away from considering patch–patch food web interactions. Increasingly we consider these interactions more in terms of individual predators interacting with individual prey, and determining whether ‘contact rate’ between predator and prey organisms is regulated mostly by the turbulent flow field or by the actions of individual predators in reponse to their sensing of prey organisms. At sea, even guided by the results of laboratory observations, we can seldom observe individual phytoplankton organisms, so conceptual models have evolved toward consideration of individual predators interacting with patches of prey. Is this the ecologically meaningful view of plankton patchiness? To resolve our conceptual uncertainties we must improve our sampling capabilities such that we can observe individual prey and predator organisms simultaneously over the dimensions of observed patches. The challenge will then be to scale up from the observations of how individual organisms interact with their environment and with other individual organisms to an understanding of how populations of organisms collectively interact with their

355

environment and with other populations of both their predators and their prey.

See also Bioluminescence. Fish Larvae. Moorings. Plankton. Small-Scale Physical Processes and Plankton Biology.

Further Reading Davis CS, Flierl GR, Wiebe PH, and Franks PJ (1991) Micropatchiness, turbulence and recruitment in plankton. Journal of Marine Research 49: 109--151. Denman KL (1994) Physical structuring and distribution of size in oceanic foodwebs. In: Giller PS, Hildrew AG, and Raffaelli DG (eds.) Aquatic Ecology: Scale, Pattern and Process. Oxford: Blackwell Scientific. Levin SA, Powell TM, and Steele JH (eds.) (1993) Patch Dynamics, Lecture Notes in Biomathematics, 96. Berlin: Springer-Verlag. Mackas DL, Denman KL, and Abbott MR (1985) Plankton patchiness: biology in the physical vernacular. Bulletin of Marine Science 37: 652--674. Okubo A (1980) Diffusion and Ecological Problems: Mathematical Models. Berlin: Springer-Verlag. Platt T (1972) Local phytoplankton abundance and turbulence. Deep-Sea Research 19: 183--188. Platt T and Denman KL (1975) Spectral analysis in ecology. Annual Review of Ecology and Systematics 6: 189--210. Powell TM and Okubo A (1994) Turbulence, diffusion and patchiness in the sea. Philosophical Transactions of the Royal Society of London B 343: 11--18. Seuront L, Schmitt F, Lagadeuc Y, Schertzer D, and Lovejoy S (1999) Universal multifractal analysis as a tool to characterize multiscale intermittent patterns: example of phytoplankton distribution in turbulent coastal waters. Journal of Plankton Research 21: 877--922. Steele JH (ed.) (1978) Spatial Pattern in Plankton Communities. New York: Plenum Press.


PELAGIC BIOGEOGRAPHY A. Longhurst, Place de I’Eglise, Cajarc, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2114–2122, & 2001, Elsevier Ltd.

Introduction The pelagic organisms discussed here inhabit the epipelagic zone where biogeographic distribution of organisms is more complex than in the interior of the oceans. The epipelagic zone comprises the surface mixed layer, the pycnocline, and several hundred meters into the deeper water masses. In principle, at such depths, water is circulated by the surface currents throughout all the interconnecting basins and marginal seas of the global ocean. Observations of watermass properties, and of natural passive tracers such as the anomalously low-salinity water that passed around the Subarctic Gyre of the North Atlantic from 1961 to 1981, support this principle. Though some large fish actively migrate over distances comparable to those of migrating birds, epipelagic plankton and small nekton have insufficient mobility to counter the flow of ocean currents. Apparently, a population of such organisms must go where the water carries it, and this conclusion led early biogeographers to believe that most plankton species must be cosmopolitan. However, we know that plankton do not behave like passive tracers, although many are widely distributed. There is repetitive pattern in the distribution both of individual species and of associations of species. Only an understanding of the ecology and behavior of individual species will explain how such distributions are maintained, and how regional oceanography determines the characteristic composition of pelagic communities. To understand why individual species are distributed as we observe them to be, it may also be necessary to consider the progressive evolution of ocean basins during geological time. It was this historical aspect of their subject that attracted much of the attention of biogeographers in the past, but pelagic biogeography is now concerned with more than that.

The Pelagic Biota and Their Diversity The biogeography of the pelagic biota reflects two simple facts. First, although most basic life-forms of

356

metazoa originated in the ocean, relatively few of these invaded the dry land. And, second, the higher plants of the terrestrial flora have no representatives in the ocean, where almost all photosynthesis is performed by single-celled algae and cyanobacteria. Consequently, the marine fauna is relatively diverse at higher taxonomic levels but has far fewer individual species than the terrestrial fauna. Insect species are more than one thousand times more numerous than copepods because of the (literally) uncounted number of specific insect–plant associations. This illustrates the important generalization that the structural complexity of the terrestrial plants induces a great diversity of ecological niches compared with the poverty of niche-space in the relatively unstructured oceanic ecosystems. The coral reef exception proves the rule. As in terrestrial ecosystems, diversity in the epipelagic zone of the ocean responds indirectly to latitude. In polar oceans, where there is only a brief, light-limited pulse of primary production near midsummer, there are fewer species, of fewer sub-phyla or classes, than elsewhere (Table 1). In the central oceanic gyres where primary production rate is relatively invariant, both species and higher taxonomic units are the most diverse. Where upwelling occurs seasonally in warm seas, diversity is intermediate. Values for ecological measures of diversity, such as the Shannon–Weaver index, are similarly distributed. These observations conform to the ecological generality that diversity is greater in ecosystems of low than in those of high latitudes. Such analyses of diversity of the ocean biota are constrained by our uncertainty about the real numbers of marine species, and by the general difficulty of defining the ecologically significant lower taxonomic units of plankton. The 3000-odd species of marine plankton that have been formally described are Linnaean or ‘taxonomic’ species: that is, forms that are distinguishable on the basis of their morphology and to which a binomial (e.g., Calanus helgolandicus) is assigned. This classical approach does not describe the reproductively isolated populations that occupy disjunct parts of the total range of many species and that are the significant ecological units. These expressions of the ‘Rassenkreis’ of Rensch, or of ‘sub-species’ of the ‘biological’ species of Mayr and Julian Huxley, have been satisfactorily investigated in only a few marine organisms, but their existence may be crucial for the persistence of species of



0.23 3.28 1.98

0.62 1.39

11 11

0.22 0.67

8 12

7 11 11

0.06 1.39 0.85

MEDUS

7 10 13

N

2.66 7.88

0.75 1.78 8.51

0.59 2.11

0.15 0.45 2.61

SIPH

6.51 9.81

8.62 8.62 9.00

3.32 5.83

3.88 3.99 6.38

CHAET

0.79 0.64

0.47 0.47 1.10

0.39 0.43

0.23 0.54 0.85

POLY

0.78 0.48

0.00 0.00 0.01

7.07 6.46

0.00 0.47 0.14

CLD

80.10 62.95

86.32 80.45 72.31 0.78 0.82

0.56 0.81 0.54

0.28 4.98

3.05 3.05 2.61 64.66 37.51

67.58 67.58 33.13 5.22 4.25

4.57 4.57 2.58

0.00 0.00

0.00 0.00 0.00

0.07 0.02

0.00 0.02 0.15

11.79 10.67

7.66 7.66 30.21

1.02 1.17

0.52 2.77 3.02

COPEP AMPH MYSI EUPH Numbers of individuals (%)

Carbon equivalent biomass (%)

0.31 5.72

2.45 3.12 3.39

OSTR

1.11 15.26

0.13 0.13 5.02

0.27 3.15

0.02 0.33 1.30

PENAE

4.83 1.15

6.73 6.73 2.66

1.50 0.60

2.44 1.76 1.70

PTER

0.17 0.18

0.18 0.18 0.10

3.51 5.45

3.32 2.79 3.85

APPEN

0.80 5.80

0.03 0.03 2.89

0.46 5.15

0.02 0.52 2.10

THALI

0.00 0.00

0.00 0.00 0.01

0.05 0.10

0.00 0.12 0.66

DOLIO

Italic type emphasizes trends in community composition from high latitudes to low. N ¼ number of major taxonomic groups (classes, sub-phyla) comprising 40.5% of individuals or biomass, MEDUS ¼ Medusae, SIPH ¼ Siphonophora, CHAET ¼ Chaetognatha, POLYPolychaeta, CLD ¼ Cladocera, OSTRO ¼ stracoda, COPEP ¼ Copepoda, AMPH ¼ Amphipoda, MYSI ¼ Mysidacae, EUPH ¼ Euphausidae, PTER ¼ Pteropoda, APPEN ¼ Appendicularia, THALI ¼ Thaliacae, DOLIO ¼ Doliolida.

Oceanic Polar Westerlies Trades Coastal Westerlies Trades

Oceanic Polar Westerlies Trades Coastal Westerlies Trades

Biomes

Table 1 Percentage composition of zooplankton communities in the Polar, Westerlies, Trades and Coastal biomes, data from 4000 plankton samples, taken randomly over all oceans and sorted into 15 major taxonomic categories

PELAGIC BIOGEOGRAPHY 357

358

PELAGIC BIOGEOGRAPHY

plankton and micronekton. Until recently, our ability to locate such self-sustaining populations was limited to groups whose morphology was amenable to careful numerical analysis – as in the counting and relative spacing of pores in the exoskeleton of some copepods. Fortunately, the sequencing of mitochondrial DNA is now available to discriminate genetically isolated populations with clarity. Even today, because this has been done in very few cases, we remain unsure to what extent Linnaean species of plankton actually have the population structure of biological species. This information will accrue only slowly, but in the interim we may assume that it is the normal case for species that are distributed widely, for reasons we shall now discuss.

Population and Expatriation: the ‘Member–Vagrant’ Hypothesis The persistence of local populations may be the mechanism by which bisexual organisms in the plankton maintain a sufficient population density for successful reproduction, as Sinclair has suggested. In the absence of this rich population structure, passive dispersal would result in the encounter rate between individuals becoming too low for sexual reproduction to occur. Organisms with asexual reproduction, such as phytoplankon and perhaps some tunicates, have no such requirement, and consequently algal species are more cosmopolitan than metazoans. Persistence can be maintained only if each local population reproduces at a rate faster than it loses individuals by passive transport out of its retention (or reproductive) area. Since there are no absolute boundaries in the ocean, all retention areas will be more or less leaky, and individuals will be lost into the general circulation. Consequently, the entire population of any species of plankton comprises both members of the persistent population, and vagrants lost to it. So any sample of plankton will comprise abundant individuals of populations that are characteristic of the region (the dominants) where the sample is taken, together with many less abundant species (the vagrants) transported from other regions, some very distant. What mechanism ensures persistence of the local self-sustaining stocks, the dominants in plankton samples? It has been widely assumed by biogeographers that the oceanic gyral circulations are sufficiently closed that the conditions discussed above are generally satisfied; certainly, this assumption is valid for some gyral circulations. The central part of the North Atlantic Subtropical Gyre notoriously

retains a floating population of macro-algae (Sargassum muticans), and the gyre within the semienclosed Norwegian Sea retains a persistent population of Calanus finmarchicus. If we examine more typical situations, we find that simple gyral retention becomes a very incomplete explanation of how populations persist. It is common for disjunct populations to partition a single more-or-less closed gyre, as do species-pairs of North Atlantic copepods. In the cold limb of the Subarctic Gyre, Calanus glacialis and Metridia longa have their centers of distribution, while the warm limb is the habitat of C. helgolandicus and M. lucens. Conversely, the related C. finmarchicus occurs much more widely throughout the gyre. How are such patterns maintained? To the extent that flow-fields of ocean currents are laminar, then purposive activity on the part of the organisms must be invoked to explain their persistence. Topex-Poseidon images of the elevation of the sea surface show that flow is nowhere laminar but instead comprises a complex field of cyclonic and anticyclonic eddies having internal flow rates greater than that of the mean current. But because the whole eddy field is itself moving at the mean velocity of the gyral current, the eddies themselves cannot increase overall retention of passively transported biota, except where an individual eddy is captured by topography. On the other hand, population persistence requiring active behavior on the part of the retained organisms does occur. The seasonal migrations of Calanus finmarchicus between the near-surface and 500–1000 m maintain enough individuals in suitable advective trajectories within the subarctic gyre for centers of persistence to be maintained. Populations of Calanus spp. and Calanoides carinatus of upwelling regions persist by migrating down to the slower and even contrary flow below the newly formed pycnocline at the end of each upwelling event, thus avoiding longshore and offshore transport. At the onset of the next upwelling event, they are carried passively surfaceward and toward the coast. A similar pattern of seasonal vertical migration enables populations of macroplankton in the Southern Ocean to persist within their optimal latitudinal zone. Most significantly, do the almost ubiquitous diel migrations of many kinds of organisms also serve the same purpose? Planktologists are reluctant to question the current paradigm that diel vertical migration of oceanic zooplankton is primarily a response to predation (as it undoubtedly is in lakes), so this possibility has been largely neglected. But by crossing the shearzone within the pycnocline and passing 12 h within the slow or even contrary transport of the deeper



circulation, many diel migrants must significantly reduce their passive transport within the surface water. This is the concept of Alister Hardy that ‘vertical migration sets them striding through the sea with sevenleague boots,’ although perhaps we should regard vertical migration as a mechanism to increase persistence rather than to enhance dispersion.

Influences of Regional Oceanography on Ecology and Distribution of Organisms The seasonal cycle of the oceanic phytoplankton is not everywhere the same, each region having a characteristic pattern determined by only a small number of factors, of which the most important are the seasonal sequences (i) of mixing and stabilization of the water column, and (ii) of irradiance in the photic zone. In coastal regions and archipelagos, the effects of tides, topography, and river effluents may force stability, nutrient supply, and irradiance. The phytoplankton association characteristic of each region differs taxonomically, and with each is associated a characteristic community of zooplankton, both herbivores and predators. It is against this background of regionally characteristic plankton communities or associations that the geographical distribution of individual species must be considered. Several suggestions have been made to account for the distribution of characteristic plankton communities by reference to regional oceanography. The simplest concept is that each surface water mass is inhabited by a characteristic association of species, some of which may be selected as ‘indicators’ for the water mass. This model dominated biogeographical research for many decades, but is of limited interest now since it does not address the seasonal ecology of each characteristic association. In recent years this need has been addressed by various proposals for a geographic partition of the epipelagos; for instance, that there are six ‘local types’ or models of phytoplankton ecology associated with the oceanographic regimes of (i) the sub-tropical gyres, (ii) the eastern boundary currents, (iii) the Southern Ocean, and so on. A more satisfactory generalization is based simply on Sverdrup’s model of the seasonal cycle of phytoplankton. This yields the required general relationship between regional oceanography and plankton ecology. Only if the depth of mixing is shallower than the critical depth, where algal cells are lightlimited, can net growth be positive and a bloom occur. This model is applicable to all oceanographic situations and responds to such diverse processes as

359

Ekman pumping, the existence of a low-salinity ‘barrier layer’ in the thermally uniform surface layer, and tidally forced mixing over continental shelves. Although local forcing of mixed layer depth is usually assumed, the model also accommodates the consequences of geostrophic forcing of mixed layer depth, as occurs widely in the tropical ocean. The physical processes required for Sverdrup’s model are not distributed continuously, but have the discontinuities we observe in regional oceanography, from which has been derived a global partition into ‘biogeochemical provinces.’ This ecological geography of the epipelagic zone comprises four biomes in the sense of Odum, for whom biomes are the largest ecological units, each having characteristic climax vegetation. These are (i) the Polar biome, where the depth of the surface mixed layer is influenced by the presence of light, low-salinity water derived from ice melt; (ii) the Westerlies biome, where seasonal changes in mixed layer depth are locally forced by winds and irradiance; (iii) the Trades biome, where mixed layer depth may be largely a geostrophic response to distant wind forcing; and (iv) the Coastal biome, where the factors determining mixed layer depth are diverse and are more significant than the effect of latitude itself. Each of the oceanic biomes has a characteristic set of biota and a characteristic seasonal ecological succession. These suggestions recall the earlier scheme of Beklemishev for Polar, Gyral, Transition and Coastal biomes based not on ecology but on recurrent pattern that he noted in the distribution of plankton species. The boundaries between biomes, located at timevarying discontinuities in regional oceanography, are not symmetric between oceans. The Trades and Westerlies biomes meet at the relatively abrupt transition at the Subtropical Convergences from the strong tropical pycnocline to the weaker subtropical pycnocline. The Polar biome is bounded by the Polar Frontal Zones, where there is an abrupt poleward increase in the stability of the pycnocline. The Coastal biome is bounded by the shelf-break front, generally identifiable above the upper continental slope. We should not expect conditions within these four biomes to be everywhere uniform, and indeed Platt and Sathyendranath have introduced the concept of dynamic biogeochemical provinces, which are a finer partition but conceptually similar – ‘dynamic’ because their boundaries are allowed to vary at all temporal and spatial scales, and ‘biogeochemical’ because each has characteristic environmental and ecological conditions and harbors a characteristic plankton population. Using boundaries defined by frontal regions and other features of regional oceanography, 50-odd biogeochemical provinces (Figure 1) have been identified globally, each of


360


which can be individually assigned to one of eight simple models of the seasonal production cycle (Table 2). This is not the only possible partition and others will undoubtedly be suggested if the interpretation of satellite imagery, which is the essential and unique tool for such investigations, becomes more widely accepted by biological oceanographers.

Concordance of Biogeographic Patterns Examination of maps of the global distribution of plankton species (and very many such maps are available) led Dunbar to comment that the biogeographic method does not exist, or at least that there are as many methods as there are biogeographers. There are no agreed terminologies for the various types of distribution, or criteria for determining biogeographic boundaries. The problem is in part a consequence of undersampling and in part of uncertainty in the identification of specimens collected at sea, and finally reflects the fact that the motives of biogeographers have been very diverse. Only in relatively small parts of the ocean has sampling been adequate to draft contoured maps of the seasonal abundance of individual species. Over most of the ocean, we have only limited information on the presence or absence of species and in any event the published maps too often lack conviction because they do not plot ‘absence.’ However, despite this all-too-frequent lack of rigour, sufficient information is now in hand that we can be confident that the general outlines of pelagic biogeography (in the taxonomic sense) are known. For instance, the

distribution maps for many Pacific plankton species published by the Scripps Institution of Oceanography carry conviction that a few consistent and repetitive patterns have been identified, for whose existence logical explanations can be suggested. Without an extraordinary and unlikely expenditure of ship-time, this situation will not change and we must do the best we can with what we have. But from satellite imagery we can now obtain regional and global maps of the near-surface chlorophyll field, from which we may compute phytoplankton biomass. These images show us for the first time how one component of the planktonic biota responds to the fine spatial detail of the eddying flow-field of ocean currents. For just a few cases we have sufficiently detailed grids of samples to infer something of the small-scale distribution of zooplankton species, and these show that this distribution in no way resembles the smooth contoured maps of abundance usually obtained from plankton surveys. One unusually close study shows how the abundance of euphausiids in the California Current matches the details of circulation around interacting eddies, and so conforms to the pattern of the surface chlorophyll field observed in simultaneous satellite images. This demonstration should cause us to ask how best to extrapolate from chlorophyll images to the geographic distribution of other organisms – initially, of course, the herbivores. All this being said, maps of the surface water masses, of the distribution of individual species, of community diversity, of individual biomes, and of metazoan and algal biomass are nevertheless all remarkably concordant at the large spatial scale. All tend to reflect the zonal distribution of oceanographic

SA SA TR

TR CN

CN

EQ (ETP) EQ

EQ

EQ

CN

CN

CN TR TR SA

SA

TR SA

Figure 1 Special biogeography: a generalized arrangement of characteristic boundaries for distributions of oceanic plankton species, excluding Polar forms. Vagrant individuals will occur outside these boundaries, sometimes even abundantly, but are unable to establish presistent populations there. Boundaries for this figure have been redrawn from the suggestions of McGowan and Backus. Sa, Sub (Ant)Aratic; TR, Transitional; CN, Central; EQ, Equatorial; ETP, Eastern Tropical Pacific.



361

Table 2 The eight models of seasonal phytoplankton growth that characterize conditions in each of the 57 biogeochemical provinces into which the ocean may be partitioned Model 1. Polar biome 2. Westerlies biome 3. Westerlies biome 4. Trade Wind biome 5. Trades and Coastal biomes 6. Coastal biome 7. Coastal biome 8. Coastal biome

Light-constrained summer bloom (6 provinces in cold polar and sub-polar seas) Winter-spring bloom terminated by nutrient limitation (15 provinces of the open ocean in mid-latitudes) Spring bloom terminated by nutrient limitation (4 oceanic provinces of the North Atlantic and Pacific) Small-amplitude algal growth response to trade wind seasonality (6 provinces of the open tropical oceans) Large-amplitude response to reversal of monsoon winds (Coastal: 4 provinces, Trades: 2 open ocean provinces) Canonical spring–autumn blooms of mid-latitude continental shelves (parts of those coastal provinces that include mid-latitude continental shelves) Topographically forced summer production (3 provinces where the continental shelf is unusually wide and level) Intermittent and seasonal blooms at coastal divergences (6 upwelling provinces, principally in eastern boundary currents)

properties, most clearly in the Southern Ocean and North Pacific, and less so in the North Atlantic and the monsoon regions of the Indo-West Pacific. Thus, the zonal Polar and Subtropical Convergences are prominent not only in maps of oceanographic properties but also in many distributions of species and biological properties. Similarly, the pattern of zonal currents in the equatorial zone is captured not only in oceanographic maps but also in satellite imagery of chlorophyll and in distributions of individual species of plankton. Biogeographic analysis at this level of abstraction does not, of course, explain the differential distribution of individual species. Generalization in this special aspect of biogeography depends on the recognition and analysis of repetitive pattern expressed in individual distribution maps, which must reflect not only the circulation and ecology of the presentday ocean but also (as already noted) the geological evolution of the ocean basins. A distribution pattern that caught the attention of the early biogeographers, and that seemed to falsify the concept that all plankton were cosmopolitan, is of species that seemed to occur in both polar oceans, and only there. But modern surveys and taxonomic analysis have shown that many of these so-called ‘bipolar’ species really have much wider cold-water distributions, by progressive submergence equatorward. Just a few truly bipolar species are presently acknowledged, and the real status of these awaits genotype analysis of the disjunct populations in boreal and austral seas. Because the Atlantic is clearly more isolated from the Pacific than from the Indian Ocean, systematists (unable clearly to discriminate between taxonomic species and populations of biological species) are

more likely to accord specific status to populations divided by the American continent, which reaches very far to the south, than to those with greater possibility of exchange around the southern tip of Africa. Consequently, many taxa widely distributed in the warm regions of the Indo-Pacific appear not to occur in equivalent zones in the Atlantic, being replaced there by a related, but apparently different species. Those few cases where critical population analysis has been performed show that the real situation may be very complex, as DNA analysis of many disjunct populations of Cyclothone demonstrate. The eastern Pacific population of this small pelagic fish has greater genetic affinity with that of the western Atlantic than with the other Pacific populations. Disjunct populations must be relatively stable, since no genetic drift has occurred since an eastern population was divided into Atlantic and Pacific stocks by the closure of the Panamanian isthmus. And it also shows that populations within a single ocean basin are able to retain their genetic isolation without intermingling. Such complexity is perhaps not unusual and may be a typical pattern of pelagic biogeography; many populations of closely related species are likely organized in this way. We may take the maps of distribution of Pacific plankton drafted at Scripps Institution of Oceanography as the best available representation of how Linnaean species are distributed in an entire ocean basin. The observed distributions of individual species of several phyla may be grouped into six repetitive patterns (or ‘biotic provinces’), each of which in turn can be associated with features of regional oceanography, or with surface water masses. Perhaps the most satisfactory group synthesis is that


362


2 _ Westerlies biome: nutrient limited spring peak 3 _ Westerlies biome: winter_ spring peak 4 _ Trades biome: low-amplitude production variability 5 _ Trades biome: large-amplitude responses to monsoons 8 _ Coastal biome: intermittent coastal upwelling

Figure 2 Biogeochemical provinces: a simplified arrangement of the 50-odd Sverdrup provinces into which the ocean may be partitioned (see text), but excluding most of the Coastal biome. The figure also illustrates the distribution of five of the eight models of the seasonal cycle of phytoplankton production, based on the Sverdrup model, which are detailed in Table 2.

for euphausiids, because it includes a vertical component, enabling us to compare distributions in the epipelagos with those in the interior of the ocean. In the open Pacific, three distribution patterns of euphausiids are recognized (Figure 2): (i) ‘Central species’ of the subtropical gyres; (ii) ‘Equatorial species’ occupying the zone from 201N–201S, extending somewhat into the western boundary current and including a further partition of a triangular eastern equatorial zone); and (iii) ‘Transitional species’ lying along a narrow belt in each hemisphere at about 401 latitude, and extending equatorward somewhat into the eastern boundary currents. These are ideal distributions, as is made clear from the degree of spatial overlap between them, and species substitutions occur between equivalent zones in the two hemispheres. The Transitional zone is associated with the oceanographic subtropical convergence zone, where eastward transport occurs in a very active eddy-field lying between the subarctic and subtropical gyres. Some species of zooplankton and nekton appear to be endemic to this zone, but phytoplankton blooms are derived opportunistically from algae from both subarctic and subtropical gyres, whichever happens to be have been advected into the Transitional zone when conditions are appropriate for a bloom to occur. So much for the great reaches of the open Pacific Ocean. Marginally, eight neritic species of euphausiids comprise a ‘Boundary’ group and about the

same number comprise a ‘Subpolar’ group, four each in the Subarctic Gyre and the Subantarctic Current. A further set is specific to the Antarctic Current and coastal waters. It is suggested with good reason that species distributions of most groups of planktonic organisms can be accommodated in only about halfa-dozen patterns, together with a residue of more apparently cosmopolitan species. Once again, however, it should be emphasized that it is very probable that modern population analysis will demonstrate genetic differentiation within these apparently cosmopolitan taxa. These distribution patterns of individual species of plankton and micronekton support – by their overlap, and their open boundaries – the member vagrant hypothesis, and they also conform very well in their general arrangement, and in some of their actual boundaries, with the arrangement of biomes and smaller ecological or biogeochemical partitions based on regional oceanography and the Sverdrup model. More remarkably, there is very close concordance between the distributions of individual species of oceanic tuna and the biogeochemical provinces of the Westerlies and Trades biomes, whose distribution is discussed above. This is all the more remarkable considering the ability of these fish to migrate rapidly over great distances. In the eastern tropical Pacific, yellowfin and skipjack are closely constrained within the boundaries of two biogeographic provinces



defined by criteria from regional oceanography, while in the North Atlantic, bluefin tuna distribution supports the east–west division of the Subtropical Gyre into two provinces. From the excellent concordance we can now observe between the distribution patterns of all kinds of pelagic organisms, both with each other and with maps of processes forced by regional oceanography, we may reasonably conclude that, after a difficult start, the way is now clear for biogeography of the pelagic habitat of the open oceans to take its place among the serious disciplines of biological oceanography. Further progress will require work both at the molecular level, in the analysis of genetic isolation of populations, and with satellite-borne sensors to understand better how the single biological variable we can reasonably hope to measure synoptically and globally – the biomass of phytoplankton – is forced by regional oceanographic processes.

See also Diversity of Marine Species. Large Marine Ecosystems. Ocean Gyre Ecosystems. Polar Ecosystems. Primary Production Processes. Upwelling Ecosystems.

363

Further Reading Backus RH (1986) Biogeographic boundaries in the open ocean. In: Pelagic Biogeography. UNESCO Technical Papers in Marine Science 49, pp. 9–10. Brinton E (1962) The distribution of Pacific euphausiids. Bulletin of the Scripps Institution of Oceanography 8: 51--270. Ekman S (1953) Zoogeography of the Sea. London: Sidgewick and Jackson. McGowan JA (1971) Oceanic biogeography of the Pacific. In: Funnell SM and Riedel WR (eds.) The Micropalaeontology of Oceans, pp. 3--74. Cambridge University Press: Cambridge. Longhurst AR (1998) Ecological Geography of the Sea. San Diego: Academic Press. Pierrot-Bults AC and van bulleted Spoel S (1998) Pelagic Biogeography. Workshop report no. 142. Intergovernmental Oceanographic Commission. Platt T and Sathyendranath S (1999) Spatial structure of pelagic ecosystem processes in the global ocean. Ecosystems 2: 384--394. Semina HJ (1997) An outline of the geographical distribution of oceanic phytoplankton. Advances in Marine Biology 32: 528--561. Sinclair M (1988) Marine Populations. An Essay on Population Regulation and Speciation. Seattle: University of Washington.


PELAGIC FISHES D. H. Cushing, Lowestoft, Suffolk, UK

Clupeoids

Copyright & 2001 Elsevier Ltd.

Herring: A Case Study of a Pelagic Fish

This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2122–2127, & 2001, Elsevier Ltd.

Herring live in the North Atlantic (Clupea harengus L.) and in the North Pacific (Clupea pallasii Val.). The edges of the scales on the belly of the Atlantic herring are rough, but in the Pacific herring they are smooth. Atlantic herring are found in the east from the Bay of Biscay to Spitzbergen and the Murman coast and in the west from Greenland and Labrador to Cape Hatteras on the eastern seaboard of the United States. The Pacific herring is found off British Columbia, off Hokkaido in Japan and off Kamchatka. Although they live in the midwater, both species lay their eggs on the seabed; the Atlantic herring usually lay their eggs at depths of 40–200 m and in some cases intertidally on gravel and small stones and the Pacific herring spawns on seaweeds between tidemarks. Both species spawn on narrow strips; for example that near the Sandettie´ Bank not far from Dover is 3000 m long and 300–360 m wide. Herring are small fish, the adults being about 25– 30 cm in length; Baltic herring (Clupea harengus/ membras) are smaller and the Norwegian herring are somewhat larger. North Sea herring live for about twelve years and Norwegian herring for about twenty years. They mature at about three or four years of age and their annual fecundity amounts to 40 000–100 000 eggs. Throughout life the total fecundity would be about ten times greater and only two survivors are needed to replace the stock. Hence the annual mortality each year of the eggs, larvae and juveniles is high; indeed, it is approximately the inverse of the annual fecundity. The natural mortality of the adults is that sustained under predation in the absence of fishing and, as might be expected, it is rather difficult to establish; that of the herring might be about 10–20% of numbers per year. They live and swim in large shoals, each of which may be many kilometers across. They migrate towards the surface at night and then the shoals tend to disperse. They make long migrations each year from spawning ground to feeding ground and back, distances of up to 2000 km. To do this they usually swim down tide or current. Herring feed on Calanus among other plankton animals. They grow quite quickly particularly during the first year of life, but as adults they grow relatively slowly, as more energy is devoted to reproduction. The herring of any spawning group spawn at the same season each year to within about a week. It is likely that they return to the grounds on

Introduction In the economy of the sea pelagic fish play a central part. The simplest food chain comprises phytoplankton, copepods and pelagic fish. They spend their lives in the midwater of coastal seas and oceans. Much of our knowledge about them comes from the fisheries that exploit them, which yield very large catches. The three groups, clupeoids, tunas and mackerels live in all parts of the ocean and many of them migrate across the seas for considerable distances. The clupeoids include such fishes as herring, sardines and sprats which have supported large fisheries since the earliest times. Anchovies are widespread and the Peruvian anchoveta once supported a very large fishery. Herring are caught in the North Atlantic and North Pacific; sardines are taken in the upwelling areas and sprats are mainly caught in the North Sea. Herring can migrate for up to 2000 km each year and their stocks (or populations) have yielded annual catches of as much as one million tonnes, wet weight. They are smallish fish ranging in length from about 12 cm (sprats) to 20 cm (sardines) and 25–30 cm or longer (herring). The tunas are larger fishes, one or two meters in length and each year they make transoceanic migrations. They include yellowfin, albacore, bigeye and bluefin. Catches amount to about a million tonnes each year and they are taken at many places in the world ocean, but particularly in upwelling areas and at fronts. Mackerels are larger than herring (up to 40 cm in length) and they also migrate across considerable distances. The common mackerel in the North Atlantic and the cosmopolitan Spanish mackerel are typical examples; horse mackerels (in a different suborder) are also widely distributed. The pelagic fish spend their lives in the near-surface layers of the ocean and they swim steadily for long periods. From their central position in the marine ecosystem they control the passage of energy up the food webs from the algal cells and copepods to fish.

364


365

PELAGIC FISHES

which they were spawned, although this cannot be shown decisively. In the Northeast Atlantic there are a number of spring spawning stocks: Norwegian, Murman, Shetland and Faroe Islands. The Icelandic spring spawners may no longer exist, but the Icelandic summer spawners flourish. There are some small stocks in the Skagerak and there are very small local stocks that spawn at the mouths of certain rivers such as the Elbe. In the North Sea there are two autumn spawning stocks, Buchan and Dogger in the northern and central North Sea and a winter spawning stock, the Downs, in the southern North Sea. The Buchan group spawns off the northeast coast of Scotland; the Dogger group spawns off Whitby and around the Dogger Bank. The Downs stock spawns in the English Channel off Dover and in the Baie de la Seine off northern France. There is another winter spawning stock in the western English Channel. Icelandic summer spawners lay their eggs on three grounds off the north coast of the island. The Norwegian herring spawn in spring between Egersund and Bergen, between Bergen and Kristiansund and off Lofoten and Westeralen. The North Sea herring migrate in a clockwise circuit in the North Sea (Figure 1). The Norwegian herring migrate from the Norwegian coast across to Iceland and then northwards to Jan Mayen and the Barents Sea before returning to their spawning grounds on the Norwegian coast (Figure 2). The Icelandic herring stocks migrate round the island in a clockwise direction. In the Northwest Atlantic there are a number of herring stocks: off Labrador, the southern and western coasts of Newfoundland, the Scotian Shelf, Georges Bank, and to the south of Georges Bank. Most of these stocks are spring spawners, but after 1950, off the south and west of Newfoundland the migrations of the stock on the Scotian Shelf are quite limited but the reason for this difference is not known. In the Pacific there are two main groups of stocks, that off British Columbia and that in the Far East. The British Columbian fishery is established off Vancouver Island, and off Queen Charlotte Is. There are four major stocks in the east, the SakhalinHokkaido (a spring spawner), the Kamchatka herring, the Kora Karazynsk herring and the Okhotsk herring. Each stock migrates from feeding ground to spawning ground over distances of about 1200 km, for the first three stocks and about 500 km for the Okhotsk herring (Figure 3). In the North Sea herring have been caught from the earliest times by drift net off Shetland, off the northeast coast of Scotland, off northeast England

Shetland Is Buchan 60˚ Fladen ground Skagerak 58˚ Aberdeen r

ge

56˚

g Do

Whitby 54˚ The Wash Downs Thames Estuary Dover

52˚

Boulogne 50˚

Baie de la Seine 4˚

2˚

0˚

2˚

4˚

6˚

8˚

Figure 1 The migration circuits of the three groups of spawners in the North Sea: Buchan, Dogger and Downs.

and off East Anglia. Later, they were caught by trawl on the Fladen Ground northeast of Aberdeen, off the Dogger Bank and on the narrow grounds not far from Dover and Boulogne. In the Norwegian Sea herring were originally caught by drift net in the fiords and later by purse seine. Finally, when the purse seine could be worked in the open sea, herring were caught there off southern Norway and also north of the Lofoten Is. Off Iceland, drift nets were also replaced by purse seines worked in the open sea. Nowadays most herring are caught by purse seine and midwater trawl. In the Northwest Atlantic the traditional fishery took place in the Gulf of Maine and in the Bay of Fundy. In recent years, trawl fisheries have been developed on Georges Bank and purse seine fisheries off Nova Scotia. There were fisheries also off Newfoundland and in the Gulf of St Lawrence. Catches were converted into fish meal and oil. In the Far East, the major fishery was that for the Hokkaido spring herring which were caught by offshore gill nets. The fishery peaked between 1895 and 1905 after which it declined to low levels. Since 1950, the offshore gill nets have been replaced by trawlers. Off Sakhalin and in the Sea of Okhotsk herring have been caught by gill nets for a long time.


366

PELAGIC FISHES 80˚

75˚

70˚

65˚

75˚

ts ren Ba ea S

60˚ 45˚ 30˚ 90˚

75˚ 60˚ 45˚ 30˚ 15˚

0˚

15˚

an rm Mu oast C

z

len

era

st We

60˚

n

Lofote

Greenland

Jan en May

er

g

n va Sta

tian Kris und s en Berg

nd Icela

55˚

Faroe Is

d

un

rs ge

50˚

E

45˚

Figure 2 The migration circuit of the spring spawning Norwegian herring; the spawning ground lies off the Norwegian coast, the nursery ground spreads north into the Barents Sea and the feeding ground lies between Iceland and Spitzbergen.

The global catches of herring have amounted to several million tonnes. In the North Sea the Downs stock was overexploited in the late 1950s by trawling on the spawning grounds and the two other groups, Buchan and Dogger, suffered from overfishing by purse seiners in the late 1960s. The Norwegian spring spawning stock of herring was overexploited by the offshore purse seiners in the early 1970s. In the Northwest Atlantic the fishery on Georges Bank became too heavy in the 1970s. It is possible that the Hokkaido stock suffered from overfishing but this cannot be shown decisively. Sprats

Sprats, Clupea sprattus (L.), are small clupeoids about 10–13 cm in length. They live for about five to seven years, but the most abundant age group is the third. Like other clupeoids, they feed on copepods and other plankton animals. They spawn in spring and summer, but in the northern Adriatic they spawn between December and March. They mature from the end of their second year to the end of their third

year and they lay between 10 000 and 40 000 eggs each year. The natural mortality is probably high. There are major spawning grounds in the southern North Sea and in the southern Norwegian fiords. The eggs, larvae and juveniles are fully pelagic. Sprats are found in the Baltic, in the North Sea, in the northern Adriatic and off Romania in the Black Sea. Fisheries were established in the Scottish firths, between Bergen and Stavanger in Norway, off Brittany in France and around the English coast, particularly in the Wash and in the Thames estuary. In the main, these are winter fisheries worked with drift nets and stow nets, i.e. bag nets hung from fixed poles in the tidal streams. Sardines

There are a number of sardines and related species: Sardina pilchardus Walb in European waters, Sardinops caeruleus (Girard) off California, Sardinops ocellatus (Pappe´) off South Africa, Sardinops melanostictus (Schlegel) off Japan and Sardinella aurita Val. off West Africa. The pilchard (S. pilchardus)


PELAGIC FISHES

150˚

160˚

y Ba

C.Navarin

4 60˚

G

60˚

180˚

170˚

rf Ko

izh Ba igin y sk

140˚

367

3 Karaginski

Sakhalin

50˚

BERING SEA

Kam

OKHOTSK SEA

cha tka

2

1 50˚

1 1 2 o

kaid

JAPAN SEA

Hok

140˚

150˚

160˚

170˚

180˚

Figure 3 The stocks of Clupea pallasi in the Far East. 1, Sakhalin-Hokkaido herring; 2, Okhotsk herring; 3, Gyzhigynsk-kamchatka herring; 4, Korto-karagynsk herring. Bold line, spawning grounds; Hatched areas, feeding grounds.

lives in the English Channel, in the Bay of Biscay and off the Iberian peninsula. Elsewhere, sardines live mainly in the upwelling areas in subtropical oceans. An upwelling area is an extensive region along western coasts in the subtropical ocean both north and south of the equator. The four main regions lie off California, off Peru and Chile, off southern Africa and off northwest Africa. There are also lesser upwellings off the west coast of the Iberian peninsula, off Ghana, off southern Arabia and off the Malabar coast of India. In an upwelling area the wind blows towards the equator. Water advects offshore and the nutrient-rich replacement is drawn up from below and in it production starts. In these rich areas sardines flourish in large populations. Sardines are smaller than herring with an average length of just over 20 cm. They live up to ten years but the abundant year classes peak at about four or five years of age. An average year class of Sardinops caeruleus off California might comprise about a billion fish. The natural mortality of the sardine is high because they are eaten by a wide range of predators. They spawn in spring and fall off Baja California; within the Baja they spawn in late winter. Their fecundity may amount to about 40 000 eggs year 1. Shoals lie along the line of the tide or current and, like herring shoals, they are much longer than they are wide. Sardines do migrate within an upwelling area, but their movements are restricted. They feed on copepods; the larvae eat copepod

nauplii and the larger fish feed on adult copepods. They tend to spawn in winter and spring in both hemispheres, perhaps a little before the upwelling strengthens (but off southern California, they can also spawn in the fall). Eggs, larvae and juveniles are all pelagic. Records of sardine catches off Japan go back to the fifteenth century. The rich periods lasted from twenty to seventy years, for example, from 1660 to 1730, 1818 to 1864 and 1917 to 1939. From five to seven local groups were recognized but they tended to flourish or decline together. One of the most remarkable events is the trend in catches peaking between the 1930s and the 1950s, off Japan, off California, off Spain and in the northern Adriatic. Catches of the Japanese sardine (Sardinops melanostictus) rose to a peak of about 2 500 000 tonnes in 1935–36 and then declined to very low levels between 1945 and 1972. With the year classes, 1977 and 1980, catches recovered to over 4 000 000 tonnes each year by 1985. Subsequently catches again declined. These events occurred at places far apart and there is, as yet, no explanation for the fact that the catches rose and declined together. Pilchards were caught by drift net off Brittany and by ring net off Cornwall. In the upwelling regions purse seines were used. Catches reached a peak quite quickly and then declined and the fishermen would switch to another pelagic stock. Off South Africa sardines were replaced by anchovies and off Peru, the


368

PELAGIC FISHES

anchoveta was replaced by sardines. The sardine stock off California collapsed and was replaced by the stock of the northern anchovy (Engraulis mordax Girard) which was subsequently exploited further south by Mexican fishermen. Off northwest Africa sardine catches remained fairly steady in the north off Cap Ghir but further south between Cap Blanc and Cap Bojador catches reached very high levels before a sudden collapse. The fisheries for sardines and anchovies were very large and they supplied a market for fish meal and oil. Anchovies (Engraulis spp.) are small fishes and they live not far from the coast. They are caught in small fisheries in European waters. Off India, a rather larger fishery was practiced on Rastrelliger kanagurta (Cuvier) and a much bigger fishery has been that for the Peruvian anchoveta (Engraulis ringens Jenyns). At its peak it was the largest fishery in the world, yielding about eleven million tonnes each year.

Tuna The tunas are large pelagic fish which grow to up to 2 m in length. The principal species are the yellowfin (Thunnus albacares Bonaterre), bigeye (Thunnus obesus Lowe), skipjack (Katsuwonus pelamis L.), albacore (Thunnus alalunga Bonaterre) and bluefin (Thunnus thynnus L.). They are, in the main, subtropical animals and are distributed across each of the oceans making transoceanic migrations. The natural mortality of the tunas is probably rather high because they do not appear to live very long. Catches of each species amount to about 100 000 or 200 000 tonnes each year; for all tuna species total catches amount to about one million tonnes each year. Tuna larvae are found all over the subtropical ocean at nearly all seasons which means that spawning is widespread and continuous, but they are also found in the upwelling areas. In three years the bluefin tuna grows to 50 kg. The yellowfin is most abundant in the divergences and convergences of the equatorial system. There is a spawning migration into the Mediterranean and tuna have been caught in small traps and sighted from vedette (look-out posts) in Dalmatia, on the Roussillon coast of France and off Algeria. Bluefin are caught in the tonnare di corsa (spawning migration) in Spain, Portugal, France, Sardinia and Sicily. The tonnara is a complex structure; that off Trapani in Sicily was manned some years ago by ninety-three men. Porters, stevedores, cooks, boxmakers and coopers supported such operations. The fish were herded into the mattanza or ‘death room’ with white palm leaves and as the nets were brought towards the surface the tuna were killed with lances.

Off Brittany, albacore are caught with hooks on long rods. The fish are most abundant at the shelf edge, associated with ‘heavy swells’ perhaps where internal waves ride from south to north. Off California, tuna are caught either by pole and line with live bait or with purse seine in the equatorial region and there is also a sport fishery off the coasts of the United States. Yellowfin are prominent in the catches perhaps because they tend to live in stocklets just north of the North Equatorial Current. Albacore tagged off California can be recovered off Japan for the fish may live right across the North Pacific gyre. The larger fish live in the North Equatorial Current and spawn east of the Philippines.

Mackerels There are a number of mackerels together with their relatives. The Atlantic or common mackerel, Scomber scombrus L., lives in the Atlantic off the North American coast and in European waters. The chub mackerel, Scomber japonicus Houttuyn, is cosmopolitan. The Indian mackerel is Rastrelliger kanagurta (Cuvier). The common horse mackerel, Trachurus trachurus L., lives in the Atlantic; the Pacific form is Trachurus japonicus. The Atka mackerel of the North Pacific is a scombrid, Pleurogrammus azonus Jordan and Metz. The Spanish mackerel, Scomberomorus commerson Lace´pe`de, is a larger animal. Mackerels are larger than herring, reaching as much as 40 cm in length or more. They spawn in the midwater in productive seas where the larvae and juveniles grow up, and they spend the rest of their lives there. They feed on copepods and other plankton animals. They swim quickly and some of the larger ones are predatory. The natural mortality of the mackerels is perhaps high, up to 30% of numbers per year. They do not shoal as herring do but there may be small and transient aggregations. In the Northeast Atlantic annual catches of mackerel have reached as much as one million tonnes from two stocks, one in the North Sea and the other to the west of the British Isles. They are found in the Mediterranean but catches have only amounted to about 30 000 tonnes each year. Considerable catches of the Atlantic mackerel have been made off the eastern seaboard of the United States. The stock of Atka mackerel in the Northwest Pacific yielded annually about 100 000 tonnes. The Pacific mackerel yielded large catches for a period, after which they declined. Tens of thousands of tonnes of Rastrelliger are taken each year off the Indian coast. The horse mackerel (Trachurus trachurus L.) has usually been caught in trawls. It is somewhat larger


PELAGIC FISHES

than the common mackerel, is pelagic and feeds on small animals in the plankton. It probably does not shoal very much and lives rather deeper than the common mackerel or the herring. In the South Atlantic about 100 000 tonnes of maasbanker (Trachurus trachurus L.) are caught. Some tens of thousands of tonnes are caught in the open ocean.

Conclusion The pelagic fish occupy a central position in the marine ecosystem as they harvest food from lower trophic levels and support the predators in higher ones. The stocks respond to climatic changes and provide near stability to the tuna and fishes like them. Catches are very large, indeed the largest sector in the world harvest.

See also Fish Migration, Horizontal. Fish Reproduction. Marine Fishery Resources, Global State of. Open

369

Ocean Fisheries for Large Pelagic Species. Small Pelagic Species Fisheries.

Further Reading Cushing DH (1982) Fisheries Biology. Madison: University of Wisconsin Press. Cushing DH (1996) Towards the Science of Recruitment in Fish Populations. Oldendorf Luhe, Germany: Ecology Institute. FAO (1983) Species Catalogues. No. 125, Vol. 2 Scombrids of the World; No. 125, Vol. 7(1) Clupeid Fishes of the World; No. 125, Vol. 7(2) Clupeid Fishes of the World. Rome: FAO. Graham M (1943) The Fish Gate. London: Faber. Gulland JA (1974) The Management of Marine Fisheries. Bristol: Scientica. Hardy A (1959) The Open Sea: Fish and Fisheries. London: Collins. Rothschild BJ (1986) The Dynamics of Marine Fish Populations. Cambridge, MA: Harvard University Press.


PELECANIFORMES D. Siegel-Causey, Harvard University, Cambridge MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2128–2136, & 2001, Elsevier Ltd.

Introduction Who does not know what a pelican looks like? Pelicans, and the other members of the Pelecaniformes, are common inhabitants of the marine littoral and are found almost everywhere along the margins of the world ocean. They are medium-sized to very large aquatic birds found commonly near coastal marine and inland waters, feeding mainly on fish and less commonly on small arthropods and mollusks. Pelicans typify two of the common features of this order of birds, including a large distensible pouch under the bill and a totipalmate foot (four toes connected by three webs). Other features common to Pelecaniformes are less exclusive, but altogether are diagnostic of the group. The feet are set far back on the body for efficient swimming, but on land most of this group are clumsy and almost immobile. The eggs and chicks are incubated on the feet, the adults lacking brood patches. In most, chicks are born blind and naked, and down feathers appear after a few days. Chicks feed on regurgitated food taken by inserting their head into the parent’s throat. Most species breed in large, dense colonies, and in places the birds are a valuable source of guano – a source of nitrates and a valuable fertilizer. Colonies are susceptible to disturbance by humans and small mammals when eggs and chicks are lost through temperature imbalances after the adults leave the nest, or by predation. Because of their proximity to humans, and dense aggregations in feeding and breeding activities, most pelecaniforms – cormorants and pelicans in particular – are persecuted by fishermen under the mistaken belief that they deplete fishing stocks.

Systematics General

On the surface, Pelecaniformes are an order of birds easily identified by a few key characters, the most noticeable being a totipalmate foot, that is, one with all four toes joined by a web. No other assemblages of birds possess this trait: gulls and waterfowl have

370

only three toes joined by a web, and some birds have only a single web between two toes. Other features have been associated with this group, including a gular pouch, or the unfeathered, loose folds of skin underneath the bill, and a lack of brood patches on the abdomen. Six families are now considered to be placed in the Order: tropicbirds (Phaethontidae), frigatebirds (Fregatidae), pelicans (Pelecanidae), gannets (Sulidae), cormorants (Phalacrocoracidae), and darters (Anhingidae). Recent studies on the morphology and molecular variation in the traditional species of pelecaniform birds suggest, however, that these groups may not be as closely related as formerly thought. For example, there are doubts that tropicbirds and frigatebirds are pelecaniform, and instead ought to be placed in other groups, such as the tubenoses (Procellariformes), herons and storks (Ciconiiformes), or their own orders. It also appears that, except for the lack of toe webbing, the shoebill (Balaeniceps rex) and Hammerhead (Scopus umbretta) herons are most closely related to pelicans, and may represent aberant members of the Pelecaniformes. The pelecaniform birds and other large waterbird orders (i.e., Sphenisciformes, Gaviiformes, Podicipediformes, Procellariiformes, Ciconiiformes) are considered by most systematists to comprise the early branching lineages of birds with flight. The earliest known fossil pelecaniform is thought to be Elopteryx from the Cretaceous (70 million years ago), but as with all early bird fossils there is considerable debate about its identity – even whether or not it is avian! Cormorants are among the earliest known pelecaniforms, with fossil elements recovered from as early as the Eocene–Oligocene boundary (approximately 55 million years ago). There are approximately an equal number of extinct and extant species described, but the fossils are often from single bones and of dubious affinities. The species-level taxonomy has been very neglected, with some families having many species and subspecific forms described (e.g., cormorants with 39 species and 57 taxa), while others have had little attention. For example, the darters are variously described as comprising one, two, or four species, but with few data to back up any of the competing schemes. As a whole, the group is drab or monochrome or both, and sexes are dimorphic only in the darters and the frigatebirds. Order Pelecaniformes

The groups now considered to be members of the Order Pelecaniformes comprise 6 families, and


PELECANIFORMES

include 6 genera, 67 species, and about 120 described subspecies, races, or types (Figure 1). Cormorants are the most diverse group, with half of the extant species; tropicbirds and anhingas are the least diverse, with 3–4 species known at present. Pelecaniform birds are found on every continent, along most marine coasts and major fresh water drainages, and are absent only from extreme deserts and the icecovered regions. The nest is built by the female with material usually brought by the male, and is situated on the ground or in trees, selected to be well protected from predators. Reproductive behavior is complex and involves primarily visual displays near the nest site; vocalizations are rare and often described as being grunts, croaks, or strident. Phaethontidae A single genus, Phaethon, and three species are found throughout the tropical waters of the world ocean. Tropicbirds differ from the typical pelecaniform birds and really only share a single feature – the totipalmate foot – with the other species. The chicks hatch fully covered with down; they are fed by the adults from the bill rather than the throat as in other species; courtship displays are aerial, noisy, and synchronized; the bill lacks a terminal hook; the gular pouch is feathered and obsolete; and there are many other distinctions. All are medium-sized – about the size of a small gull – white with black eye and wing barring, with long tapering wings and gull-shaped bills, and with very elongated tail feathers forming streamers often longer than the rest of the body. The white parts of the plumage are often tinted pink or gold; the sexes are alike in appearance. Species range widely over the warm tropical and subtropical waters, and represent the most oceanic of all pelecaniform birds. They only come to land for breeding, on remote oceanic islands, and select nest sites that are inaccessible to terrestrial predators and have some shade, such as might be provided by overhanging rocks on shrubs, or within cavities. Little is known about the ecology and behavior of the family; the birds are rarely seen in groups or flocks, often only singly, and usually in flight. Feeding seems to be in low light (i.e., early morning or late evening) and by plunge-diving into local concentrations of cephalopods and flying-fish. Hovering over schools and ship-following have been noted, as well as gathering with multispecies feeding flocks. More commonly, though, tropicbirds tend to forage solitarily, dive abruptly, capture food close to surface, and then take to the air soon after swallowing their prey. Fregatidae Frigatebirds are found throughout tropical and subtropical oceans of the world,

371

commonly associated with the regions of trade winds. They are large – roughly the size of large gulls – and predominantly black, with some species having white breasts or bellies. Five species are described, but there is substantial confusion on the status of certain island forms, plumage variants, and size differences. The wings are large, elongated, and pointed, with a distinctive open W-shaped flight silhouette, and all species are quite deft in the air. During breeding, males inflate their scarlet gular pouch to nearly half the size of the body, and do so in groups and alone, on ground and in the air, and at times will vibrate it with the bill, as though drumming. Frigatebirds will leave the roosts at dawn and spend the rest of the day foraging alone or in small groups. They are known for feeding by theft (‘kleptoparasitism’; hence their common names of frigate birds or man-of-war birds) but commonly feed on shoals of flying fish, cephalopods, and rarely krill. They rarely dive in the water, preferring instead to feed by surface dipping or even aerial captures of emergent flying fish. Every tropical seabird colony seems to be attended by several frigatebirds, and unattended eggs and chicks are taken by swooping after an aerial dive. Fish are snatched from the bills of other seabirds, whether at sea or on the nest. The breeding season seems to be determined by local conditions, such as seasonal onset of oceanographic or weather patterns, abundance of food, and appearance (or nonappearance) of neighboring breeding species. The reproductive cycle is longer than for most of the other Pelecaniformes, with incubation lasting as long as 60 days, fledging as long as 7 months, postfledging dependence on adults lasting on average between 9 and 12 months, and age of first breeding from 6 to 11 years. This seems to be related to the unpredictable nature of the food supply, low chick productivity, and low breeding success. While the birds are not threatened at present except in a few localities, these aspects of natural history make them particularly vulnerable to human disturbance and habitat destruction. Pelecanidae Pelicans are the largest members of the order, and are distinguished by their large body, long, heavy bill, and huge distensible gular pouch. Seven species are recognized, and all are found not far from water and rarely out of sight of land. Plumages are generally light colored – only the brown pelican is dark – and in the breeding season the bare facial and gular skin, and often bills, become brightly colored in both sexes. Because of size and large broods, pelicans are voracious feeders. They predominately eat fish, and feed by


372

PELECANIFORMES

Figure 1 Examples of Pelecaniform Seabirds. (1) Northern gannet (Sula bassana). Other names: North Atlantic gannet, gannet. Length: 93 cm; wingspan: 172 cm; approximate body mass: 3090 g. Range: North Atlantic Ocean. (2) Brown booby (Sula leucogaster). Length: 69 cm; wingspan: 141 cm; approximate body mass: 1150 g. Range: Tropical Pacific, Atlantic and Indian Ocean. (3) Brown pelican (Pelecanus occidentalis). Length: 114 cm; wingspan: 203 cm; approximate body mass: 3500 g. Range: west coast of the Americas from British Columbia to Ecuador; east coast of the Americas from New Jersey to the Amazon River mouth, and Caribbean waters. (4) Red-tailed tropicbird (Phaethon rubicauda). Other names: Bosunbird. Length: 46 cm; wingspan: 104 cm; approximate body mass: 715 g. Range: Tropical Pacific and Indian Oceans. (5) Great frigatebird (Fregata minor). Other names: Man o’War Bird. Adult female. Length: 93 cm; wingspan: 218 cm; approximate body mass: 1375 g. Range: Tropical Pacific, Atlantic and Indian Oceans. (6) Great cormorant (Phalacrocorax carbo). Other names: common cormorant. (a) Adult; (b) immature. Length: 90 cm; wingspan: 140 cm; approximate body mass: 2280 g. Range: Widely distributed in coastal waters of world except not found along west coast of the Americas or east coast of Central and South America. (7) Imperial shag (Phalacrocorax atriceps). (a) Adult breeding; (b) non-breeding plumage. Length: 72 cm; wingspan: 124 cm; approximate body mass: 3120 g. Range: Southern Ocean. Illustrations from Harrison P (1985) Seabirds, an identification guide. Revised edition. Boston, Massachusetts; Houghton Mifflin.


PELECANIFORMES

plunge-diving into surface shoals of fish with the bill opened. Most pelicans are found near lakes and large water bodies, and only the brown pelican is primarily marine. Pelicans are gregarious, and are rarely encountered alone. They breed almost exclusively in large colonies; the prefledged young form large creˆches within the breeding colony; and roosting and flight are done in large groups. Pelicans have the best-documented communal feeding, in which large groups will coordinate fishing in fresh or brackish waters as a group. From 10 to 50 birds will alight on the water in a row, and drive surface-schooling fish toward shallow water by paddling, flapping wings, and plunging their bills in the water, sometimes feeding and sometimes not. Other birds plunge-dive into the water after fish, or swim from beneath; when captured, the fish are retained in the gular pouch as the water drains away. They are then eaten, often after being flipped up into the air to be swallowed head-first. Pelicans, along with cormorants, are among the birds most affected by human disturbance and environmental pollution, particularly by chlorinated pesticides such as DDT and dieldrin. Pelicans and humans have long been associated, and freshwater species are vulnerable to habitat loss and pollution. Fishermen have persecuted pelicans and other fisheating birds in times of low fish catch, but most studies indicate that the take by pelicans is primarily of noncommercial fish. In every case, low fishing yields are due to poor management or poor environmental stewardship, and not due to pelican–human competition. Pesticide contamination of the marine foodweb led to drastic reductions in brown pelicans throughout their range midway through the twentieth century, but pollution abatement and environmental protection efforts have allowed populations to recover. Pelicans now serve as an icon for conservation success, but many populations are still under threat. Sulidae The sulids, gannets and boobies (genera Morus and Sula) include 9 species found throughout the tropical and temperate regions of the marine environment. Generally, they are duck-sized, with a long conical bill finely serrated along the edges. Gannets are found in high-latitude waters, while boobies tend to be more tropical; both are exclusively marine and piscivorous. Sulids feed by plunge-diving, then often shallow pursuit (usually to no greater than 15 m depth) after fish in the upper water levels. The plumage is thought to reflect this behavior, as the light chest and abdomen will provide little contrast when seen from below, and prove indistinct when seen from above. Whether related to camouflage or swimming ability, sulids are very efficient in capturing small

373

schooling fish such as anchovy (Engralis) and sardines (Sardinops). The timing of breeding seems mostly associated with local conditions of food and nest site availability, and occurs annually in most species. In very good conditions, the tropical boobies will breed at intervals less than 12 months, and the red-footed booby in the Galapagos often breeds once in 15–18 months. These nonannual periods have been thought to be a response to bad (or good) oceanographic conditions related to the El Ninõ Southern Oscillation, or conversely La Ninã, or as a means to desynchronize breeding with potential nest predators such as frigatebirds, or both. In both gannets and boobies, the reproductive cycle is similar to other seabirds, in that chicks fledge, leave the nest, and are independent before the end of the first year. Sulids usually do not breed for the first time until they are 3–6 years old, and live 10–20 years. The largest clutches are found in the species breeding in upwelling regions with high fish production (e.g., Peruvian and blue-footed boobies), and average 2–3 eggs. Other tropical boobies such as brown and masked boobies will lay two eggs, laid 4– 6 days apart; the larger and stronger chick – not always the first one hatched – will often push the weaker chick from the nest to its ultimate demise. All other sulids lay a single egg. The most aberrant and presumed the most primitive sulid is Abbott’s booby, which is now restricted to Christmas Island, Indian Ocean. Chicks take 20–24 weeks to fledge, and are fed by adults for an additional 20–40 weeks. This slow development is attributed as an evolutionary response to the unpredictable and distant feeding grounds of the adults, but the evidence is not clear. Abbott’s booby is also the most threatened as a result of habitat loss, constriction of breeding range, and low productivity in raising chicks to breeding adults. Phalacrocoracidae There are nearly as many species of cormorants and shags as all of the remaining members of the Pelecaniformes. This family has about 60 described taxa; these are sometimes lumped into as few as 26 species or as many as 40 or more. The discrepancy lies in the many related island forms found throughout the Southern Ocean, including Antarctic, Subantarctic, and Southern Subtemperate waters. There are two distinct subfamilies, the cormorants (Phalacrocoracinae) and the shags (Leucocarboninae), although the common names often do not correspond so neatly: the pied shag of the Antipodes is in fact a cormorant, and the pelagic cormorant of the North Pacific is a shag. The two groups are distinguished by clear morphological and genetic differences, and represent an early


PELECANIFORMES

divergence as far back as the Miocene, perhaps 15–20 million years ago. Cormorants are generally heavybodied birds with labored flight, commonly found on fresh water systems, often far into the interior. Shags are smaller and slimmer, are more powerful fliers, and are common inhabitants of the marine littoral, neritic, and pelagic waters. In many species, particularly those of the Northern Hemisphere, the plumage is basically black or very dark, often with metallic or iridescent sheens to the feathers. Most shags and cormorants of the Southern Hemisphere have in addition a white abdomen; in a few species such as the red-legged shag and spotted shag (Stictocarbo gaimardi and S. punctatus), the plumage is predominantly gray. In breeding, adults will often undergo dramatic changes in color and appearance that affect both bare parts and the plumage. The naked skin of the face, gular pouch, bill lining, and rarely feet, become more intensely colored, change color, or even assume color, and these changes are often accompanied by enlarged or increased carunculations near the eyes and face. Many species produce long white filoplumes and breeding plumes on the head and neck (great cormorant, Phalacrocorax carbo sinensis), and in a few species (red-faced, pelagic cormorants Stictocarbo urile, S. pelagicus) a conspicuous white patch appears on the thigh. Unlike many other Pelecaniformes, cormorants and shags have very flexible requirements for nesting and nest construction. They are known to breed in diverse habitats including on level ground, on cliffs and embankments, in trees, on bridge supports, and on wharfs: generally, on objects large and sturdy enough to support the weight of the nest and occupants while affording protection from ground predators. Proximity to water is an absolute requirement because cormorants are nearly exclusively piscivorous, and the size of breeding colonies in continental interiors correlates with how close they are to feeding areas (Figure 2). By virtue of wing morphology and aerodynamics, cormorants are indifferent fliers and do not range far from roosting or breeding areas. Colony and perch sites as a consequence are located near foraging areas, tend to be patchily distributed throughout the landscape, and concentrate large numbers of birds. There is ongoing debate whether cormorants in colonies share information in some way about feeding areas, but the phenomenon of social feeding in cormorants – as in pelicans – is well observed. Two patterns are known: line hunting, in which cormorants move through the water in a straight line in a rolling flock, and zigzag hunting, in which individuals search and change directions. Line hunting is usually associated with

Maximum size of colony

8

6

7 Sustainable size of colony

1 6

Number of breeding pairs (1000s)

374

5

4 2

5 3

4

3

2

Expanding colonies

1

0 0

200

400

600

800 2

Area of feeding grounds (km ) Figure 2 Colony size (breeding pairs) of Great Cormorants Phalacrocorax carbo in relation to available feeding habitat (0– 20 m water depth) without overlap with their colonies with a range of 20 km from the colony. Maximum colony size (open circles) and, if different, most recent colony size (dark circles) indicate effects of oversaturation of feeding resources. The recent colonies in Denmark (shaded circles) are undersaturated with respect to available feeding resources and are fast expanding in size. Both effects suggest the existence of ‘sustainable’ and a ‘maximum’ density. The oversaturated effect is particularly noticeable in the Brændega˚rd colony (1) in Denmark. Numbers refer to monitored colonies in Denmark (1–3) and The Netherlands (4–6). (Redrawn with permission from van Eerden and Gregersen (1995).)

smaller fish like smelt (Osmerus) and ruffe (Gymnocephalus), whereas zigzag fishing is seen when cormorants pursue larger fish like roach (Rutilus) and Perch (Perca). Anhingidae Darters are fairly large birds (‘goosesized’) with a very long, slender neck with a pronounced ‘S-shaped’ kink, a long spearlike bill, and a large fanlike tail resembling that of a turkey. The taxonomy is contentious: variously one, two,


PELECANIFORMES

four or more species have been described for the family. Darters are primarily aquatic and spend their life in water or on branches overhanging protected, usually fresh water, streams and ponds, especially swamps and marshy areas. Darters are widely distributed throughout tropical and subtropical zones, and are sometimes found in warm temperate habitats. They are the least marine of the Pelecaniformes, but can be observed in brackish coastal waters, mangrove and cypress swamps, and coastal lagoons. They are similar in appearance to cormorants, but are easily distinguished in flight, on roosts, and on the water. Darters are good fliers and often alternate powered flight with glides; cormorants use sustained flapping and seem always at risk of not quite keeping airborne. Darters will use thermals for soaring and will soar for extended periods, perhaps while surveying new areas for feeding or roosting. Their long tail and tapered wings allow a much greater agility in air than shown by shags or cormorants, and they are deft at flying through enclosed swamps and thickets. While it is swimming on the surface, the body of a darter is usually submerged, with only the head and snakelike neck visible, making it obvious why the term ‘snake bird’ is often used for them. The wettable plumage of darters results in considerable loss of body heat under water, and they therefore spend large amounts of time sunning and drying feathers. Darters have several unique anatomical features apparently related to feeding. The distal portions of the upper and lower bill have fine, backward-pointing serrations for holding fish. Modifications to the eighth and ninth cervical vertebrae allow a rightangled kink in the neck, and the extensor ligaments of the neck pass through a ‘pulleylike’ loop that allows a fulcrum for the straight-line stabbing motion with which this species spears its prey. Darters have an unusual bony articulation at the base of skull, allowing an even greater forward acceleration of the head and neck, and a hairlike pyloric valve in the stomach. Unlike all other pelecaniform birds, darters will stab fish with the bill, rather than grasping them between the bill, and then flip the fish into the throat head first. This type of fishing technique is well-suited for the shallow, murky waters they inhabit, and deadly to the slow moving, laterally compressed fish the birds pursue. In North America, centrarchid and cyprinodontid fish are most preferred, but freshwater decapods often are eaten when abundant. Darters are monogamous, as are most Pelecaniformes, with the pair bond lasting many years and breeding pairs returning to the same nest each season. Nesting is usually solitary, but small aggregations are known, and darters will commonly breed

375

in mixed colonies with other waterbirds such as cormorants, herons, egrets, ibises, and storks. Darters are sedentary, rarely dispersing after breeding, but some of the populations breeding in temperate regions will disperse to warmer areas in winter. Owing to their wariness, solitary nature, and preference for impenetrable swampy areas, comparatively little is known about their behavior or natural history.

Ecology Color has been implicated as an important factor in pelecaniform feeding; it is commonly stated that light-colored plumage affords a minimal contrast against daylight or sky, while dark plumage is especially cryptic in dim light. Pelecaniform birds range in color as adults from mostly white (e.g., gannets, tropicbirds) to mostly black (e.g., cormorants, frigatebirds), and with many intermediate shades, colors, and contrasts (e.g., grey plumage in red-footed cormorants, mottled plumage in boobies and brown pelicans). However, juvenile and immature frigatebirds, sulids, and pelicans are generally darker than adults, while in cormorants and darters, and in some frigatebirds and pelicans, the juveniles are lighter. Shags, boobies, and some other pelecaniforms, have white and black contrasting patterns on their abdomens and chests, and in species such as brown booby (Sula variegata) and rock shag (Stictocarbo magellanicus), there are morphological variants with the same patterns. Many have speculated that these plumage patterns relate to greater efficiency in capturing fish or avoiding detection, but other factors seem as important. Melanistic plumage is often more durable than light-colored plumage, and thermoregulation is affected strongly by overall coloration. Much more study is required to understand what may be important in plumage color. Primarily fish-eating birds, the Pelecaniformes nonetheless utilize a diversity of feeding methods, marine and aquatic habitats, and food items (Table 1). Gregariousness is prominent in the group, with some being groups well-known for social feeding (pelicans, cormorants), and with most breeding in large, dense colonies. Nearly every type of fresh water and marine habitat is exploited, but, except for tropicbirds and frigatebirds, the pelecaniform birds are limited to littoral and neritic waters, rarely encountered far from land. Their preference for small schooling fish, and for breeding in large, dense colonies near the feeding grounds, makes their guano production a desirable attribute from the human viewpoint, but interactions with humans are for the most part neutral or negative. Cormorants and


376

PELECANIFORMES

Table 1

Summary of ecological and behavioral characteristics of Pelecaniforms

Sociality Colonial Flocking Solitary Feeding method Pursuit Deep plunge Shallow plunging Surface feeding Skimming Kleptoparasitism Feeding period Day Crepuscular Night Feeding depth Bottom Midwater Surface Feeding habitat Littoral Neritic (marine) Pelagic (marine) Diet Marine v. freshwater Fish Cephalopood Crustacean Scavenging

Phaethontidae

Fregatidae

Pelecanidae

Sulidae

Phalacrocoracidae

Anhingidae

þ ? þ

þ þ þ

þ þ þ

þ þ þ

þ þ þ

þ

þ þ ? þ

þ

þ

þ þ þ

þ þ ?

þ þ ?

þ

þ þ

þ ? ?

þ þ þ

þ þ ?

þ

þ þ ?

þ

þ

þ

M þ þ ?

M þ þ ? þ

pelicans are often – wrongly – implicated in the depredation of fingerling and fry of economically important fish such as trout and salmon, and consequently suffer culling and killing campaigns to reduce their numbers.

Behavior Mostly nonvocal and colonial, pelecaniform birds utilize ground-based visual displays much more prominantly than other waterbirds (see Figure 3). When birds live and feed together in close proximity in colonies, the interactions between individuals is much greater than found in solitary nesting or dispersed breeders. Further, because space is limited at most sites suitable for nesting, colonial birds breed in extremely dense conditions, with territories squeezed to a minimal area around the nest, or even nonexistent. Most of the pelecaniform birds have a basic

þ

þ þ

M/F þ

M þ þ

þ þ

þ

þ

þ þ

þ þ ?

þ þ

þ þ

þ

M/F þ

F þ

þ

þ

courtship sequence that serves to establish the pair bond and maintain it, often for many years. In nearly every case, the male selects the nest site and advertises for a mate; the females choose among males through assessment of no doubt many factors, including the nature of the displays, the excellence of the nest site, and prior history. Copulation occurs on the nest site, usually between bouts of nest building by the male, sometimes by both partners. After the pair bond has been established, both birds take turns guarding the nest, incubating eggs and chicks, and feeding young. With exceptions noted above, parental care stops soon after fledging, and young birds will forage independently soon after leaving the nest. Some pelecaniform birds will begin breeding as soon as the second year (e.g., cormorants), although three years of age seems to be the norm, while in other species (e.g., Abbott’s booby) many years may pass before individuals breed.


PELECANIFORMES

(A)

(B)

(D)

(H)

(C)

(E)

(I)

377

(F)

(J)

(K)

(G)

(L)

Figure 3 Behavioral displays in the rock shag (Stictocarbo magellanicus). Appeasement displays: (A) landing display approaching a nesting cliff; (B) postlanding or hopping; (C) nest-touching after a threat display. Courtship displays: (D) beginning and (E) ending phase of darting by the male; (F) beginning and (G) ending of throat-clicking by a pair. Pairing displays that terminated in copulation: (H) initial phase of wing-waving by the male (bottom) after approach by the female (top); (I) beginning phase of hop by female, male is wing-waving; (J) conclusion of hop by female followed by nest-indicating, male is wing-waving; (K) full neck extension by male during wing-waving, female is gaping; (L) bill-biting by female and bowing by male. (Redrawn with permission from Siegal-Causey (1986).)


378

PELECANIFORMES

Status and Conservation Few of the Pelecaniformes are globally threatened, although some island and inland populations of pelecaniform birds have become extinct in the face of habitat loss and contamination. In some species (e.g., double-crested cormorant, white pelican), numbers are increasing with near-exponential growth, while in others (e.g., Dalmatian pelican) population abundance is rapidly decreasing. Island species and populations have been especially hard-hit by the introduction of terrestrial predators such as cats, rats, and pigs; loss of traditional breeding habitat has played a greater role in aquatic and inland-nesting species. Some islands such as Christmas Island (Indian Ocean) and Ascension Island (Atlantic Ocean) are breeding grounds for several pelecaniform species, and their recent protected status has promised a much better future than might have been possible only a few years earlier.

See also Seabird Conservation. Seabirds as Indicators of Ocean Pollution. Seabirds: An Overview.

Further Reading Ashmole NP (1971) Sea bird ecology and the marine environment. In: Farner DS, King JR, and Parkes KC (eds.) Avian Biology, Vol. 1. New York: Academic Press.

Cairns DK (1986) Plumage colour in pursuit diving seabirds: why do penguins wear tuxedos? Bird Behaviour 6: 58--65. Diamond AW (1975) The biology of tropicbirds at Aldabra Atoll, Indian Ocean. Auk 92: 16--39. Frederick PC and Siegel-Causey D (2000) Anhinga (Anhinga anhinga). In: Poole A and Gill F (eds.) The Birds of North America, No. 522. Philadelphia, PA: The Birds of North America, Inc. Johnsgard PA (1993) Cormorants, Darters, and Pelicans of the World. Washington, DC: Smithsonian Institution Press. Nelson JB (1978) The Sulidae: Gannets and Boobies. Aberdeen: Oxford University Press. Nelson JB (1985) Frigatebirds, aggression, and the colonial habit. Noticias Galapagos 41: 16--19. Siegel-Causey D (1986) Behaviour and affinities of the Magellanic Cormorant. Notornis 33: 249--257. Siegel-Causey D (1988) Phylogeny of the Phalacrocoracidae. Condor 90: 885--905. Siegel-Causey D (1996) The problem of the Pelecaniformes: molecular systematics of a privative group. In: Mindell DL (ed.) Avian Molecular Evolution and Systematics. New York: Academic Press. Van Eerden MR and Voslamber B (1995) Mass fishing by cormorants Phalacrocorax carbo sinensis at Lake IJsselmeer, The Netherlands: a recent successful adaptation to a turbid environment. Ardea 83: 199--212. Van Tets GF (1965) A comparative study of some social communication patterns in the Pelecaniformes. Ornithological Monographs, No. 2. Washington, DC: American Ornithological Union.


PENETRATING SHORTWAVE RADIATION C. A. Paulson and W. S. Pegau, Oregon State University, Corvallis, OR, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2136–2141, & 2001, Elsevier Ltd.

Introduction The penetration of solar radiation into the upper ocean has important consequences for physical, chemical, and biological processes. The principal physical process is the heating of the upper layers by the absorption of solar radiation. To estimate the solar radiative heating rate, the net downward shortwave irradiance entering the ocean and the rate of absorption of this energy as a function of depth must be determined. Shortwave irradiance is the flux of solar energy incident on a plane surface (W m2). Given the downward shortwave radiance field just above the sea surface, the rate of shortwave absorption as a function of depth is governed primarily by sea surface roughness, molecular structure of pure sea water, suspended particles, and dissolved organic compounds. The optical properties of pure sea water are considered a baseline; the addition of particles and dissolved compounds increases absorption and scattering of sunlight. The dissolved organic compounds are referred to as ‘colored dissolved organic matter’ (CDOM) or ‘yellow matter’ because they color the water yellowish-brown. The source of CDOM is decaying plants; concentrations are highest in coastal waters. Suspended particles may be of biological or geological origin. Biological (organic) particles are formed as the result of the growth of bacteria, phytoplankton, and zooplankton. The source of geological (inorganic) particles is primarily weathering of terrestrial soils and rocks that are carried to the ocean by the wind and rivers. Phytoplankton particles are the main determinant of optical properties in much of the ocean and the concentration of chlorophyll associated with these plants is used to quantify the effect of phytoplankton on optical properties. Case 1 waters are defined as waters in which the concentration of phytoplankton is high compared with inorganic particles and dissolved compounds; roughly 98% of the world ocean falls into this category. Case 2 waters are waters in which inorganic particles or CDOM are the dominant influence on optical properties. Case 2 waters are usually coastal, but not all coastal water is case 2.

Inherent optical properties (IOPs), such as attenuation of a monochromatic beam of light, depend only on the medium, i.e. IOPs are independent of the ambient light field. It is often assumed that the inherent optical properties of the upper ocean are independent of depth. To the extent that the upper ocean is well-mixed, the assumption of homogeneous optical properties is reasonable. However, in the stratified layers below the mixed layer, the concentration of particles is likely to vary with depth. The consequences of this variation on radiant heating are expected to be small because the magnitude of the downward irradiance decreases rapidly with depth. Apparent optical properties (AOPs) depend both on the medium and on the directional properties of the ambient light field. Some AOPs, such as the ratio of downward irradiance in the ocean to the surface value, are sufficiently independent of directional properties of the light field to be useful for characterizing the optical properties of a water body.

Albedo Albedo, A, is the ratio of upward to downward short-wave irradiance just above the sea surface and is defined by: A

Eu Ed

where Eu and Ed are the upwelling and downward irradiances just above the sea surface, respectively. The upwelling irradiance is composed of two components: emergent irradiance due to back-scattered light from below the sea surface; and irradiance reflected from the sea surface. Emergent irradiance is typically o10% of reflected irradiance. The rate at which net short-wave irradiance penetrates the sea surface is the rate at which the sea absorbs solar energy and is given by: ð1 AÞ Ed ðWm2 Þ R.E. Payne analyzed observations to represent albedo as a function of solar altitude y and atmospheric transmittance G defined by: G ¼ Ed r2 =S siny where S is the solar constant (1370 W2) and r is the ratio of the actual to mean Earth–sun separation. The transmittance is a measure of the effect of the


379

380

PENETRATING SHORTWAVE RADIATION

Table 1 Latitude 801N 701N 601N 501N 401N 301N 201N 101N 01 101S 201S 301S 401S 501S 601S

Mean albedos for the Atlantic Ocean by month and latitude Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

0.28 0.11 0.10 0.09 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06

0.41 0.12 0.10 0.09 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07

0.33 0.15 0.09 0.08 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.08

0.14 0.10 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.08 0.08 0.11

0.10 0.08 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.08 0.09 0.10 0.13

0.09 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.09 0.11 0.13

0.08 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.08 0.10 0.11 0.27

0.08 0.09 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.08 0.08 0.07

0.12 0.11 0.07 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.08 0.08

0.25 0.10 0.08 0.08 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.07 0.07 0.07

0.16 0.11 0.10 0.08 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06 0.06

0.44 0.12 0.11 0.09 0.07 0.07 0.06 0.06 0.06 0.06 0.06 0.06 0.06

(Reproduced with permission from Payne, 1972.)

Earth’s atmosphere, including clouds, on the radiance distribution at the Earth’s surface. If there were no atmosphere, the transmittance would equal one and the radiance would be a direct beam from the sun. For very heavy overcast, the transmittance can be o0.1 and the downward radiance distribution may be approximately independent of direction. Payne’s observations were taken from a fixed platform off the coast of Massachusetts from 25 May to 28 September. Solar altitude ranged up to 721 and the mean wind speed was 3.7 m s1. The transmittance varied from near zero to about 0.75. Payne fitted smooth curves to the albedo as a function of transmittance for observations in intervals of 21 of solar altitude and 0.1 in transmittance. The smoothed albedos ranged from 0.03 to 0.5. Payne extrapolated the curves to values of solar altitude and transmittance for which there were no observations by use of theoretical calculations of reflectance for a sea surface roughened by a wind speed of 3.7 m s1. Albedo was obtained for the limiting case of G ¼ 1 by adding 0.005 to the calculated reflectance to account for the irradiance emerging from beneath the surface. Wind speed affects albedo through its influence on the surface roughness. The clear-sky reflectivity from a flat water surface is a strong function of solar altitude for altitudes o301. As wind speed and roughness increase, reflectivity decreases because of the nonlinear relationship between reflectivity and the incidence angle. Payne investigated the variation of albedo with wind speed for solar elevations from 171 to 251 and found that albedo decreased at the rate of 2% per meter per second increase in wind speed.

Wind speed may also affect albedo by the generation of breaking waves that produce white caps. The albedo of whitecaps is higher, on average, than the whitecap-free sea surface. Hence the qualitative effect of whitecaps is to increase the albedo. Monahan and O’Muircheartaigh estimate an increase of 10% in albedo due to whitecaps for a wind speed of 15 m s1 and of 20% for a wind speed of 20 m s1. Payne estimated monthly climatological values of albedo at 101 latitude intervals for the Atlantic Ocean (Table 1). For the range of latitudes from 401S to 401N, mean albedo varies from a minimum of 0.06 at the equator in all months to a maximum of 0.11 at 401S and 401N for the months containing the winter solstice. The symmetry exhibited by mean albedo values for the winter solstice in the North and South Atlantic suggests that the values in Table 1 may be reasonable estimates for the world ocean.

Spectrum of Downward Irradiance The spectrum of downward short-wave irradiance at various depths in the ocean (Figure 1) illustrates the strong dependence of absorption on wavelength. The shape of the spectrum at the surface is determined primarily by the temperature of the sun (Wien displacement law) and wavelength-dependent absorption by the atmosphere. The area under the spectrum at each depth is proportional to the downward irradiance. The downward irradiance at a depth of 1 m is less than half the surface value because of preferential absorption at wavelengths in excess of 700 nm. The downward irradiance below 10 m depth is in a relatively narrow band centered



381

_

_

Downward irradiance (W m 2 nm 1)

1.5

1.0

0.5 10 m Surface 1m 1 cm

100 m

500

1000

1500

2000

2500

Wavelength Figure 1 The spectrum of downward irradiance Ed(z, l) in the sea at different depths. (Adapted with permission from Jerlov, 1976.)

Modeled Irradiance 10

4

Radiative transfer models are useful tools for investigating the characteristics of underwater light fields and their dependence on suspended particles and dissolved organic matter. The Hydrolight model, constructed by Mobley, is used to illustrate the diffuse attenuation coefficient for downward irradiance for three cases with different concentrations of chlorophyll and values of CDOM beam attenuation at 440 nm (Figure 2). The diffuse attenuation coefficient for downward irradiance Kd is defined by

_ Kd (m 1)

103 102 101 100 _ 10 1

200

500

800

1100 1400 1700 2000 2300 Wavelength (nm)

Figure 2 The diffuse attenuation coefficient for downward irradiance in sea water versus wavelength. Data for the wavelength band from 350 to 800 nm are from the Hydrolight radiative transfer model with the following conditions: depth 10 m, solar altitude 601, cloudless sky, wind speed 2 m s1. The three lines (thin, thick, dashed) result from specified concentrations of chlorophyll (0.05, 1, 10 mg m3) and the beam attenuation coefficient at 440 nm for CDOM (0.01, 0.05, 0.1 m1). Data for the wavelength band from 800 to 2200 nm are for pure water. (Tabulations taken from Kuo et al., 1993.)

near 470 nm (blue-green). Pure sea water is most transparent near a wavelength of 450 nm which, by coincidence, is close to the peak in the downward irradiance spectrum at the surface.

Kd ðz;lÞ ¼

d lnEd ðz;lÞ dz

where z is the vertical space coordinate, zero at the surface and positive upward, and l is the wavelength. If Kd is independent of z, monochromatic irradiance decreases exponentially with depth, consistent with Beer’s law. Kd increases five orders of magnitude as wavelength increases from 500 to 2000 nm (Figure 2), consistent with the strong dependence of absorption on wavelength shown in Figure 1. In the wavelength band centered around 500 nm, Kd varies by a factor of 10 between the least absorbent and most absorbent cases. The order of magnitude variation in Kd at 500 nm among the three cases (Figure 2) has a dramatic effect


382


0

0

10 60

Depth (m)

Depth (m)

30

90

120

150 _ 10 4

20

30 _3

_1

_ 10 2 E / E(0)

10

10

100

40 0.01

Figure 3 Net irradiance E ¼ (Ed Eu) versus depth from the Hydrolight model for the conditions specified in the caption for Figure 2.

on net irradiance (Figure 3) because the peak of the solar spectrum is near 500 nm (Figure 1). Net irradiance, E, is the difference between net (wavelength integrated) downward and net upward irradiance. The difference between downward and net irradiance is negligible for most purposes because upward irradiance is typically 3% of downward irradiance within the ocean. At depths where the net downward irradiance is o10% of its surface value, the decay with z is approximately exponential because light at these depths is roughly monochromatic. The three cases with different optical properties (Figures 2 and 3) can be characterized biologically as oligotrophic, mesotrophic, and eutrophic (ranging from least to most absorbent). Oligotrophic water has low biological production and low nutrients. Eutrophic water has high biological production and high nutrients and mesotrophic water is moderate in both respects. The oligotrophic case illustrated in Figures 2 and 3 is typical of open-ocean water. Mesotrophic and eutrophic water are likely to be found near the coast.

Parameterized Irradiance versus Depth Observations show that downward short-wave irradiance decreases exponentially with depth below a depth of about 10 m (Figure 4) as the result of absorption by the overlying sea water of all irradiance except for blue-green light. This suggests that Ed can be approximated by a sum of n exponential terms: Ed =E0 ¼

n X

Fi expðKi zÞ

i¼l n X i¼l

Fi ¼ 1

½1

1.0

0.1 Ed / Ed(0)

Figure 4 Measurements in the upper 40 m of downward irradiance normalized by downward irradiance just below the surface. The measurements were made in the North Pacific (351N, 1551W) in February. The open circles are the average of five sets of observations with similar irradiance profiles; solar altitude ranged from 301 to 381 and the sky was overcast. The plus signs show one set of observations for which solar altitude was 161 and the sky was clear. The curves are the sum of two exponential terms fitted to the observations (see eqn [1]). (Adapted with permission from Paulson and Simpson, 1977.)

Table 2 Values of parameters determined by fitting the sum of two exponential terms (see eqn [1]) to values of downward irradiance which define Jerlov’s (1976) water types Water type

F1

K1 (m1)

K2 (m1)

C (mg m1)

I IA IB II III

0.32 0.38 0.33 0.23 0.22

0.036 0.049 0.058 0.069 0.13

0.8 1.7 1.0 0.7 0.7

0–0.01 B0.05 B0.1 B0.5 B1.5–2.0

Values of downward irradiance in the upper 100 m were used, except that values were limited to the upper 50 m for type I because of a change in slope below 50 m. F2 is 1 F1. (Adapted from Paulson and Simpson, 1977.) The column labeled C is the approximate chlorophyll concentration for each water type as determined by Morel (1988).

where Fi is the fraction of downward irradiance in a wavelength band i and Ki is the diffuse attenuation coefficient for the same band. The leading term in eqn [1] is defined as the short wavelength band that describes the exponential decay below 10 m (Figure 4). At least one additional term is required. A total of two terms fits the observations in Figure 4 reasonably well, although accuracy in the upper few meters is lacking.



383

Depth (m)

0

20

III

II

I

40

0.001

0.01

0.1

1.0

Ed / Ed(0) Figure 5 Normalized downward irradiance versus depth for water types I, II, and III. The data (open circles, squares, and triangles) are from Jerlov (1976) and the curves are a fit to the data with the parameters given in Table 2. (Adapted with permission from Paulson and Simpson, 1977.)

Jerlov has proposed a scheme for classifying oceanic waters according to their clarity. He defined five types (I, IA, IB, II, and III) ranging from the clearest open-ocean water (type I) to increasingly turbid water. Parameters for the sum of two exponential terms (eqn [1]) fit to the values of downward irradiance which define the Jerlov water types are given in Table 2 and plots of values and fitted curves are shown in Figure 5. Apart from systematic disagreement in the upper few meters, the fit is good. Differences in the values of K2 in Table 2 are not significant. Water types IA and IB are not shown in Figure 5. However, the fits to the observations shown in Figure 4 yield parameters very similar to those for types IA and IB (open circles and plus signs, respectively, in Figure 4). Most open-ocean water is intermediate between types I and II. The approximate chlorophyll concentration for each of the Jerlov water types is given in Table 2. The Jerlov water types can be compared to the oligotrophic, mesotrophic, and eutrophic cases illustrated in Figures 2 and 3. The oligotrophic case is intermediate between types IA and IB. The mesotrophic case is similar to type III and the eutrophic case is similar to Jerlov’s coastal type 5 water. The sum of two exponential terms (eqn [1]) is adequate for modeling purposes when the required vertical resolution is a few meters or greater. For a vertical resolution of 1 m or less, additional terms are

required. These additional terms can be constructed with knowledge of the surface irradiance spectrum (Figure 1) and the diffuse attenuation coefficient versus wavelength (Figure 2).

See also Bio-Optical Models. Coastal Circulation Models. General Circulation Models. Heat and Momentum Fluxes at the Sea Surface. Inherent Optical Properties and Irradiance. Photochemical Processes. Radiative Transfer in the Ocean. Upper Ocean Heat and Freshwater Budgets. Upper Ocean Time and Space Variability. Wind- and BuoyancyForced Upper Ocean.

Further Reading Dera J (1992) Marine Physics. Amsterdam: Elsevier. Jerlov NG (1976) Marine Optics. Amsterdam: Elsevier. Kou L, Labrie D, and Chylek P (1993) Refractive indices of water and ice in the 0.65- to 2.5-mm spectral range. Applied Optics 32: 3531--3540. Kraus EB and Businger JA (1994) Atmosphere–Ocean Interaction, 2nd edn. New York: Oxford University Press. Mobley CD (1994) Light and Water. San Diego: Academic Press. Mobley CD and Sundman LK (2000) Hydrolight 4.1 User’s Guide. Redmond, WA: Sequoia Scientific.


384


Monahan EC and O’Muircheartaigh IG (1987) Comments on glitter patterns of a wind-roughened sea surface. Journal of Physical Oceanography 17: 549--550. Morel A (1988) Optical modeling of the upper ocean in relation to its biogenous matter content. Journal of Geophysical Research 93: 10749--10768. Paulson CA and Simpson JJ (1977) Irradiance measurements in the upper ocean. Journal of Physical Oceanography 7: 952--956.

Payne RE (1972) Albedo of the sea surface. Journal of Atmospheric Sciences 29: 959--970. Thomas GE and Stamnes K (1999) Radiative Transfer in the Atmosphere and Ocean. Cambridge: Cambridge University Press. Tyler JE and Smith RC (1970) Measurements of Spectral Irradiance Underwater. New York: Gordon and Breach.


PERU–CHILE CURRENT SYSTEM J. Karstensen, Universita¨t Kiel (IFM-GEOMAR), Kiel, Germany O. Ulloa, Universidad de Concepcioń, Concepcioń, Chile & 2009 Elsevier Ltd. All rights reserved.

the wind stress. As the wind field has a pronounced seasonal signal it is useful to contrast the situation for austral summer (January to March) and austral winter (July to September) (Figure 2). On the large scale, the wind variability over the PCCS originates

0° S

Introduction The Peru–Chile Current System (PCCS) comprises the surface and subsurface flows in the eastern boundary current system along the Chilean and Peruvian coasts (Figure 1). The PCCS hosts four major currents: two flowing equatorward at the surface and two flowing poleward, one at the surface and one at subsurface. For individual currents of the PCCS a variety of names exist: the offshore equatorward flow has been called Peru Current (the name that relates the current into a regional geographical context is indicated in italic for purposes of clarity here), Oceanic Peru Current, Mentor Current, Humboldt Current (in honor of the German naturalist Alexander von Humboldt), or PeruHumboldt Current. The equatorward flow near the coast has been called Peru Coastal Current, Chile Coastal Current, and sometimes the Inshore Peru Current. The poleward surface flow has been called the Peru–Chile Countercurrent while the poleward subsurface flow has been called Peru–Chile Undercurrent or Gunther Current (in honor of the British investigator Eustace Rolfe Gunther). The PCCS has the largest meridional extent of all eastern boundary currents stretching from about 51 S to south of 421 S. The PCCS is similar to the other eastern boundary current regions in a number of aspects, for example, wind forcing, strong upwelling, and enhanced productivity. The PCCS has a rather tight connection to the equatorial Pacific and the globally strongest mode of interannual variability; the ‘El Ninõ/Southern Oscillation (ENSO)’ propagates via atmospheric and oceanic pathways into the PCCS and provokes specific physical and ecological responses. In fact, ENSO has become well known due to its severe socioeconomic consequences on the fishery activity off Peru and Chile.

The Currents The primary driver of the currents in the PCCS is the frictional force of the wind on the ocean’s surface,

EUC

PCoastalC 10°

PCCC

20°

PCC

CCoastalC 30°

SPC

40° S

85° W

80°

75°

70° W

Figure 1 The principal surface flows path in the PCCS and associated temperature anomaly (red indicates warm current, blue indicates cold): Chile Coastal Current (CCoastalC), Peru Coastal Current (PCoastalC), Peru–Chile Countercurrent (PCCC), and Peru–Chile Current (PCC). On the large scale, supplies stem from the South Pacific Current (SPC) and the Equatorial Undercurrent (EUC). The circles at about 301 S symbolize intensive eddy formation. The subsurface poleward Peru–Chile Undercurrent (PCUC) is not drawn but approximately follows the CCoastalC and PCosatalC.


385

386

PERU–CHILE CURRENT SYSTEM

Austral Summer (Jan.−Mar.) (a)

(b)

(c)

(d)

0° S

0° S

0° S

0° S

10°

10°

10°

10°

20°

20°

20°

20°

30°

30°

30°

30°

40° S

40° S

40° S

5 m s−1 40° S

85° W

80°

75°

70° W

85° W

80°

75° −1

70° W

85° W

80°

−375 −300 −225 −150 −75

0

75°

70° W

85° W

SST (°C)

w (m yr )

75

150

10

15

20

80°

75° −3

70° W

Chl a (mg m ) 25

0.05 0.1 0.2

Austral Winter (Jul.−Sep.) (e)

(f)

(g)

(h)

0° S

0° S

0° S

0° S

10°

10°

10°

10°

20°

20°

20°

20°

30°

30°

30°

30°

40° S

40° S

40° S

0.5

1

2

5

10

25

5 m s−

40° S

85° W

80°

75°

70° W

85° W

80°

75°

w (m yr−1)

−375 −300 −225 −150 −75

70° W

85° W

80°

75°

70° W

0

75

150

10

15

20

85° W

80°

75° −3

70° W

Chl a (mg m )

SST (°C) 25

0.05 0.1 0.2

0.5

1

2

5

10

25

Figure 2 Wind field vectors (QuikSCAT) (a, e), Ekman pumping velocity (QuikSCAT) (b, f; upwelling negative), sea-surface temperature (AVHRR) (c, g), and surface chlorophyll a concentration (SeaWiFS/MODIS) (d, h) for austral summer (upper) and austral winter (lower).



from the meridional movement of the eastern Pacific Subtropical Anticyclone from approx. 321 S in austral summer to 281 S in austral winter (Figures 2(a) and 2(e)). In the northern part of the region, off Peru, the movement of the Intertropical Convergence Zone (ITCZ) from approx. 51 to 101 N in austral winter is important. For the southern part of the region, in particular south of about 301 S, the alongshore propagation of atmospheric low-pressure systems is of importance. They originate from the west wind zone and upon impinging the coast propagate northward or southward steered by the Andes mountain range. Over the PCCS the wind field has its dominant component parallel to the coast. One of the reasons for this is the configuration of the continent and the ocean which maintains through differential heating a large-scale pressure difference between the South American continent and the Southeast Pacific Ocean. The pressure difference leads to equatorward winds along the South American west coast. In addition, the Andes Mountain range, which closely follows the coastline, steers the low-level winds parallel to the coast. The along-shore equatorward wind stress generates an offshore Ekman transport and therefore intensive upwelling occurs along the coast. For the region off Peru, upwelling is strongest in austral winter (Figure 2(f)) while for the region between 301 and 401 S, off Chile, almost continuous upwelling is strongest in austral summer (Figure 2(b)). There is weaker but almost continuous upwelling for the region in between. South of about 421 S, winds are predominately poleward and associated with coastal downwelling. The offshore transport of surface water causes a trough in the sea surface height (SSH) along the coast (Figure 3, left). Such a trough has dynamical consequences as it causes a zonal (west to east) pressure gradient force. On a rotating Earth and in cases where ocean bottom friction can be neglected, the pressure gradient force is balanced by the Coriolis force acting perpendicular (in the Southern Hemisphere to the left) to the movement. Oceanographers (and meteorologists) refer to the equilibrium of pressure gradient force and Coriolis force as geostrophy. To bring the zonal pressure gradient force along the coast in the PCCS region into a geostrophic balance an equatorward flow is required. This flow, which is in the same direction as the wind, is here associated with the Peru Coastal Current (PCoastC) and the Chile Coastal Current (CCoastC) (Figure 1). The coastal currents exist over the shelf area and reach only to shallow depths (less than 80 m). They have typical flow speeds of about 0.1 m s 1 in the northern part of the PCCS, decreasing to 0.02 m s 1

387

in the southern part. In austral winter, when atmospheric low-pressure systems impinge on the coast at about 301 S, episodic flow reversals of the CCoastC have been observed. The coastal currents are identified through a minimum in sea surface temperature. This is because they carry colder water from the south northward but more important is the entrainment of the cold upwelling waters into the flow. Further offshore, at the boundary between the upwelling and downwelling favorable regions (Figures 2(b) and 2(f)), the Peru–Chile Current (PCC) is found (Figure 1). This flow is independent of the coastal current and is sometimes considered to be the eastern branch of the subtropical gyre circulation of the South Pacific. Here the contrasting Ekman regimes modify the internal density field such that an equatorward flow is required. Again the PCC can be recognized at the surface from lower-than-average sea surface temperatures associated with the equatorward advection of colder waters. About 100–300 km offshore, and between the equatorward flowing PCC and the coastal currents (CCoastalC and PCoastalC) is a poleward flow, the Peru–Chile Countercurrent (PCCC). The PCCC is most pronounced off the Peruvian Coast and transports warm and saline water of equatorial origin poleward. The driving mechanisms for the PCCC are not fully understood. It has been suggested that the flow is driven by Sverdrup dynamics as a result of the cyclonic wind stress curl. Other mechanisms, such as eddy/mean flow interaction, buoyancy forcing, or the multiple pattern of Ekman-driven up- and down welling, in particular off the Peruvian Coast, may contribute to the forcing of the current. Upwelling in the PCCS coastal region brings cold (Figures 2(c) and 2(g)) and nutrient-rich water from depths below 150 m to the surface. Local upwelling rates are higher than 600 m yr 1 (Figures 2(b) and 2(f)), but it is clear that the transport is not only vertical but also ‘horizontal’ or, better to say, lateral, along surfaces of constant density, minimizing the required energy for the flow. The most intense upwelling in the PCCS occurs along the coast and the lateral transport at depths has a pronounced component toward the coast. This onshore transport must be in geostrophic balance and hence a north/ south (meridional) pressure gradient along the coast is required for balancing the transport, with higher pressure toward the equator. This equilibrium of onshore transport and meridional pressure gradient is disturbed at the shelf break, acting as a barrier for the onshore flow and the flow stops. When the flow comes to a halt, the pressure gradient force is no longer balanced by the Coriolis force but drives a poleward meridional flow near and above the shelf.


388


0° S

200 m 500 m

10°

20°

30°

40° S

85° W

80°

75°

70° W

0

Mean dynamic ocean topography (cm) −20 −15 −10 −5

0

5

10

15

100

200

Shelf width (km)

20 25

Figure 3 (Left) Absolute sea surface height (based on joint analysis of drifter data, satellite altimeter data, wind data, and a model geoid) and (right) width of the shelf with depth less than 200 m (blue) and less than 500 m (black). Black dotted line indicates political borders.

Now bottom-friction forces balance the pressure gradient force. This subsurface flow along the shelf in the PCCS is associated with the Peru–Chile Undercurrent (PCUC). The PCUC is typically found between 50- and 4400-m depth, but it may sporadically even outcrop at the sea surface. The PCUC has flow speeds between 0.1 and 0.7 m s 1. It is ultimately fed by the Pacific Equatorial Undercurrent (EUC) and thus transports warm and saline water, which is low in oxygen and high in nutrients content, southward. These unique characteristics of the water

in the PCUC allow us to trace the flow to south of 421 S all along the coast. The divergence of the eastward-flowing South Pacific Current (SPC), when it approaches the coast at about 401 S, also drives upwelling that may reach several hundreds of kilometers offshore. This upwelling is weaker than the coastal one and its location and intensity move southward in the Southern Hemisphere summer following the movement and intensification of the South Pacific Current.


389


The Ecosystem The rich biological productivity of the PCCS depends primarily on the wind-driven coastal upwelling that brings colder and nutrient-rich subsurface waters into the euphotic zone. The upwelling is, in part, fed by water from the oxygen-minimum zone, located below the PCCS and the water has particularly high nutrient levels due to the remineralization of organic matter. When this nutrient-rich water is upwelled into the surface layer it is utilized by phytoplankton along with dissolved CO2 (carbon dioxide) and light energy from the Sun. The productivity can be seen from the high chlorophyll a (Chl a) concentrations at the surface (Figures 2(d) and 2(h)). The phytoplankton is, in turn, available for other components of the marine food web, including zooplankton and fish, making the region exceptionally productive. The high phytoplankton biomass resulting from the fertilizing effect of the upwelling is evident in the coastal band off Peru and off southern Chile (south of about 301 S). For both regions, the temporal variability in near-shore sea surface temperature and Chl a concentrations are in phase with each other, suggesting that upwelling-favorable winds are the main driver for productivity here. However, it is not only the local upwelling intensity but also the supply of nutrient-rich waters to the region via the PCUC that maintains the productivity. The shelf is particularly wide in these regions (Figure 3, right) and the PCUC broadens and, in turn, facilitates the entrainment of the nutrient-rich waters into the euphotic zone over a wider area. In contrast, upwelling is weaker in the northern Chile region and there is virtually no shelf which makes the nutrient supply to the euphotic zone less effective and the productivity lower. Offshore Chl a variability is not in phase with the coastal variability (cf. Figures 2(d) and 2(h)). The reason for this is not yet clear, but possible mechanisms include differences in the timing of the maximum in nutrient supply and the effect of photoadaptation to the seasonal changes in the light field. During the upwelling season, waters over the broad continental shelves off northern and central Peru as well as south of 301 S off Chile have extremely high Chl a concentrations (420 mg m 3), with large diatoms dominating the phytoplankton community. In contrast, over the narrow shelves in southern Peru and northern Chile, the chlorophyll concentrations are lower (o2 mg m 3) and smaller phytoplankton species, including prymnesiophytes and cyanobacteria, dominate. Measurements of dissolved iron and experiments with iron addition off Peru have shown that primary productivity in the low-chlorophyll areas of the PCCS are limited by iron. This situation is similar

to that found in the so-called high-nutrient lowchlorophyll (HNLC) regions, such as the equatorial eastern Pacific, the North Pacific, and the Southern Ocean. Thus, variability in the productivity of the PCCS cannot be due to changes in the intensity of the upwelling winds alone, but also due to changes in the availability of iron. Zooplankton is considered the next trophic level in the food chain, going from phytoplankton to fish and top predators. In the PCCS, zooplankton is usually dominated by copepods and euphasids, but other groups, including gelatinous organisms, can become important in certain periods of the year. Nearly 60 species of copepods have been identified in the PCCS. Some of them are endemic, like the abundant Calanus chilensis, which is found to be associated with the areas of intensive upwelling. Within the euphasids, the most abundant one is the endemic species Euphausia mucronata, which has been shown to migrate vertically through the oxygen-minimum

25 Jan. to 01 Feb. 2003 24° S

26°

28°

30°

32°

34°

36°

38° S 78° W

74°

76°

70° W

72°

Chl a (mg m−3) 0.05 0.1 0.2

0.5

1

2

5

10

25

Figure 4 Weekly composite of MODIS satellite-derived Chl a concentrations for the period 25 Jan. to 1 Feb. 2003. White areas indicate data gaps (e.g., cloud cover).


390


zone. Another important one, appearing especially during El Ninõ years, is Euphausia eximia. In order to avoid being permanently advected offshore and away from the food-rich upwelling regions along the coast, zooplankton often make use of their vertical migratory capacities to remain at the coast and ‘travel back’ on shore with the aid of the onshore subsurface flow. However, this mechanism does not appear to be widely ‘used’ by zooplankton species in the PCCS, presumably due to the presence of the intense and shallow oxygen-minimum zone that prevents many of them from vertical migration. Similarly, in other coastal systems, copepods show seasonal reproductive cycles with a few well-defined generations per year. In the PCCS, in contrast, they show continuous reproduction and multiple generations throughout the year. As there are virtually no food limitations in the upwelling areas along the coast, even during El Ninõ years, the near-coastal zooplankton growth rates appear to be mainly

temperature-controlled. Recent evidence suggests that food quality can also be important for zooplankton reproduction and recruitment. In the benthic environment of the continental shelf and upper slope with oxygen-deficient conditions, extensive mats of the giant, sulfide-oxidizing, nitratereducing bacteria Thioploca spp. can develop. This contrasts with the low abundance and diversity of the macrofauna found there. However, the benthic fauna that inhabits this particular environment, like certain polychaetes, presents particular adaptations to cope with the suboxia (dissolved oxygen content below B10 mmol kg 1) or even anoxia (no oxygen available), and the building up of toxic hydrogen sulfide in the sediments.

The Variability Snapshots of the ocean’s surface property fields (Figure 4) reveal the existence of swirls and

2

0

PDO

1

−1 −2 0.8 Anchoveta (2 × 107 t ) Sardines (1 × 107 t )

Fish production (Chile + Peru)

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

1960

1965

1970

1975

1980 Year

1985

1990

1995

2000

Figure 5 Time series of (upper) PDO index based on North Pacific sea surface temperatures and (lower) anchoveta (Peruvian anchovy; blue) and sardines (red) fish production as given by the Food and Agriculture Organization of the United Nations (FAO).



filaments, manifested in sharp gradients. They are associated with the generation and propagation of mesoscale eddies. Eddies may form from flow instabilities of the coastal currents, from westward propagating Rossby waves and from other coastal trapped waves (CTWs) propagating along the coast. The eddies act as an efficient distributor of nutrients and biomass from the coast to the open ocean into the so-called ‘coastal transition zone’. This zone is particularly large between 251 and 401 S and may reach as far as 600–800 km off shore. One important reason for the flow instability is the intermittent character of the upwelling intensity, with timescales from 3 to 10 days. This is true for the whole coast, but in the zone from 251 to 401 S a strong density gradient is formed between the coast and the open waters, in particular in summer through the increase in coastal upwelling. During intermittent upwelling events, instabilities are generated which, in turn, release the filaments of cold,

391

nutrient-rich, and potentially productive waters far from the coast. After relaxation of the wind, these filaments are dissipated. Interannual and decadal variability also plays an important role in the region. In particular, the effects of ENSO are evident. This influences the large-scale atmospheric as well as oceanic circulation. During an ENSO event, the wind field changes at the equator and warm waters of the western equatorial Pacific ‘warm pool’ region propagate to the eastern Pacific by means of a Kelvin wave and radiates in CTW poleward along the eastern boundary. As a consequence, the temperature (as well as the dissolved oxygen contents) of the surface ocean in the PCCS are significantly higher with a simultaneous deepening of the main thermocline. This deepening causes a reduction of the nutrient supply to the euphotic zone and thus a decrease in primary productivity, which has a negative impact all the way through the food chain.

Oceanic waters Coastal waters

Wi

nd

Mesoscale eddies

Eddy pumping ~50 m

re flow Co ard ew Pol

Oxygenminimum zone

g

llin

we

Up

~500 m

Figure 6 Schematic of physical processes and peculiarities of the hydrographic stratification in the PCCS. Printed with permission of Samuel Hormaza´bal.


392


It has been documented that the offshore extension of the high-productivity band reduces during El Ninõ years and a deepening of the oxycline (the transition between the surface water region with high oxygen content and the oxygen-minimum zone below) occurs. Species that are dependent on the oxygenated conditions of the surface waters then occupy a broader depth range or concentrate in deeper depths. Changes in migratory routes and nursery distributions have also been observed in some species such as jack mackerel. The marked differences between the ecological effects of particularly strong El Ninõ events, like the 1982–83 versus the 1997–98, suggests that longer variability (e.g., Pacific Decadal Oscillation (PDO)) plays an important role in regulating ecological responses in the PCCS (Figure 5). As a first step, and associated with the PDO, two environmental stages have been identified: a warmer stage, termed ‘El Viejo’, where sardine abundance is promoted and a colder stage, termed ‘La Vieja’, where anchovy abundance is promoted. It is the interaction of different time- and space scales which makes it difficult to fully understand the processes that control the physical and ecological functioning of the PCCS, as well as those of other eastern boundary current systems. A schematic of important physical processes for the PCCS is shown in Figure 6. The spectrum of variability is very wide. It covers spatial variability of a few kilometers, associated with filaments and mesoscale eddies, but the variability of basin-scale currents or large-scale atmospheric circulation is also of importance. Likewise, a wide range of temporal scales from the diurnal cycle to decadal times are important.

Waves. Ekman Transport and Pumping. El Ninõ Southern Oscillation (ENSO). El Ninõ Southern Oscillation (ENSO) Models. Mesoscale Eddies. Rossby Waves. Upwelling Ecosystems.

Further Reading Bakun A and Nelson CS (1991) The seasonal cycle of wind-stress curl in subtropical eastern boundary current regions. Journal of Physical Oceanography 21: 1815--1834. Chavez FP, Ryan J, Lluch-Cota SE, and Niquen M (2003) From anchovies to sardines and back: Multidecadal change in the Pacific Ocean. Science 299: 217--221. Hutchins DA, Hare CE, Weaver RC, et al. (2002) Phytoplankton iron limitation in the Humboldt Current and Peru upwelling. Limnology and Oceanography 47: 997--1011. Montecino V, Strub PT, Chavez F, Thomas AC, Tarazona J, and Baumgartner TR (2006) Chapter 10: Bio-physical interactions off Western South America. In: Robinson AR and Brink KH (eds.) The Sea, Vol. 14A: The Global Coastal Ocean – Interdisciplinary Regional Studies and Syntheses (Part 2. Regional Interdisciplinary Oceanography). Cambridge, MA: Harvard University Press. Philander SGH (1990) El Nino, La Nina and the Southern Oscillation, 289pp. San Diego, CA: Academic Press. Strub PT, Mesias JM, Montecino V, Rutllant J, and Salinas S (1998) Coastal circulation off western South America. In: Robinson AR and Brink KH (eds.) The Sea, Vol. 11: The Global Coastal Ocean – Regional Studies and Syntheses, pp. 273--313. New York: Wiley.

See also Benguela Current. California and Alaska Currents. Canary and Portugal Currents. Coastal Trapped


PHALAROPES M. Rubega, University of Connecticut, Storrs, CT, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2142–2149, & 2001, Elsevier Ltd.

Introduction Phalaropes are shore birds that have largely abandoned the shore. Rather than wading at the boundary of water and land, as most shore birds do, red (Phalaropus fulicaria) and red-necked (P. lobatus) phalaropes spend up to 9 months of the year swimming on the open ocean. (Red phalaropes are commonly called gray phalaropes in Europe.) A third species, Wilson’s phalarope (P. tricolor), does not have a marine life history (and thus will not be much considered here) but swims most of the year on saline lakes in the interior of North and South America, where its biology is similar to that of the marine phalaropes. During the breeding season, all three species frequent fresh-water wetlands. At around 20 cm long, phalaropes are among the smallest of oceanic birds. Phalaropes display un-waderlike adaptations for swimming, including lobed toes, legs that are flattened to reduce drag when paddling, and dense, waterproof feathers. Those air-trapping feathers, combined with small body size, result in the corklike buoyancy of phalaropes, as a result of which it is almost comically difficult for the birds to dive. Restricted to surface waters, phalaropes are famous among birdwatchers and biologists for frenetic spinning in small circles. This behavior generates a miniature upwelling that brings prey from deep in the water column within reach. Once grasped, prey are rapidly moved up the beak into the mouth by a process that uses the surface tension of water. The sex roles are reversed in phalaropes: larger, brightly colored females lay eggs for their mates, while smaller males with dull feathers perform all care of eggs and chicks. This unusual breeding system is further distinguished by polyandry: female phalaropes sometimes have multiple mates in a single season. Committed to an aquatic lifestyle, phalaropes do virtually everything but lay eggs on the water, nesting on land near pools or ponds. The marine phalaropes breed circumpolarly in the Arctic and sub-Arctic and migrate to pelagic wintering areas by flying out to sea. Some red-necked phalaropes also migrate overland by

a series of short flights visiting every imaginable body of water, but especially saline lakes. The population status of the two marine phalaropes is uncertain at best. The pelagic biology of these birds is poorly known, compared to other seabirds, and this hampers efforts to monitor populations and obtain data about their numbers. Breeding biology is better understood, but few breeding populations are being monitored. Large flocks at sea and on saline lakes imply that phalaropes are abundant, but human disturbance has reduced the breeding range of phalaropes. More disturbing, massive flocks have simply disappeared from former migratory staging sites, for example, the western Bay of Fundy on the north-eastern coast of Canada. We do not know whether this represents a real reduction in the population, or a shift to a new staging location caused by a change in the availability of plankton.

Phalaropes at Sea Appearance

Phalaropes are small sandpipers, with the small heads, needle-shaped beaks, and elongate necks typical of the family Scolopacidae (Figure 1). As they paddle on the ocean surface, their long wader’s legs are generally hidden from an observer. Their small size and shape make them unmistakable for any other kind of bird at sea, although they are easily mistaken for each other. Red and red-necked phalaropes are more similar to each other, and more closely related (see Systematics below), than either is to Wilson’s phalarope. These degrees of relatedness show in appearance, structure, and behavior. When seen at sea in their white and gray nonbreeding plumage (Figure 1B) red and red-necked phalaropes are difficult to distinguish. Red-necked phalaropes are slightly smaller, and more lightly built, with an exceptionally fine needlelike black bill; red phalaropes are more heavily built, with a broader, deeper, more yellow bill. Either species is easily distinguished from Wilson’s phalarope (which is the largest of the three species and lacks a dark eye patch and wing bars), which is only rarely seen in the marine environment, and never pelagically. Female phalaropes are slightly larger than males and have brighter feathers during the breeding season (Figure 1a) a condition known as reverse sexual dimorphism (in most birds males are the larger and


393

394

PHALAROPES

Figure 1 Examples of Phalaropes. Red phalarope (Phalaropus fulicarius). Other names: Grey Phalarope. (a) Adult female, breeding; (b) adult, non-breeding, both sexes; (c) adult, late summer/autumn. Length: 20 cm; wingspan: 37 cm; approximate body mass: 50 g. Range: Breeds throughout the Arctic and into the subarctic; migrates and winters throughout the North and South Atlantic Oceans, the North Pacific Ocean and the Eastern South Pacific Ocean. Illustrations from Harrison P (1985) Seabirds, an identification guide. Revised edition. Boston, Massachusetts: Houghton Mifflin.

more colorful sex). These physical differences are accompanied in phalaropes by reversed sex roles (see ‘Phalaropes on land’ below). During most of the time they are at sea, however, males and females do not differ in their plumage and cannot be distinguished (Figure 1b and c). Some researchers have suggested that measures of body size can be used to discriminate between male and female phalaropes; however, the discriminant functions developed for this purpose will only reliably identify large females. Small females overlap males in body size and cannot be distinguished from them with a high degree of certainty on the basis of body size alone. Aquatic Adaptations

All phalaropes have similar adaptations for an aquatic lifestyle that are not shared by the other sandpipers. The form of the adaptations tends to be less pronounced in Wilson’s phalarope, which is less aquatic in its habits. Phalaropes swim by paddling, and their toes are bordered by flaps or lobes of flesh (Figure 2). These toe lobes are flexible: when the foot is drawn forward through the water, the lobes fold flat behind the toe, reducing drag; on the backstroke the lobes flare out, boosting thrust. Phalarope legs are laterally flattened, another modification that reduces drag as they swing their legs through the water. As befits birds that spend their time swimming, phalaropes have waterproof feathers, with particularly

Figure 2 The lobed toes of phalaropes. These flaps of skin flare out to provide thrust when the bird is paddling through the water; they fold back around the toe on the upstroke to reduce resistance. Photo by author.

dense belly plumage. Immersion in salt water may present a constant challenge for them: red-necked phalaropes at saline lakes avidly seek freshwater sources in which to bathe and drink. It is unknown whether they seek out microhabitats with less saline water at sea for the same purpose. Retaining the hollow bones of more terrestrial birds, rather than the denser bones common in diving


PHALAROPES

seabirds, phalaropes are very buoyant and ride lightly on the surface of the water. This buoyancy, combined with the air trapped in their dense belly feathers, makes diving very difficult for them. They dive shallowly and very seldom, generally only when pursued persistently by a predator at close range. When forced to dive, they quickly pop back to the surface. Distribution

Phalaropes are pelagic during all but the short breeding season. Both red and red-necked phalaropes winter in or near the Humboldt Current, off the western coast of South America. Significant numbers of red phalaropes also winter off western Africa, and are found in the Pacific off the western United States from June through March. Red-necked phalaropes mix with reds off the western United States during migration, from July through early November. The Eurasian population of Red-necked phalaropes winters in the Arabian Sea, and from central Indonesia to western Melanesia. Food and Feeding

Diet Phalaropes are planktivores, specializing on copepods, euphausiids, and amphipods. The marine phalaropes are size-selective; copepods taken apparently do not exceed 6 mm long by 3 mm wide. They also take almost anything else that is small and floats, including other crustaceans, insects, invertebrates (including hydrozoans, molluscs, polychaetes, and gastropods), small fish, and fish eggs. In addition, phalaropes regularly ingest nonnutritious materials, including small quantities of seeds, sand, feathers, and plastic particles. Feeding behavior and mechanics Phalaropes are visual hunters. They typically swim in meandering, sinusoidal tracks, leaning forward and peering into the water. They peck at prey on, or just beneath, the water’s surface, and where prey densities are high their peck rates may climb to 180 pecks per minute. They will occasionally seize a flying insect from the air or, rarely, catch aquatic organisms by rapidly swiping the bill sideways through the water in a motion known as scything. Unlike other planktivores, phalaropes are not filter feeders; they capture zooplankton one at a time, tweezering them out of the water between the tips of their beak. When prey are successfully seized, red-necked phalaropes use the surface tension of water to move their tiny catch from the tip of their beaks to their mouths (Figure 3). They accomplish this by suspending a drop of water containing the prey between their upper and lower jaws. Since water molecules are attracted to one another, drops of water tend to

395

assume shapes with the tightest packing of molecules, i.e., shapes with the least possible surface area. Work is required to pull enough molecules out of the center of the drop to make new surface area; the amount of work required in any particular instance of water temperature and salinity is a measure of the surface tension of water. Feeding red-necked phalaropes open their jaws; this action stretches the prey-containing drop, increasing its free surface area. Once the drop is stretched, the surface tension of the water drives the drop, and the prey it contains, to the back of the jaws, where the free surface area is minimized and the prey can be swallowed. This method of transporting prey, called ‘surface tension feeding’, can be completed in as little as 0.02 s. Wilson’s phalaropes also use this method of prey transport; the feeding mechanics of red phalaropes have not been studied in detail. When prey are below the surface, a bird may submerge its head and neck or (even more rarely) upend, but a more typical response to prey deep in the water column is a conspicuous toplike spinning. All the phalaropes engage in this behavior; indeed, it is probably their best-known characteristic. Old accounts of their behavior attributed spinning variously to courtship behavior or stimulation of prey in cold water, but most authors suspected that spinning functioned to ‘stir up’ prey from the bottom of ponds and pools. This explanation was based on observations of birds on small ponds near breeding areas, and did not account for phalaropes spinning while at sea, when the bottom was hundreds of meters below. Spinning does produce subsurface water flow that concentrates and lifts prey nearer the bird, but not by creating flow against the bottom. The whirling motion is produced by kicking harder and with higher frequency with the outer leg than with the inner leg. (Observations of captive phalaropes indicate that individuals are ‘handed’ — birds spin both clockwise and counterclockwise, but each individual only spins in one direction.) Birds spin around at about one complete rotation per second. This rapid cycling kicks water at the surface away from the axis of the bird’s rotation. This deflection of surface water generates an upward-momentum jet of subsurface water; in other words, phalaropes make their own small upwellings in the center of the area they are circling by pushing surface water away so quickly that subsurface water must flow upward to replace it. Birds watch this area of upwelling for rising prey in essentially the same way that a spinning ballerina keeps from falling over. They fix their gaze on one spot, holding their heads immobile until their rotating bodies force head movement, then they snap their heads one-quarter of a turn while still looking at the same spot. Birds can generate flows to as deep


396

PHALAROPES

Figure 3 Surface tension feeding in a red-necked phalarope. Phalaropes use this feeding mechanism to transport tiny prey to the mouth, where it can be swallowed. Jaw-spreading (also called mandibular spreading) stretches the drop; the surface tension of the water drives the drop to the back of the beak, where it will have the smallest possible free surface area. (Reproduced with permission from Rubega MA and Obst BS (1993) Surface tension feeding in phalaropes: discovery of a novel feeding mechanism. Auk 110: 169–178.)

as 0.5 m, and each cycle of the spin is slightly off to the side of the previous one, so that birds slowly progress, and process water, along a track about one body length wide. As can be imagined by anyone who has ever carried water in a bucket, lifting water is an energetically expensive proposition for a phalarope: about twice as expensive, per unit of water inspected, as simply swimming in a straight line to feed. Hence, they should only spin when absolutely necessary, and they generally select pelagic habitats in which spinning is unnecessary. Habitat

The pelagic biology of phalaropes appears to revolve around food. Although the marine phalaropes spend

the majority of their life cycle at sea, they do not breed in oceanic environments. Thus, unlike other marine birds such as petrels or penguins, their use of the ocean is not influenced by factors such as access to islands with nesting sites. This freedom to concentrate on prey shows in their relationship to physical structure in the ocean. Although they occasionally take small fish, phalaropes are fundamentally planktivores, and their habitats at sea are characterized by features of the ocean that concentrate and bring to the surface dense concentrations of zooplankton. They are commonly found at fronts, thermal gradients, convergences, upwellings, and slicks. Up to two million migratory birds have been estimated at a single upwelling near Mount Desert Rock off the Maine coast. Red-necked phalaropes in


PHALAROPES

migration use a wide variety of aquatic habitats and are consistently associated within them with smallscale hydrographic features. At Mono Lake, in eastern California, red-necked and Wilson’s phalaropes concentrate their feeding activities in areas where currents help to raise prey to the surface, and at drift lines where prey are concentrated. When physical oceanography fails them, phalaropes make use of other marine organisms to locate and gain access to food. The few phalaropes found in the outer continental and slope domains of the South Atlantic Bight off the eastern coast of the United States are associated with mats of Sargassum seaweed, which are themselves the product of convergences. Red phalaropes associate with feeding whales and schools of fish that force plankton to the surface incidental to their feeding activities. The relationship of red phalaropes to whales is sufficiently dependable that whalers formerly used them as an indicator of the presence of whales. There are even reports of phalaropes picking parasites off the backs of whales. They are sometimes parasites themselves: red-necked phalaropes commonly dash in to seize prey that have been spun to the surface by the effort of another phalarope.

Phalaropes on Land All oceanic birds must make landfall to breed and reproduce, and phalaropes are no exception. The marine phalaropes are essentially on land only during the brief breeding season; red-necked phalaropes also migrate over land to some extent, and are especially numerous on saline lakes (see ‘Movements’ below). In contrast, Wilson’s phalaropes are almost entirely continental in their distribution, albeit more aquatic in their habits than most shore birds. What follows is a brief summary of the biology of phalaropes on land, and its bearing on their life at sea. Readers with an interest in the breeding biology of phalaropes will find more detail in works listed in the bibliography.

397

Breeding Behavior

The breeding behavior of phalaropes is notable for the degree to which it is conducted on water. The nest may be on land, but most significant behavioral components of breeding are carried out on the waters of small pools and ponds. Breeding displays and fights over mates usually occur on the water, as do the resulting copulations. This is in contrast to most oceanic birds, which return to land in order to engage in these behaviors. Phalaropes also differ from other oceanic birds in their unusual breeding system. First, in an arrangement that is rare for any bird, the roles of the sexes are reversed; females compete vigorously with one another for males, and provide no care for eggs or chicks, while males build the nest, brood the eggs, and care for the young when they hatch. Second, phalaropes, like most seabirds, are usually monogamous, but they differ in being polyandrous when the opportunity arises. When more males than females are present, females will breed with more than one male in a single season; female red phalaropes have laid second clutches of eggs within a few hundred meters of their first mate’s nest. Whether monogamous or polyandrous, females normally leave a male after having laid the last of the 3–4 eggs in the clutch. Chicks hatch after about 20 days of incubation, can walk and swim a few hours later, and fledge within about 20 days. Although the male tends them, they feed themselves, and may become completely independent before they are able to fly. Unlike many oceanic birds, phalaropes do not breed colonially, although pairs may breed in relatively close proximity when habitat is limited; the density of nests at any one site is highly variable from year to year. They show little tendency to return to any particular breeding site, or to the site where they were hatched. Red phalaropes will exploit the aggressive nature of colonial birds to ward off predators, by nesting in the colonies of other seabirds such as Arctic terns (Sterna paradisaea).

Distribution

Appearance Each species of phalarope has an extremely colorful, distinctive appearance during the breeding season, with striking patterns of black, gray, and red or rusty markings on the neck and face (Figure 1a). They are easily distinguished from one another and from other shore birds. All phalaropes exhibit reverse sexual dimorphism; females are more sharply and brightly colored than males when breeding. On average females are also larger.

The marine phalaropes breed in tundra habitats circumpolarly in Arctic and sub-Arctic, along coasts of the Arctic Ocean. The red-necked phalarope breeds farther south, and further into the interior of Eurasia, Alaska, and northern Canada than does the red phalarope. Red-necked phalaropes also breed in small numbers in the Aleutian Islands, Scotland, and Ireland. The nonpelagic Wilson’s phalarope, in contrast, breeds only in the Nearctic, primarily in the interior of western North America.


398

PHALAROPES

Food and Feeding

Diet During the breeding season phalaropes essentially remain planktivores, eating small aquatic prey, especially the larvae of dipteran flies. However, on the breeding grounds their diet expands to contain adult dipteran flies (which they snap out of the air), mosquito larvae, dragonfly nymphs, water boatmen, backswimmers, caddisflies, beetles, bugs, ants, spiders, mites, snails, crustaceans, molluscs, and annelid worms. When food is limited they may eat seeds and other plant materials. The diets of red-necked phalaropes at saline lakes consist almost entirely of brine flies, with third-instar larvae predominating, plus adult and larval dipterans. Brine shrimp are very abundant at saline lakes but are rarely taken by red-necked phalaropes. This lack of interest is the product of a nutrient limitation; captive birds are reluctant to eat brine shrimp, and those restricted to a brine shrimp diet lost about 5% of their body weight while eating three times their body weight over a 12-hour period. In contrast, Wilson’s phalaropes do eat significant amounts of brine shrimp at saline lakes; their ability to extract nutrition from them has not been investigated in detail.

Movements All phalaropes are migratory; they differ chiefly in the degree to which they move over land versus sea. Red phalaropes virtually always migrate pelagically; red-necked phalaropes migrate over both land and ocean; Wilson’s phalaropes are thought to migrate southward over the Pacific after leaving North America, apparently without ever landing on the water, while the north-bound migration occurs almost entirely over land. In all three species the timing of movements is similar: nonbreeding birds of both sexes leave the breeding grounds first, followed consecutively by females, males, and juveniles. The migration of red phalaropes is perhaps least well understood, since it is least easily observed. What is known about their migratory routes is inferred as much from information about where they are not seen as from sightings of them on the move. Nearctic breeders winter off western and southwestern Africa. Until recently large flocks occurred in the western Bay of Fundy during migration (see Conservation and Threats), but few have been seen farther south near the western Atlantic coasts. Thus, they presumably fly directly across the Atlantic to their destinations off the African coast after leaving the Bay of Fundy. Many red phalaropes winter in the Humboldt Current, off western South America, and

large flocks are present during migration off the Pacific coast of North America. Sightings of red phalaropes in the central Pacific Ocean during the migratory period indicate that those breeding in the Siberian Arctic cross the Pacific to winter in the Humboldt current. Red-necked phalaropes mix travel over land and sea to a much greater extent. Those breeding in Europe and western Siberia move through the Caspian Sea and overland via lakes across the former Soviet Union and Iran to arrive at their Arabian Sea wintering grounds through the Gulf of Oman. Birds that have bred in Fenno-Scandinavia move southeast through the gulfs of Bothnia and Finland. Breeding populations from eastern Siberia migrate overland and offshore of Japan to winter in the East Indies. Nearctic populations move south across Canada and the western United States, where tens of thousands stop at hypersaline lakes, along with smaller numbers at every conceivable body of water in their southward path; from these lakes they move out to sea and south to their wintering grounds at the northern edge of the Humboldt Current. Huge flocks estimated at 2 000 000 individuals occurred in the Bay of Fundy until recently (see Conservation and Threats), and these may include birds from Greenland, Iceland, and the Nearctic. Where these birds winter is a mystery; they do not winter off western Africa with the red phalaropes with which they mingle in the Bay of Fundy, and only small flocks of red-necked phalaropes have been seen farther south in the western Atlantic in the winter. It has been suggested that they may fly to the Humboldt Current via routes crossing the Caribbean and Central America, but no more than a few red-necked phalaropes have been seen in these areas.

Systematics The names and scientific classifications of phalaropes have histories nearly as colorful as the birds themselves (in South America, red-necked phalaropes are called ‘pollito del mar,’ roughly ‘little chicken of the sea’). Both common and scientific names have changed repeatedly (Table 1). Each species was once considered to form a distinct genus, and was given a separate Latin first name (Phalaropus, Lobipes, and Steganopus for red, red-necked, and Wilson’s phalarope, respectively). Subsequently, and for many years, the phalaropes have been placed in a single group in the family Scolopacidae (sandpipers), subfamily Phalaropodinae, genus Phalaropus. The idea that all the phalaropes descended from a single common ancestor is sometimes disputed.


PHALAROPES

Table 1

399

The diversity of names for phalaropes Red phalarope

Red-necked phalarope

Other common names

Grey phalarope Red coot-footed tringa

Present scientific name Previous scientific names

Phalaropus fulicaria Tringa fulicaria

Northern phalarope Hyperborean phalarope Cock coot-footed tringa Phalaropus lobatus Tringa tobata/lobata Lobipes lobatus

Synonymous scientific names

Phalaropus glacialis, rufus, platyrhynchus, rufescens, griseus, cinereus

Analyses of both their genetic material and skeletons showed that red and red-necked phalaropes are each other’s closest relatives. Analyses of skeletal materials united the more distantly related Wilson’s phalarope to the other two in a single group, but genetic analyses do not unambiguously support the idea that all three phalaropes belong to a single group. Both kinds of evidence indicate that phalaropes are either closely related to the Tringine or Scolopacine sandpipers. Thus, their scientific classification has recently depended on the authority consulted. For example, Sibley and Monroe’s 1993 taxonomy of birds, based on DNA–DNA hybridization studies, put the phalaropes in the subfamily Tringinae, with red and red-necked phalaropes in the genus Phalaropus, and Wilson’s phalarope returned to the genus Steganopus. The American Ornithologists’ Union continues to classify them in a single genus Phalaropus, in the subfamily Phalaropidinae.

Conservation and Threats

Lobipes hyperboreus, anguirostris, antarcticus, fuscus, ruficollis, tropicus, vulgaris

Wilson’s phalarope

Phalaropus tricolor Steganopus tricolor some authorities have resurrected this name for Wilson’s phalarope Steganopus wilsoni, incanus, frenatus, stenodactylus

longer breed anywhere in Britain apart from in Scotland and Ireland (and there only irregularly) because of egg collecting in the nineteenth century. Breeding populations elsewhere are largely unstudied and, where they are monitored, information about population trends is equivocal. The population of male red-necked phalaropes at LaPerouse Bay, in Churchill, Manitoba declined by 94% between 1980 and 1993, but nesting densities have increased since 1981 near Prudhoe Bay, Alaska. At sea, apparent declines are even more difficult to understand. The number of phalaropes staging for fall migration in the western Bay of Fundy declined from estimates of two million to almost nothing in the mid-1990s. This disappearance may represent a true population decline, or simply a shift of currents and prey, and thus of birds, to some as-yet undiscovered area of the Bay or the western Atlantic. Similar declines have been reported in the number of phalaropes seen off coastal Japan in spring. Limited evidence suggests that the numbers of birds passing through the Bay of Fundy during spring migration is unchanged.

Population Status

Phalarope populations are not thought to be threatened on a global scale, but their status is poorly known. Thus, significant population declines would be difficult to document with any degree of certainty. This lack of information arises from the peculiar life history of phalaropes; most seabirds are counted at their breeding colonies, while most shore birds are counted at migratory staging areas or wintering grounds because they do not nest in colonies. Phalaropes do not nest colonially, and in the nonbreeding season gather far out at sea, where it is difficult to find and count them. Local declines of breeding birds have been documented; for instance, red-necked phalaropes no

Threats

Compared to many oceanic birds, phalaropes probably face relatively few threats. Their breeding populations are widely distributed and thus, unlike those of many colonially nesting seabirds breeding on islands, are resistant to depredations of introduced predators. Their predators on the breeding grounds include raptorial birds such as pomarine and parasitic jaegers, mammals such as arctic and red foxes and short-tailed weasels, and chick and egg predators such as glaucous gulls, sandhill cranes, and arctic ground squirrels. They are safe from most of these when at sea. However, they are not invulnerable even at sea: four red-necked phalaropes were once found in the stomach of a common dolphin taken off Baja California, Mexico.


400

PHALAROPES

With the exception of minor subsistence hunting by indigenous northerners, phalaropes are not hunted by humans and as surface-swimming planktivores, are not incidentally taken in fishing nets, as so many seabirds are. They are potentially vulnerable to spilled oil, particularly since oil and food particles may be concentrated at the same convergence zone. As for all oceanic organisms, human-caused disruption and destruction of marine environments is likely the most serious threat facing phalarope populations.

See also Baleen Whales. Copepods. Plankton and Climate. Seabird Foraging Ecology. Seabird Migration. Seabird Population Dynamics. Seabirds and Fisheries Interactions. Seabirds as Indicators of Ocean Pollution. Upwelling Ecosystems.

Further Reading

The Birds of North America, No. 83. Philadelphia: The Academy of Natural Sciences; Washington, D.C: The American Ornithologists’ Union. Cramp S and Simmons KEL (eds.) (1983) Handbook of the Birds of Europe, the Middle East and North Africa: The Birds of the Western Palearctic, vol. 3. Oxford: Oxford University Press. Del Hoyo J, Elliott A, and Sargatal J (eds.) (1992) Handbook of the Birds of the World, vol. 1. Barcelona: Lynx Edicions. Ho¨hn EO (1971) Observations on the breeding behaviour of grey and red-necked phalaropes. Ibis 113: 335--348. Murphy RC (ed.) (1936) Oceanic Birds of South America, vol. II. New York: American Museum of Natural History. Rubega MA, Schamel DS, and Tracy D (2000) Red-necked phalarope (Phalaropus lobatus). In: Poole A and Gill F (eds.) The Birds of North America, No. 538. Philadelphia: The Academy of Natural Sciences; Washington, D.C.: The American Ornithologists’ Union. Tinbergen N (1935) Field observations of East Greenland birds. I. The behavior of the red-necked phalarope (Phalaropus lobatus L.) in spring. Ardea 26: 1--42.

Colwell MA and Jehl JR, Jr (1994) Wilson’s phalarope (Phalaropus tricolor). In: Poole A and Gill F (eds.)


PHOSPHORUS CYCLE K. C. Ruttenberg, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2149–2162, & 2001, Elsevier Ltd.

Introduction The global phosphorus cycle has four major components: (i) tectonic uplift and exposure of phosphorus-bearing rocks to the forces of weathering; (ii) physical erosion and chemical weathering of rocks producing soils and providing dissolved and particulate phosphorus to rivers; (iii) riverine transport of phosphorus to lakes and the ocean; and (iv) sedimentation of phosphorus associated with organic and mineral matter and burial in sediments (Figure 1). The cycle begins anew with uplift of sediments into the weathering regime. Phosphorus is an essential nutrient for all life forms. It is a key player in fundamental biochemical reactions involving genetic material (DNA, RNA) and energy transfer (adenosine triphosphate, ATP), and in structural support of organisms provided by membranes (phospholipids) and bone (the biomineral hydroxyapatite). Photosynthetic organisms utilize dissolved phosphorus, carbon, and other essential nutrients to build their tissues using energy from the sun. Biological productivity is contingent upon the availability of phosphorus to these organisms, which constitute the base of the food chain in both terrestrial and aquatic systems. Phosphorus locked up in bedrock, soils, and sediments is not directly available to organisms. Conversion of unavailable forms to dissolved orthophosphate, which can be directly assimilated, occurs through geochemical and biochemical reactions at various stages in the global phosphorus cycle. Production of biomass fueled by phosphorus bioavailability results in the deposition of organic matter in soil and sediments, where it acts as a source of fuel and nutrients to microbial communities. Microbial activity in soils and sediments, in turn, strongly influences the concentration and chemical form of phosphorus incorporated into the geological record. This article begins with a brief overview of the various components of the global phosphorus cycle. Estimates of the mass of important phosphorus reservoirs, transport rates (fluxes) between reservoirs,

and residence times are given in Tables 1 and 2. As is clear from the large uncertainties associated with these estimates of reservoir size and flux, there remain many aspects of the global phosphorus cycle that are poorly understood. The second half of the article describes current efforts underway to advance our understanding of the global phosphorus cycle. These include (i) the use of phosphate oxygen isotopes (d18O-PO4) as a tool for identifying the role of microbes in the transformer of phosphate from one reservoir to another; (ii) the use of naturally occurring cosmogenic isotopes of phosphorus (32P and 33 P) to provide insight into phosphorus-cycling pathways in the surface ocean; (iii) critical evaluation of the potential role of phosphate limitation in coastal and open ocean ecosystems; (iv) reevaluation of the oceanic residence time of phosphorus; and (v) rethinking the global phosphorus-cycle on geological timescales, with implications for atmospheric oxygen and phosphorus limitation primary productivity in the ocean.

The Global Phosphorus Cycle: Overview The Terrestrial Phosphorus Cycle

In terrestrial systems, phosphorus resides in three pools: bedrock, soil, and living organisms (biomass) (Table 1). Weathering of continental bedrock is the principal source of phosphorus to the soils that support continental vegetation (F12); atmospheric deposition is relatively unimportant (F82). Phosphorus is weathered from bedrock by dissolution of phosphorus-bearing minerals such as apatite (Ca10(PO4)6(OH, F, Cl)2), the most abundant primary phosphorus mineral in crustal rocks. Weathering reactions are driven by exposure of minerals to naturally occurring acids derived mainly from microbial activity. Phosphate solubilized during weathering is available for uptake by terrestrial plants, and is returned to the soil by decay of litterfall (Figure 1). Soil solution phosphate concentrations are maintained at low levels as a result of absorption of phosphorus by various soil constituents, particularly ferric iron and aluminum oxyhydroxides. Sorption is considered the most important process controlling terrestrial phosphorus bioavailability. Plants have different physiological strategies for obtaining phosphorus despite low soil solution concentrations.


401

W

m Che

ica

l

the

ring

• lithified sediments • crystalline rocks

BEDROCK • deep soil (> 60 cm) • marine sediments

ea

R IVER

Crops

Mining/ Fertilizer-P

hν

TERRESTRIAL

Coastal Upwelling

Algae

Pauth

Porg PFe Pdetr

MARINE

Pauth

Porg PFe

PHOTIC ZONE

ZONE

APHOTIC

ABYSSAL SEDIMENTS

Benthic Return P-flux

Particulate-P

Dissolved-P

Decomposition

Upwelling

hν

Algae

Dust

Particle Settling

ng

li Sedimentation

Benthic Return P-flux

Photosynthetic P-uptake Sedimentation Downwelling Benthic Return P-flux

hν

SLOPE SEDIMENTS

SHELF SEDIMENTS

Chemical Weathering: (dissolved-P)

SOIL U Biomass PNO Decompo- uptake FF sition Groundwater l R ecyc Erosion: (biomass, soil particles)

Litter fall

Biomass Storage

g

THE GLOBAL PHOSPHORUS CYCLE

R R ecyc

(c) 2011 Elsevier Inc. All Rights Reserved. 0

2

_

3

4

Pacific

PO4 (µmol l 1)

1

Atlantic

Indian

Characteristic deep-sea dissolved phosphate profiles for three ocean basins

4000

3000

2000

1000 m

Figure 1 Cartoon illustrating the major reservoirs and fluxes of phosphorus described in the text and summarized in Tables Table 1 and 2. The oceanic photic zone, idealized in the cartoon, is typically thinner in coastal environments owing to turbidity from continental terrigenous input, and deepens as the water column clarifies with distance away from the continental margins. The distribution of phosphorus among different chemical/mineral forms in marine sediments is given in the pie diagrams, where the abbreviations used are: Porg, organic phosphorus; PFe, iron-bound phosphorus; Pdetr, detrital apatite; Pauth, authigenic/biogenic apatite. The Porg, PFe, and Pauth reservoirs represent potentially reactive phosphorus pools (see text and Tables 2 and 5 for discussion), whereas the Pdetr pool reflects mainly detrital apatite weathered off the continents and passively deposited in marine sediments (note that Pdetr is not an important sedimentary phosphorus component in abyssal sediments, far from continents).Continental margin phosphorus speciation data were compiled from Louchouarn P, Lucotte M, Duchemin E and de Vernal A (1997) Early diagenetic processes in recent sediments of the Gulf of St-Lawrence: Phosphorus, carbon and iron burial rates. Marine Geology 139(1/4): 181–200, and Ruttenberg KC and Berner RA (1993) Authigenic apatite formation and burial in sediments from non-upwelling continental margin environments. Geochimica et Cosmochimica Acta 57: 991–1007. Abyssal sediment phosphorus speciation data were compiled from Filippelli GM and Delaney ML (1996) Phosphorus geochemistry of equatorial Pacific sediments. Geochimica et Cosmochimica Acta 60: 1479}1495, and Ruttenberg KC (1990) Diagenesis and burial of phosphorus in marine sediments: implications for the marine phosphorus budget. PhD thesis, Yale University. The global phosphorus cycle cartoon is from Ruttenberg (2000). The global phosphorus cycle. In: The Encyclopedia of Global Change, Oxford University Press. (in Press), with permission. The vertical water column distributions of phosphate typically observed in the three ocean basins are shown in the panel to the right of the global phosphorus cycle cartoon, and are from Sverdrup HV, Johnson MW and Fleming RH (1942) The Oceans, Their Physics, Chemistry and General Biology. New York: Prentice Hall, ^ 1942 Prentice Hall; used with permission.

U P L I F T

Ph E

l ica on ys ro s i

in

402 PHOSPHORUS CYCLE

PHOSPHORUS CYCLE

Table 1

403

Major reservoirs active in the global phosphorus cycle and associated residence times

For example, some plants can increase root volume and surface area to optimize uptake potential. Alternatively, plant roots and/or associated fungi can produce chelating compounds that solubilize ferric iron and calcium-bound phosphorus, enzymes and/or acids that solubilize phosphate in the root vicinity. Plants also minimize phosphorus loss by resorbing much of their phosphorus prior to litterfall, and by efficient recycling from fallen litter. In extremely unfertile soils (e.g., in tropical rain forests) phosphorus recycling is so efficient that topsoil contains virtually no phosphorus; it is all tied up in biomass. Systematic changes in the total amount and chemical form of phosphorus occur during soil

development. In initial stages, phosphorus is present mainly as primary minerals such as apatite. In midstage soils, the reservoir of primary apatite is diminished; less-soluble secondary minerals and organic phosphorus make up an increasing fraction of soil phosphorus. Late in soil development, phosphorus is partitioned mainly between refractory minerals and organic phosphorus (Figure 2). Transport of Phosphorus from Continents to the Ocean

Phosphorus is transferred from the continental to the oceanic reservoir primarily by rivers (F24).


404

PHOSPHORUS CYCLE

Table 2

Fluxes between the major phosphorus reservoirsa


PHOSPHORUS CYCLE


405

Weight of P/unit area of soil

406

PHOSPHORUS CYCLE

Pmineral Ptotal

Pnonoccluded

Poccluded

Porganic Time

Figure 2 The fate of phosphorus during soil formation can be viewed as the progressive dissolution of primary mineral P (dominantly apatite), some of which is lost from the system by leaching (decrease in Ptotal), and some of which is reincorporated into nonoccluded, occluded and organic fractions within the soil. Nonoccluded P is defined as phosphate sorbed to surface of hydrous oxides of iron and aluminum, and calcium carbonate. Occluded P refers to P present within the mineral matrix of discrete mineral phases. The initial build-up in organic P results from organic matter return to soil from vegetation supported by the soil. The subsequent decline in Porganic is due to progressive mineralization and soil leaching. The time-scale over which these transformations occur depends upon the initial soil composition, topographic, and climatic factors. Figure is after Walker and Syers (1976). The fate of phosphorus during pedogenesis. Geoderma 15: 1–19, with permission.

Deposition of atmospheric aerosols (F84) is a minor flux. Groundwater seepage to the coastal ocean is a potentially important but undocumented flux. Riverine phosphorus derives from weathered continental rocks and soils. Because phosphorus is particle-reactive, most riverine phosphorus is associated with particulate matter. By most estimates, over 90% of the phosphorus delivered by rivers to the ocean is as particulate phosphorus (F24(p)). Dissolved phosphorus in rivers occurs in both inorganic and organic forms. The scant data on dissolved organic phosphorus suggest that it may account for 50% or more of dissolved riverine phosphorus. The chemical form of phosphorus associated with riverine particles is variable and depends upon the drainage basin geology, on the extent of weathering of the substrate, and on the nature of the river itself. Available data suggest that approximately 20–40% of phosphorus in suspended particulate matter is organic. Inorganic forms are partitioned mainly between ferric oxyhydroxides and apatite. Aluminum oxyhydroxides and clays may also be significant carriers of phosphorus.

The fate of phosphorus entering the ocean via rivers is variable. Dissolved phosphorus in estuaries at the continent–ocean interface typically displays nonconservative behavior. Both negative and positive deviations from conservative mixing can occur, sometimes changing seasonally within the same estuary. Net removal of phosphorus in estuaries is typically driven by flocculation of humic-iron complexes and biological uptake. Net phosphorus release is due to a combination of desorption from fresh water particles entering high-ionic-strength marine waters, and flux of diagenetically mobilized phosphorus from benthic sediments. Accurate estimates of bioavailable riverine phosphorus flux to the ocean must take into account, in addition to dissolved forms, the fraction of riverine particulate phosphorus released to solution upon entering the ocean. Human impacts on the global phosphorus cycle The mining of phosphate rock (mostly from marine phosphorite deposits) for use as agricultural fertilizer (F72) increased dramatically in the latter half of the twentieth century. In addition to fertilizer use, deforestation, increased cultivation, and urban and industrial waste disposal all have enhanced phosphorus transport from terrestrial to aquatic systems, often with deleterious results. For example, elevated phosphorus concentrations in rivers resulting from these activities have resulted in eutrophication in some lakes and coastal areas, stimulating nuisance algal blooms and promoting hypoxic or anoxic conditions that are harmful or lethal to natural populations. Increased erosion due to forest clear-cutting and widespread cultivation has increased riverine suspended matter concentrations, and thus increased the riverine particulate phosphorus flux. Dams, in contrast, decrease sediment loads in rivers and therefore diminish phosphorus-flux to the oceans. However, increased erosion below dams and diagenetic mobilization of phosphorus in sediments trapped behind dams moderates this effect. The overall effect has been a 50–300% increase in riverine phosphorus flux to the oceans above pre-agricultural levels. The Marine Phosphorus Cycle

Phosphorus in its simplest form, dissolved orthophosphate, is taken up by photosynthetic organisms at the base of the marine food web. When phosphate is exhausted, organisms may utilize more complex forms by converting them to orthophosphate via enzymatic and microbiological reactions. In the open ocean most phosphorus associated with biogenic particles is recycled within the upper water column.


PHOSPHORUS CYCLE

Efficient stripping of phosphate from surface waters by photosynthesis combined with build-up at depth due to respiration of biogenic particles results in the classic oceanic dissolved nutrient profile. The progressive accumulation of respiration-derived phosphate at depth along the deep-water circulation trajectory results in higher phosphate concentrations in Pacific Ocean deep waters at the end of the trajectory than in the North Atlantic where deep water originates (Figure 1). The sole means of phosphorus removal from the oceans is burial with marine sediments. The phosphorus flux to shelf and slope sediments is larger than the phosphorus flux to the deep sea (Table 2) for several reasons. Coastal waters receive continentally derived nutrients via rivers (including phosphorus, nitrogen, silicon, and iron), which stimulate high rates of primary productivity relative to the deep sea and result in a higher flux of organic matter to continental margin sediments. Organic matter is an important, perhaps primary, carrier of phosphorus to marine sediments. Owing to the shorter water column in coastal waters, less respiration occurs prior to deposition. The larger flux of marine organic phosphorus to margin sediments is accompanied by a larger direct terrigenous flux of particulate phosphorus (organic and inorganic), and higher sedimentation rates overall. These factors combine to enhance retention of sedimentary phosphorus. During high sea level stands, the sedimentary phosphorus reservoir on continental margins expands, increasing the phosphorus removal flux and therefore shortening the oceanic phosphorus residence time. Terrigenous-dominated shelf and slope (hemipelagic) sediments and abyssal (pelagic) sediments have distinct phosphorus distributions. While both are dominated by authigenic Ca-P (mostly carbonate fluorapatite), this reservoir is more important in pelagic sediments. The remaining phosphorus in hemipelagic sediments is partitioned between ferric ironbound phosphorus (mostly oxyhydroxides), detrital apatite, and organic phosphorus; in pelagic sediments detrital apatite is unimportant. Certain coastal environments characterized by extremely high, upwelling-driven biological productivity and low terrigenous input are enriched in authigenic apatite; these are proto-phosphorite deposits. A unique process contributing to the pelagic sedimentary Fe-P reservoir is sorptive removal of phosphate onto ferric oxyhydroxides in mid-ocean ridge hydrothermal systems. Mobilization of sedimentary phosphorus by microbial activity during diagenesis causes dissolved phosphate build-up in sediment pore waters, promoting benthic efflux of phosphate to bottom waters

407

or incorporation in secondary authigenic minerals. The combined benthic flux from coastal (sFcbf) and abyssal (sFabf) sediments is estimated to exceed the total riverine phosphorus flux (F24(dþp)) to the ocean. Reprecipitation of diagenetically mobilized phosphorus in secondary phases significantly enhances phosphorus burial efficiency, impeding return of phosphate to the water column. Both processes impact the marine phosphorus cycle by affecting the primary productivity potential of surface waters.

Topics of Special Interest in Marine Phosphorus Research Phosphate Oxygen Isotopes (d18O-PO4)

Use of the oxygen isotopic composition of phosphate in biogenic hydroxyapatite (bones, teeth) as a paleotemperature and climate indicator was pioneered by Longinelli in the late 1960s–early 1970s, and has since been fairly widely and successfully applied. A novel application of the oxygen isotope system in phosphates is its use as a tracer of biological turnover of phosphorus during metabolic processes. Phosphorus has only one stable isotope (31P) and occurs almost exclusively as orthophosphate ðPO43 Þ under Earth surface conditions. The phosphorus–oxygen bond in phosphate is highly resistant to nonenzymatic oxygen isotope exchange reactions, but when phosphate is metabolized by living organisms, oxygen isotopic exchange is rapid and extensive. Such exchange results in temperature-dependent fractionations between phosphate and ambient water. This property renders phosphate oxygen isotopes useful as indicators of present or past metabolic activity of organisms, and allows distinction of biotic from abiotic processes operating in the cycling of phosphorus through the environment. Currently, the d18O-PO4 system is being applied in a number of studies of marine phosphorus cycling, including (i) application to dissolved sea water inorganic phosphate as a tracer of phosphate source, water mass mixing, and biological productivity; (ii) use in phosphates associated with ferric iron oxyhydroxide precipitates in submarine ocean ridge sediments, where the d18O-PO4 indicates microbial phosphate turnover at elevated temperatures. This latter observation suggests that phosphate oxygen isotopes may be useful biomarkers for fossil hydrothermal vent systems. Reevaluating the Role of Phosphorus as a Limiting Nutrient in the Ocean

In terrestrial soils and in the euphotic zone of lakes and the ocean, the concentration of dissolved


PHOSPHORUS CYCLE

orthophosphate is typically low. When bioavailable phosphorus is exhausted prior to more abundant nutrients, it limits the amount of sustainable biological productivity. Phosphorus limitation in lakes is widely accepted, and terrestrial soils are often phosphorus-limited. In the oceans, however, phosphorus limitation is the subject of controversy and debate. The prevailing wisdom has favored nitrogen as the limiting macronutrient in the oceans. However, a growing body of literature convincingly demonstrates that phosphate limitation of marine primary productivity can and does occur in some marine systems. In the oligotrophic gyres of both the western North Atlantic and subtropical North Pacific, evidence in the form of dissolved nitrogen : phosphorous (N:P) ratios has been used to argue convincingly that these systems are currently phosphate-limited. The N(:)P ratio of phytoplankton during nutrientsufficient conditions is 16N:1P (the Redfield ratio) (see Redfield Ratio). A positive deviation from this ratio indicates probable phosphate limitation, while a negative deviation indicates probable nitrogen limitation. In the North Pacific at the Hawaiian Ocean Time Series (HOT) site, there has been a shift since the 1988 inception of the time series to N:P ratios exceeding the Redfield ratio in both particulate and surface ocean dissolved nitrogen and phosphorus (Figure 3). Coincident with this shift has been an increase in the prevalence of the nitrogen-fixing cyanobacterium Trichodesmium (Table 3). Currently, it appears as though the supply of new nitrogen has shifted from a limiting flux of upwelled nitrate from below the euphotic zone to an unlimited pool of atmospheric N2 rendered bioavailable by the action of nitrogen fixers. This shift is believed to result from climatic changes that promote water column stratification, a condition that selects for N2fixing microorganisms, thus driving the system to phosphate limitation. A similar situation exists in the subtropical Sargasso Sea at the Bermuda Ocean Time Series (BATS) site, where currently the dissolved phosphorus concentrations (especially dissolved inorganic phosphorus (DIP)) are significantly lower than at the HOT site, indicating even more severe phosphate limitation (Table 3). A number of coastal systems also display evidence of phosphate limitation, sometimes shifting seasonally from nitrogen to phosphate limitation in concert with changes in environmental features such as upwelling and river runoff. On the Louisiana Shelf in the Gulf of Mexico, the Eel River Shelf of northern California (USA), the upper Chesapeake Bay (USA), and portions of the Baltic Sea, surface water column dissolved inorganic N:P ratios indicate seasonal

24

Total dissolved 22

20 18 16

(A) 14

N:P ratio (mol:mol)

408

(B)

Suspended particles

30

20

10

0

Exported particles

40

20

(C)

0 J A J O J A J O J A J O J A J OJ A J OJ A J OJ A J OJ A J OJ A J O 1989

1990

1991 1992 1993 1994 1995 1996 1997 Sampling date

Figure 3 Time-series molar N:P ratios in (A) the dissolved pool, (B) suspended particulate matter, and (C) exported particulate matter from the HOT time series site at station ALOHA in the subtropical North Pacific near Hawaii. (A) The 3-point running mean N:P ratios for 0–100 m (circles), 100–200 m (squares), and 200–500 m (triangles). (B) The 3-point running mean (71 SD) for the average suspended particulate matter in the upper water column (0–100 m). (C) The 3-point running mean (71 SD) for the average N:P ratio of sediment trap-collected particulate matter at 150 m. The Redfield ratio (N:P16) is represented by a dashed line in all three panels. Particulate and upper water column dissolved pools show an increasing N:P ratio throughout the time-series, with a preponderance of values in excess of the Redfield ratio. (After Karl DM, Letelier R, Tupas L et al. (1997) The role of nitrogen fixation in the biogeochemical cycling in the subtropical North Pacific Ocean. Nature 388: 533–538, with permission.)

phosphate limitation. The suggestion of phosphate limitation is reinforced in the Louisiana Shelf and Eel River Shelf studies by the occurrence or presence of alkaline phosphatase activity, an enzyme induced only under phosphate-limiting conditions. Alkaline phosphatase has also been observed seasonally in Narragansett Bay. Although these coastal sites are recipients of anthropogenically derived nutrients (nitrogen and phosphate) that stimulate primary productivity above ‘natural’ levels, the processes that result in shifts in the limiting nutrient are not


PHOSPHORUS CYCLE

Table 3 Parameters affecting nutrient limitation: comparison between North Atlantic and North Pacific Gyres

409

weathered off the continents and delivered by rivers, with some minor atmospheric input. As a consequence of continental weathering control on phosphorus supply to the oceans, phosphorus limitation has been considered more likely than nitrogen limitation on geological timescales. As the recent studies reviewed in this section suggest, phosphorus limitation may be an important phenomenon in the modern ocean as well. Overall, current literature indicates that there is a growing appreciation of the complexities of nutrient limitation in general, and the role of phosphorus limitation in particular, in both fresh water and marine systems. Cosmogenic 32P and 33P as Tracers of Phosphorus Cycling in Surface Waters

necessarily related to anthropogenic effects. Other oceanic sites where phosphorus limitation of primary productivity has been documented include the Mediterranean Sea and Florida Bay (USA). One key question that studies of nitrogen and phosphorus limitation must address before meaningful conclusions may be drawn about phosphorus versus nitrogen limitation of marine primary productivity is the extent to which the dissolved organic nutrient pools are accessible to phytoplankton. In brief, this is the question of bioavailability. Many studies of nutrient cycling and nutrient limitation do not include measurement of these quantitatively important nutrient pools, even though there is indisputable evidence that at least some portion of the dissolved organic nitrogen (DON) and dissolved organic phosphorus (DOP) pools is bioavailable. An important direction for future research is to characterize the DON and DOP pools at the molecular level, and to evaluate what fraction of these are bioavailable. The analytical challenge of identifying the molecular composition of the DOP pool is significant. Recent advances in 31P-nuclear magnetic resonance (NMR) spectroscopy have permitted a look at the high molecular weight (41 nm) fraction of the DOP pool. However, this fraction represents only one-third of the total DOP pool; the other twothirds of this pool, made up of smaller-molecularweight DOP compounds, remains outside our current window of analytical accessibility. Debate continues among oceanographers about the most probable limiting nutrient on recent and on long, geological timescales. While there is an abundant reservoir of nitrogen (gaseous N2) in the atmosphere that can be rendered bioavailable by nitrogen-fixing photosynthetic organisms, phosphorus supply to the ocean is limited to that

There are two radioactive isotopes of phosphorus, 32P (half-life ¼ 14.3 days) and 33P (half-life ¼ 25.3 days). Both have been widely used in the study of biologically mediated phosphorus cycling in aquatic systems. Until very recently, these experiments have been conducted by artificially introducing radiophosphorus into laboratory incubations or, far more rarely, by direct introduction into natural waters under controlled circumstances. Such experiments necessarily involve significant perturbation of the system, which can complicate interpretation of results. Recent advances in phosphorus sampling and radioisotope measurement have made it possible to use naturally produced 32P and 33P as in situ tracers of phosphorus recycling in surface waters. This advance has permitted studies of net phosphorus recycling in the absence of experimental perturbation caused by addition of artificially introduced radiophosphorus. 32 P and 33P are produced naturally in the atmosphere by interaction of cosmic rays with atmospheric argon nuclei. They are then quickly scavenged onto aerosol particles and delivered to the ocean surface predominantly in rain. The ratio of 33P/32P introduced to the oceans by rainfall remains relatively constant, despite the fact that absolute concentrations can vary from one precipitation event to another. Once the dissolved phosphorus is incorporated into a given surface water phosphorus pool (e.g., by uptake by phytoplankton or bacteria, grazing of phytoplankton or bacteria by zooplankton, or abiotic sorption), the 33P/32P ratio will increase in a systematic way as a given pool ages. This increase in the 33 32 P/ P ratio with time results from the different halflives of the two phosphorus radioisotopes. By measuring the 33P/32P ratio in rain and in different marine phosphorus pools – e.g., DIP, DOP (sometimes called soluble nonreactive phosphorus, or SNP), and particulate phosphorus of various size classes


410

PHOSPHORUS CYCLE

corresponding to different levels in the food chain – the net age of phosphorus in any of these reservoirs can be determined (Table 4). New insights into phosphorus cycling in oceanic surface waters derived from recent work using the cosmogenically produced 33 32 P/ P ratio include the following. (1) Turnover rates of dissolved inorganic phosphorus in coastal and oligotrophic oceanic surface waters ranges from 1 to 20 days. (2) Variable turnover rates in the DOP pool range from o1 week to >100 days, suggesting differences in either the demand for DOP, or the lability of DOP toward enzymatic breakdown. (3) In the Gulf of Maine, DOP turnover times vary seasonally, increasing from 28 days in July to >100 days in August, suggesting that the DOP pool may evolve compositionally during the growing seasons. (4) Comparison of the 33P/32P ratio in different particulate size classes indicates that the age of phosphorus generally increases at successive levels in the food chain. (5) Under some circumstances, the 33P/32P ratio can reveal which dissolved pool is being ingested by a particular size class of organisms. Utilization of this new tool highlights the dynamic nature of phosphorus cycling in surface waters by revealing the rapid rates and temporal variability of phosphorus turnover. It further stands to provide new insights into ecosystem nutrient dynamics by revealing, for example, that low phosphorus concentrations can support high primary productivity through rapid turnover rates, and that there is preferential utilization of particular dissolved phosphorus pools by certain classes of organisms.

The Oceanic Residence Time of Phosphorus

As phosphorus is the most likely limiting nutrient on geological timescales, an accurate determination of its oceanic residence time is crucial to understanding how levels of primary productivity may have varied in the Earth’s past. Residence time provides a means of evaluating how rapidly the oceanic phosphorus inventory may have changed in response to variations in either input (e.g., continental weathering, dust flux) or output (e.g., burial with sediments). In its role as limiting nutrient, phosphorus will dictate the amount of surface-ocean net primary productivity, and hence atmospheric CO2 drawdown that will occur by photosynthetic biomass production. It has been suggested that this so-called ‘nutrient–CO2’ connection might link the oceanic phosphorus cycle to climate change due to reductions or increases in the atmospheric greenhouse gas inventory. Oceanic phosphorus residence time and response time (the inverse of residence time) will dictate the timescales

over which such a phosphorus-induced climate effect may operate. Over the past decade there have been several reevaluations of the marine phosphorus cycle in the literature, reflecting changes in our understanding of the identities and magnitudes of important phosphorus sources and sinks. One quantitatively important newly identified marine phosphorus sink is precipitation of disseminated authigenic carbonate fluoroapatite (CFA) in sediments in nonupwelling environments. CFA is the dominant phosphorus mineral in economic phosphorite deposits. Its presence in terrigenous-dominated continental margin environments is detectable only by indirect methods (coupled pore water/solid-phase chemical analyses), because dilution by terrigenous debris renders it below detection limits of direct methods (e.g. X-ray diffraction). Disseminated CFA has now been found in numerous continental margin environments, bearing out early proposals that this is an important marine phosphorus sink. A second class of authigenic phosphorus minerals identified as a quantitatively significant phosphorus sink in sandy continental margin sediments are aluminophosphates. Continental margins in general are quantitatively important sinks for organic and ferric iron-bound phosphorus, as well. When newly calculated phosphorus burial fluxes in continental margins, including the newly identified CFA and aluminophosphate (dominantly aluminum rare-earth phosphate) sinks are combined with older estimates of phosphorus burial fluxes in the deep sea, the overall burial flux results in a much shorter residence time than the canonical value of 100 000 years found in most text books (Table 5). This reduced residence time suggests that the oceanic phosphorus-cycle is subject to perturbations on shorter timescales than has previously been believed. The revised, larger burial flux cannot be balanced by the dissolved riverine input alone. However, when the fraction of riverine particulate phosphorus that is believed to be released upon entering the marine realm is taken into account, the possibility of a balance between inputs and outputs becomes more feasible. Residence times estimated on the basis of phosphorus inputs that include this ‘releasable’ riverine particulate phosphorus fall within the range of residence time estimates derived from phosphorus burial fluxes (Table 5). Despite the large uncertainties associated with these numbers, as evidenced by the maximum and minimum values derived from both input and removal fluxes, these updated residence times are all significantly shorter than the canonical value of 100 000 years. Revised residence times on the order of 10 000–17 000 y make phosphorus-perturbations of the ocean–atmosphere CO2


PHOSPHORUS CYCLE

Table 4

Turnover rates of dissolved inorganic phosphorus (DIP)

reservoir on the timescale of glacial–interglacial climate change feasible. Long Time-scale Phosphorus Cycling, and Links to Other Biogeochemical Cycles

The biogeochemical cycles of phosphorus and carbon are linked through photosynthetic uptake and release during respiration. During times of elevated marine biological productivity, enhanced uptake of surface water CO2 by photosynthetic organisms results in increased CO2 evasion from the atmosphere, which persists until the supply of the least abundant nutrient is exhausted. On geological timescales, phosphorus is likely to function as the limiting nutrient and thus play a role in atmospheric CO2 regulation by limiting CO2 drawdown by oceanic photosynthetic activity. This connection between nutrients and atmospheric CO2 could have played a role in triggering or enhancing the global cooling that resulted in glacial episodes in the geological

a

and dissolved organic phosphorus (DOP)

b

411

in surface sea water

past. It has recently been proposed that tectonics may play the ultimate role in controlling the exogenic phosphorus mass, resulting in long-term phosphorus-limited productivity in the ocean. In this formulation, the balance between subduction of phosphorus bound up in marine sediments and underlying crust and creation of new crystalline rock sets the mass of exogenic phosphorus. Phosphorus and oxygen cycles are linked through the redox chemistry of iron. Ferrous iron is unstable at the Earth’s surface in the presence of oxygen, and oxidizes to form ferric iron oxyhydroxide precipitates, which are extremely efficient scavengers of dissolved phosphate. Resupply of phosphate to surface waters where it can fertilize biological productivity is reduced when oceanic bottom waters are well oxygenated owing to scavenging of phosphate by ferric oxyhydroxides. In contrast, during times in Earth’s history when oxygen was not abundant in the atmosphere (Precambrian), and when expanses of the deep ocean were anoxic (e.g., Cretaceous), the


412

PHOSPHORUS CYCLE

Table 5

Revised oceanic phosphorus input fluxes, removal fluxes, and estimated oceanic residence time

potential for a larger oceanic dissolved phosphate inventory could have been realized due to the reduced importance of sequestering with ferric oxyhydroxides. This iron–phosphorus–oxygen coupling produces a negative feedback, which may have kept atmospheric O2 within equable levels throughout the Phanerozoic. Thus, it is in the oceans that the role of phosphorus as limiting nutrient has the greatest repercussions for the global carbon and oxygen cycles.

Summary The global cycle of phosphorus is truly a biogeochemical cycle, owing to the involvement of phosphorus in both biochemical and geochemical reactions and pathways. There have been marked advances in the last decade on numerous fronts of phosphorus research, resulting from application of new methods as well as rethinking of old assumptions and paradigms. An oceanic phosphorus residence time on the order of 10 000–20 000 y, a factor of 5–10 shorter than previously cited values, casts phosphorus in the role of a potential player in climate change through

the nutrient–CO2 connection. This possibility is bolstered by findings in a number of recent studies that phosphorus does function as the limiting nutrient in some modern oceanic settings. Both oxygen isotopes in phosphate (d18O-PO4) and in situ-produced radiophosphorus isotopes (33P and 32P) are providing new insights into how phosphorus is cycled through metabolic pathways in the marine environment. Finally, new ideas about global phosphorus cycling on long, geological timescales include a possible role for phosphorus in regulating atmospheric oxygen levels via the coupled iron–phosphorus–oxygen cycles, and the potential role of tectonics in setting the exogenic mass of phosphorus. The interplay of new findings in each of these areas is providing us with a fresh look at the marine phosphorus cycle, one that is sure to evolve further as these new areas are explored in more depth by future studies.

See also Carbon Cycle. Eutrophication. Nitrogen Cycle. Phytoplankton Blooms. Redfield Ratio


PHOSPHORUS CYCLE

Further Reading Blake RE, Alt JC, and Martini AM (2001) Oxygen isotope ratios of PO4: an inorganic indicator of enzymatic activity and P metabolism and a new biomarker in the search for life. Proceedings of the National Academy of Sciences of the USA 98: 2148--2153. Benitez-Nelson CR (2001) The biogeochemical cycling of phosphorus in marine systems. Earth-Science Reviews 51: 109--135. Clark LL, Ingall ED, and Benner R (1999) Marine organic phosphorus cycling: novel insights from nuclear magnetic resonance. American Journal of Science 299: 724--737. Colman AS, Holland HD, and Mackenzie FT (1996) Redox stabilization of the atmosphere and oceans by phosphorus limited marine productivity: discussion and reply. Science 276: 406--408. Colman AS, Karl DM, Fogel ML, and Blake RE (2000) A new technique for the measurement of phosphate oxygen isotopes of dissolved inorganic phosphate in natural waters. EOS, Transactions of the American Geophysical Union F176. Delaney ML (1998) Phosphorus accumulation in marine sediments and the oceanic phosphorus cycle. Global Biogeochemical Cycles 12(4): 563--572. Falkowski PG (1997) Evolution of the nitrogen cycle and its influence on the biological sequestration of CO2 in the ocean. Nature 387: 272--275. Fo¨llmi KB (1996) The phosphorus cycle, phosphogenesis and marine phosphate-rich deposits. Earth-Science Reviews 40: 55--124.

413

Guidry MW, Mackenzie FT and Arvidson RS (2000) Role of tectonics in phosphorus distribution and cycling. In: Glenn CR, Pre´vt-Lucas L and Lucas J (eds) Marine Authigenesis: From Global to Microbial. SEPM Special Publication No. 66, pp. 35–51. (See also in this volume Compton et al. (pp. 21–33), Colman and Holland (pp. 53–75), Rasmussen (pp. 89–101), and others.) Howarth RW, Jensen HS, Marino R and Postma H (1995) Transport to and processing of P in near-shore and oceanic waters. In: Tiessen H (ed.) Phosphorus in the Global Environment. SCOPE 54, pp. 232–345. Chichester: Wiley. (See other chapters in this volume for additional information on P-cycling.) Karl DM (1999) A sea of change: biogeochemical variability in the North Pacific Subtropical Gyre. Ecosystems 2: 181--214. Longinelli A and Nuti S (1973) Revised phosphate–water isotopic temperature scale. Earth and Planetary Science Letters 19: 373--376. Palenik B and Dyhrman ST (1998) Recent progress in understanding the regulation of marine primary productivity by phosphorus. In: Lynch JP and Deikman J (eds.) Phosphorus in Plant Biology: Regulatory Roles in Molecular, Cellular, Organismic, and Ecosystem Processes, pp. 26--38. American Society of Plant Physiologists Ruttenberg KC (1993) Reassessment of the oceanic residence time of phosphorus. Chemical Geology 107: 405--409. Tyrell T (1999) The relative influences of nitrogen and phosphorus on oceanic productivity. Nature 400: 525--531.


PHOTOCHEMICAL PROCESSES N. V. Blough, University of Maryland, College Park, MD, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2162–2172, & 2001, Elsevier Ltd.

Introduction Life on Earth is critically dependent on the spectral quality and quantity of radiation received from the sun. The absorption of visible light (wavelengths from 400 to 700 nm) by pigments within terrestrial and marine plants initiates a series of reactions that ultimately transforms the light energy to chemical energy, which is stored as reduced forms of carbon. This complex photochemical process, known as photosynthesis, not only provides all of the chemical energy required for life on Earth’s surface, but also acts to decrease the level of a major greenhouse gas, CO2, in the atmosphere. By contrast, the absorption of ultraviolet light in the UV-B (wavelengths from 280 to 320 nm) and UV-A (wavelengths from 320 to 400 nm) by plants (as well as other organisms) can produce seriously deleterious effects (e.g. photoinhibition), leading to a decrease in the efficiency of photosynthesis and direct DNA damage (UV-B), as well as impairing or destroying other important physiological processes. The level of UV-B radiation received at the Earth’s surface depends on the concentration of ozone (O3) in the stratosphere where it is formed photochemically. The destruction of O3 in polar regions, leading to increased levels of surface UV-B radiation in these locales, has been enhanced by the release of man-made chlorofluorocarbons (CFCs), but may also be influenced in part by the natural production of halogenated compounds by biota. These biotic photoprocesses have long been recognized as critical components of marine ecosystems and air–sea gas exchange, and have been studied extensively. However, only within the last decade or so has the impact of abiotic photoreactions on the chemistry and biology of marine waters and their possible coupling with atmospheric processes been fully appreciated. Light is absorbed in the oceans not only by phytoplankton and water, but also by colored dissolved organic matter (CDOM), particulate detrital matter (PDM), and other numerous trace lightabsorbing species. Light absorption by these constituents, primarily the CDOM, can have a number of important chemical and biological consequences

414

including: (1) reduction of potentially harmful UV-B and UV-A radiation within the water column; (2) photo-oxidative degradation of organic matter through the photochemical production of reactive oxygen species (ROS) such as superoxide (O2), hydrogen peroxide (H2O2), the hydroxyl radical (OH) and peroxy radicals (RO2); (3) changes in metal ion speciation through reactions with the ROS or through direct photochemistry, resulting in the altered biological availability of some metals; (4) photochemical production of a number of trace gases of importance in the atmosphere such as CO2, CO, and carbonyl sulfide (COS), and the destruction of others such as dimethyl sulfide (DMS); (5) the photochemical production of biologically available low molecular weight (LMW) organic compounds and the release of available forms of nitrogen, thus potentially fueling the growth of microorganisms from a biologically resistant source material (the CDOM). These processes provide the focus of this article.

Optical Properties of the Abiotic Constituents of Sea Waters CDOM is a chemically complex material produced by the decay of plants and algae. This material, commonly referred to as gelbstoff, yellow substance, gilvin or humic substances, can be transported from land to the oceans by rivers or be formed directly in marine waters by as yet poorly understood processes. CDOM is the principal light-absorbing component of the dissolved organic matter (DOM) pool in sea waters, far exceeding the contributions of discrete dissolved organic or inorganic light-absorbing compounds. CDOM absorption spectra are broad and unstructured, and typically increase with decreasing wavelength in an approximately exponential fashion (Figure 1). Spectra have thus been parameterized using the expression [1]. aðlÞ ¼ aðl0 Þ eSðll0 Þ

½1

aðlÞ and aðl0 Þ are the absorption coefficients at wavelength l and reference wavelength l0 , respectively, and S defines how rapidly the absorption increases with decreasing wavelength. Absorption coefficients are calculated from relation [2], where A is the absorbance measured across pathlength, r.


a ð lÞ ¼

2:303 AðlÞ r

½2

PHOTOCHEMICAL PROCESSES

415

6.0 (A)

(B)

0.40

0.50 4.0

_

_

a (m 1)

a (m 1)

0.30

_

a (m 1)

0.25

0.20 0.10

0.00 300

2.0

0.00 300

400 500 600 Wavelength (nm)


0.0 (C)

0.15

0.03 _

a (m 1)

_

a (m 1)

1.5

0.10

1.0

0.02 0.01

0.05

_

a (m 1)

0.04

(D)

0.20

0.00 300

0.00 300



0.5

0.0 300

400

500 600 Wavelength (nm)

300

400

500

600

Wavelength (nm)

Figure 1 Absorption spectra of CDOM (—), PDM (– – –) and phytoplankton (– – –) from surface waters in the Delaware Bay and Middle Atlantic Bight off the east coast of the USA in July 1998: (A) mid-Delaware Bay at 39o 9.070 N, 75o 14.290 W; (B) Mouth of the Delaware Bay at 381 48.610 N, 751 5.070 W; (C) Mid-shelf at 381 45.550 N, 741 46.600 W; (D) Outer shelf at 381 5.890 N, 741 9.070 W.

Due to the exponential increase of aðlÞ with decreasing l, CDOM absorbs light strongly in the UVA and UV-B, and thus is usually the principal constituent within marine waters that controls the penetration depth of radiation potentially harmful to organisms (Figure 1). Moreover, for estuarine waters and for coastal waters strongly influenced by river inputs, light absorption by CDOM can extend well into the visible wavelength regime, often dominating the absorption by phytoplankton in the blue portion of the visible spectrum. In this situation, the amount and quality of the photosynthetically active radiation available to phytoplankton is reduced, thus decreasing primary productivity and potentially affecting ecosystem structure. High levels of absorption by CDOM in these regions can also seriously

compromise the determination of phytoplankton biomass through satellite ocean color measurements. As described below, the absorption of sunlight by CDOM also initiates the formation of a variety of photochemical intermediates and products. The photochemical reactions producing these species ultimately lead to the degradation of the CDOM and the loss, or bleaching, of its absorption. This process can act as a feedback to alter the aquatic light field. Particulate detrital material (PDM), operationally defined as that light-absorbing material retained on a GFF filter and not extractable with methanol, is a composite of suspended plant degradation products and sediment that also exhibits an exponentially rising absorption with decreasing wavelength (Figure 1); eqn [2] has thus been used to parameterize this


416


material as well. However, the values of S acquired for this material are usually smaller than those of the CDOM. In estuarine and near-shore waters, and in shallow coastal waters subject to resuspension of bottom sediments, PDM can contribute substantially to the total water column absorption. However, in most marine waters, the PDM is a rather minor constituent. Little is known about its photochemical reactivity. Other light-absorbing trace organic compounds such as flavins, as well as inorganic compounds such as nitrate, nitrite, and metal complexes, do not contribute significantly to the total water column absorption. However, many of these compounds are quite photoreactive and will undergo rapid transformation under appropriate light fields.

Photochemical Production of Reactive Oxygen Species CDOM is stituent in gests that dominated

the principal abiotic photoreactive conmarine waters. Available evidence sugthe photochemistry of this material is by reactions with dioxygen (O2) in a

process known as photo-oxidation. In this process, O2 can act to accept electrons from excited states, radicals (highly reactive species containing an unpaired electron) or radical ions generated within the CDOM by the absorption of light. This leads to the production of a variety of partially reduced oxygen species such as superoxide (O 2 , the one-electron reduction product of O2), hydrogen peroxide (H2O2, the two-electron reduction product of O2), peroxy radicals (RO2, formed by addition of O2 to carboncentered radicals, R) and organic peroxides (RO2H), along with the concomitant oxidation of the CDOM (Figure 2). Many of these reduced oxygen species as well as the hydroxyl radical (OH), which is generated by other photochemical reactions, are also quite reactive. These reactive oxygen species or ROS can undergo additional secondary reactions with themselves or with other organic and inorganic seawater constituents. The net result of this complex series of reactions is the light-induced oxidative degradation of organic matter by dioxygen (Figure 2). This process leads to the consumption of O2, the production of oxidized carbon gases (CO2, CO, COS), the formation of a variety of LMW organic compounds, the release of biologically available forms of nitrogen,

(fluorescence) h ′ + heat + CDOM _

+h

_

NO2 , NO3

+h

+ Br

+ HCO3

+

+ O2 ±:

R·

_

CDOM Radical ions

Carbon-centered radicals (eg. CH3, CH3CO)

_

RO2 Peroxy radicals

CO3·

_

Secondary reactions

+ O2 _

2O2 + 2H +

_ H+

+ O2

R'H

_

_

CDOM· + e (aq)

·OH

NO, NO2

Br2

CDOM*

H2O2 + O2 Hydrogen peroxide

Superoxide (?)

CDOM· Carbon-centred radicals

Secondary reactions

CDOMoxidized + H+

+ O2 RO Alkoxy radicals

+

RO2H Organic peroxides

CDOM-OO· Peroxy radicals

Secondar y reactions Secondary reactions

Secondary reactions Oxidized products (CO, CO2,COS, LMW organic compounds) Figure 2 Schematic representation of the photochemical and secondary reactions known or thought to occur following light absorption by CDOM. For a more detailed description of these reactions see the text, Blough and Zepp (1995), and Blough (1997). Not shown in this diagram are primary and secondary reactions of metal species; for a description of these processes, see Helz et al. (1994) and Blough and Zepp (1995).



NO3 þ hn-O þ NO2

½I

NO2 þ hn-O þ NO

½II

O þ H2 O-OH þ OH

½III

2.0 Lake Valkea-kotiten

Φ ( × 103)

1.5

1.0

0.5 CO2 0.0 River water Suwanee

3.0 Φ ( × 104)

Houghton Kinoshe

2.0

Okefenokee 1.0 CO 0.0 Sea water 4.0

Φ ( × 107)

and the loss of CDOM absorption. Through direct photochemical reactions and reactions with the ROS, the speciation of metal ions is also affected. These photochemical intermediates and products are produced at relatively low efficiencies. About 98–99% of the photons absorbed by CDOM are released as heat, while another B1% are re-emitted as fluorescence. These percentages (or fractions) of absorbed photons giving rise to particular photoresponses are known as quantum yields ðFÞ. The F for the production of H2O2 and O 2 (the two reduced oxygen species produced with highest efficiency), are approximately one to two orders of magnitude smaller than those for fluorescence, ranging from B0.1% at 300 nm to B0.01% at 400 nm. The F for other intermediates and products range even lower, from B0.01% to 0.000 0001% (see below). The F for most of the intermediates and products created from the CDOM are highest in the UV-B and UV-A, and fall off rapidly with increasing wavelength; yields at visible wavelengths are usually negligible (see for example, Figure 3). The hydroxyl radical, a very powerful oxidant, can be produced by the direct photolysis of nitrate and nitrite (eqns [I]–[III]).

417

2.0

COS

The F values for these reactions are relatively high, about 7% for nitrite and about 1–2% for nitrate. However, because of the relatively low concentrations of these compounds in most marine surface waters, as well as their low molar absorptivities in the ultraviolet, the fraction of light absorbed is generally small and thus fluxes of OH from these sources also tend to be small. Recent evidence suggests that OH, or a species exhibiting very similar reactivity, is produced through a direct photoreaction of the CDOM; quinoid moieties within the CDOM may be responsible for this production. Quantum yields are low, B0.01%, and restricted primarily to the ultraviolet. In estuarine and near-shore waters containing higher levels of iron, the production of OH may also occur through the direct photolysis of iron–hydroxy complexes or through the Fenton reaction (eqn [IV]).

0.0 300

320

340 360 380 400 Wavelength (nm)

420

440

Figure 3 Wavelength dependence of the quantum yields (F) for the photochemical production of CO2, CO, and COS. Data have been replotted from those dependencies originally reported in Va¨ha¨talo et al. (2000), Valentine and Zepp (1993), and Weiss et al. (1995) for CO2, CO, and COS, respectively.

produced by direct photoreactions of a light-absorbing constituent such as CDOM can react secondarily with the nonabsorbing compounds. DMS and COS, two trace gases of some importance to the atmosphere, are thought to be destroyed and created, respectively, by sensitized photoreactions in marine surface waters.

Photochemical Production and Consumption of Low Molecular ½IV Fe2þ þ H2 O2 -Fe3þ þ OH þ OH Weight Organic Compounds and Compounds that do not absorb light within the surTrace Gases face solar spectrum are also subject to photochemical modification through indirect or ‘sensitized’ photoreactions. In this case, the ROS or intermediates

The photolysis of CDOM produces a suite of LMW organic compounds and a number of trace gases. The


418


production of these species presumably occurs through radical and fragmentation reactions arising from the net oxidative flow of electrons from CDOM to O2 (see above), although the exact mechanism(s) have yet to be established. Most LMW organic compounds produced contain three or fewer carbon atoms and include such species as acetaldehyde, acetate, acetone, formaldehyde, formate, glyoxal, glyoxalate, methylglyoxal, propanal, and pyruvate. The F values for the production of individual compounds are low, B0.001–0.0001%, with wavelengths in the UV-B the most effective; efficiencies decrease rapidly with increasing wavelength. Available evidence indicates that the F for O2 and H2O2 production are about one to two orders of magnitude larger than those for the LMW organic compounds, so it appears that the sum of the production rates for the known LMW organic compounds is small with respect to the flux of photochemical equivalents from CDOM to O2. Most, if not all, of these products are rapidly taken up and respired by bacteria to CO2. Numerous investigators have presented evidence supporting enhanced microbial activity in waters exposed to sunlight, with bacterial activities increasing from 1.5to almost 6-fold depending presumably on the length and type of light exposure, and the concentration and source of the CDOM. Recently, biologically labile nitrogen-containing compounds such as ammonia and amino acids have also been reported to be produced photochemically from CDOM. Because CDOM is normally considered to be biologically refractive, this recent work highlights the important role that abiotic photochemistry plays in the degradation of CDOM, not only through direct photoreactions, but also through the formation of biologically available products that can be respired to CO2 or used as nutrients by biota. A recent estimate suggests that the utilization of biologically labile photoproducts could account for as much as 21% of the bacterial production in some near-surface waters. Carbon dioxide and carbon monoxide are major products of the direct photolysis of CDOM (Figure 3). Quantum yields for CO production are about an order of magnitude smaller than those for O 2 and H2O2 production, ranging from B0.01% at 300 nm to B0.001% at 400 nm. Available data indicate that the F for CO2 range even higher, perhaps as much as 15–20-fold. The yields for CO2 production must thus approach, if not exceed, those for O 2 and H2O2. This result is somewhat surprising, since it implies that about one CO2 is produced for each electron transferred from the CDOM to O2, further implying a high average redox state for CDOM. Although CDOM (i.e. humic substances) is

known to contain significant numbers of carboxyl moieties that could serve as the source of the CO2, the yield for CO2 production, relative to O2 and H2O2, would be expected to fall rapidly as these groups were removed photochemically; available evidence suggests that this does not occur. An alternative explanation is that other species, perhaps the CDOM itself, is acting as an electron acceptor. Regardless of mechanism, existing information indicates that CO2 is the dominant product of CDOM photolysis (Figure 3). A recent estimate suggests that the annual global photoproduction of CO in the oceans could be as high as 0.82 1015 g C. Assuming that CO2 photoproduction is 15–20 times higher than that for CO, values for CO2 formation could reach from 12 to 16 1015 g C y1. To place these numbers in perspective, the estimated annual input of terrestrial dissolved organic carbon to the oceans (0.2 1015 g C y1) is only 1.3–1.7% of the calculated annual CO2 photoproduction, which is itself about 2–3% of the oceanic dissolved organic carbon pool. These calculated CO2 (and CO) photoproduction rates may be high due to a number of assumptions, including (1) the complete absorption of UV radiation by the CDOM throughout the oceans, (2) constant quantum yields (or action spectra) for production independent of locale or light history, and (3) neglecting mass transfer limitations associated with physical mixing. Nevertheless, these estimates clearly highlight the potential impact of abiotic photochemistry on the oceanic carbon cycle. Moreover, the products of this photochemistry are generated in near-surface waters where exchange with the atmosphere can take place readily. Like the LMW organic compounds, bacteria can oxidize CO to CO2; this consumption takes place in competition with the release of CO to the atmosphere. Due to its photochemical production, CO exists at supersaturated concentrations in the surface waters of most of the Earth’s oceans. Recent estimates indicate that global oceanic CO emissions could range from 0.013 1015 g y1–1.2 1015 g y1 (see above). The upper estimate is based on calculated photochemical fluxes (see below) and the assumption that all CO produced is emitted to the atmosphere. The lower estimate was calculated using air–sea gas exchange equations and extensive measurements of CO concentrations in the surface waters and atmosphere of the Pacific Ocean. The source of the significant discrepancy between these two estimates has yet to be resolved. Depending on the answer, CO emitted to the atmosphere from the oceans could play a significant role in controlling OH levels in the marine troposphere.



Carbonyl sulfide (COS) is produced primarily in coastal/shelf waters, apparently by the CDOM-photosensitized oxidation of organosulfur compounds. UV-B light is the most effective in its formation, with F decreasing rapidly from B6 107 at 300 nm to B1 108 by 400 nm (Figure 3). The principal sinks of seawater COS are release to the atmosphere and hydrolysis to CO2 and H2S. Accounting for perhaps as much as one-third of the total source strength, the photochemical production of COS in the oceans is probably the single largest source of COS to the atmosphere, although more recent work has revised this estimate downward. Smaller amounts of carbon disulfide (CS2) are also generated photochemically in surface waters through CDOM sensitized reaction(s); F values decrease from B1 107 at 313 nm to 5 109 at 366 nm. The CS2 emitted to the atmosphere can react with OH to form additional COS in the troposphere. Although it was previously thought that the oxidation of COS in the stratosphere to form sulfate aerosol could be important in determining Earth’s radiation budget and perhaps in regulating stratospheric ozone concentrations, more recent work suggests that other sources contribute more significantly to the background sulfate in the stratosphere. Dimethyl sulfide (DMS), through its oxidation to sulfate in the troposphere, acts as a source of cloud condensation nuclei, thus potentially influencing the radiative balance of the atmosphere. DMS is formed in sea water through the microbial decomposition of dimethyl sulfonioproprionate (DMSP), a compound believed to act as an osmolyte in certain species of marine phytoplankton. The flux of DMS to the atmosphere is controlled by its concentration in surface sea waters, which is controlled in turn by the rate of its decomposition. Estimates indicate that 7–40% of the total turnover of DMS in the surface waters of the Pacific Ocean is due to the photosensitized destruction of this compound, illustrating the potential importance of this pathway in controlling the flux of DMS to the atmosphere. In addition to these compounds, the photochemical production of small amounts of nonmethane hydrocarbons (NMHC) such as ethene, propene, ethane, and propane has also been reported. Production of these compounds appears to result from the photolysis of the CDOM, with F values of the order of 107–109. The overall emission rates of these compounds to the atmosphere via this source are negligible with respect to global volatile organic carbon emissions, although this production may play some role in certain restricted locales exhibiting stronger source strengths, or in the marine environment remote from the dominant terrestrial sources.

419

The photolysis of nitrate and nitrite in sea water produces nitrogen dioxide (NO2) and nitric oxide (NO), respectively (eqns [I] and [II]). Previous work indicated that the photolysis of nitrite could act as a small net source of NO to the marine atmosphere under some conditions. However, this conclusion seems to be at odds with estimates of the steady-state concentrations of superoxide and the now known rate constant for the reaction of superoxide with nitric oxide (6.7 109 M1 s1) to form peroxynitrite in aqueous phases (eqn [V]). O 2 þ NO- OONO

½V

The peroxynitrite subsequently rearranges in part to form nitrate (eqn [VI]).

OONO-NO3

½VI

Even assuming a steady-state concentration of O2 (1012 M) that is about two orders of magnitude lower than that expected for surface sea waters (B1010 M), the lifetime of NO in surface sea waters would be only B150 s, a timescale too short for significant exchange with the atmosphere except for a thin surface layer. Moreover, even in this situation, the atmospheric deposition of additional HO2 radicals to this surface layer (to form O2) would be expected to act as an additional sink of the NO (flux capping). It appears that most if not all water bodies exhibiting significant steady-state levels of O2, produced either photochemically or thermally, should act as a net sink of atmospheric NO and probably of NO2 as well. Further, although less is known about the steady-state levels of peroxy radicals in sea waters due largely to their unknown decomposition routes, their high rate constants for reaction with NO (1– 3 109 M1 s1) indicate that they should also act as a sink of NO. In fact, methyl nitrate, a trace species found in sea waters, may in part be produced through the aqueous phase reactions (eqns [VII] and [VIII]) with the methylperoxy radical (CH3OO) generated through a known photochemical reaction of CDOM (or through atmospheric deposition) and the NO arising from the photolysis of nitrite (or through atmospheric deposition). CH3 OO þ NO-CH3 OONO

½VII

CH3 OONO-CH3 ONO2

½VIII

The concentrations of NO and NO2 in the troposphere are important because of the involvement of these gases in the formation of ozone. The atmospheric deposition of ozone to the sea surface can cause the release of volatile iodine


420


compounds to the atmosphere. There is also evidence that methyl iodide can be produced (as well as destroyed) by photochemical processes in surface sea waters. The release of these volatile iodine species from the sea surface or from atmospheric aqueous phases (aerosols) by these processes may act as a control on the level of ozone in the marine troposphere via iodine-catalyzed ozone destruction.

Here Fðl; zÞ is the photochemical production (or consumption) rate; ED ðl; zÞ is the downwelling irradiance at wavelength, l, and depth, z, within the water column; aDi is the diffuse absorption coefficient for photoreactive constituent i; Fi ðlÞ is the quantum yield of this ith constituent. ED ðl; zÞ is well approximated by eqn [4]. ED ðl; zÞ ¼ ED0 ðlÞ eKd ðlÞz

½4

Trace Metal Photochemistry A lack of available iron is now thought to limit primary productivity in certain ocean waters containing high nutrient, but low chlorophyll concentrations (the HNLC regions). This idea has spurred interest in the transport and photochemical reactions of iron in both seawaters and atmospheric aerosols. Very little soluble Fe(II) is expected to be available at the pH and dioxygen concentration of surface seawaters due to the high stability of the colloidal iron (hydr)oxides. The photoreductive dissolution of colloidal iron oxides by CDOM is known to occur at low pH; this process is also thought to occur in seawaters at high pH, but the reduced iron appears to be oxidized more rapidly than its detachment from the oxide surface. However, some workers have found that CDOM-driven cycles of reduction followed by oxidation increases the chemical availability, which was strongly correlated with the growth rate of phytoplankton. Significant levels of Fe(II) are also known to be produced photochemically in atmospheric aqueous phases (at lower pH) and could serve as a source of biologically available iron upon deposition to the sea surface. Manganese oxides are also subject to reductive dissolution by light in surface seawaters. This process produces Mn(II), which is kinetically stable to oxidation in the absence of bacteria that are subject to photoinhibition. These two effects lead to the formation of a surface maximum in soluble Mn(II), in contrast to most metals which are depleted in surface waters due to biological removal processes. Other examples of the impact of photochemical reactions on trace metal chemistry are provided in Further Reading.

Photochemical Calculations Global and regional estimates for the direct photochemical production (or consumption) of a particular photoproduct (or photoreactant) can be acquired with knowledge of the temporal and spatial variation of the solar irradiance reaching the Earth’s surface combined with a simple photochemical model (eqn [3]). Fðl; zÞ ¼ ED ðl; zÞ Fi ðlÞ aDi ðlÞ

½3

ED0 ðlÞ is the downwelling irradiance just below the sea surface and Kd ðlÞ is the vertical diffuse attenuation coefficient of downwelling irradiance. Kd ðlÞ can be approximated by eqn [5]. P Kd ðlÞE

P ai ðlÞ þ bbi ðlÞ mD

½5

P P where ai ðlÞ and bbi ðlÞ are the total absorption and backscattering coefficients, respectively, of all absorbing and scattering constituents within the water column, and mD is the average cosine of the angular distribution of the downwelling light. This factor accounts for the average pathlength of light in the water column, and for direct solar light is approximately equal to cos y, where y is the solar zenith angle (e.g. mD B1 when the sun is directly overhead). The diffuse absorption coefficient, aDi , is given by eqn [6]. aDi ¼

ai mD

½6

This model assumes that the water column is homogeneous, that Kd ðlÞ is constant with depth, and that upwelling irradiance is negligible relative to ED ðl; zÞ. Combining eqns [3], [4] and [6] gives eqn [7]. Fðl; zÞ ¼

ED0 ðlÞ:eKd ðlÞz Fi ðlÞ ai ðlÞ mD

½7

This equation allows calculation of the spectral dependence of the production (consumption) rate as a function of depth in the water column, assuming knowledge of ED0 ðlÞ, Kd ðlÞ, ai ðlÞ and Fi ðlÞ, all of which can be measured or estimated (Figure 4). Integrating over wavelength provides the total production (consumption) rate at each depth. Integration of eqn [7] from the surface to depth z provides the spectral dependence of the photochemical flux ðYÞ over this interval Y ðl; zÞ ¼

ED0 ðlÞ: 1 eKd ðlÞz Fi ðlÞ ai ðlÞ=Kd ðlÞ mD


½8


2.0 x 1014

0.03 Solar irradiance

_1

a CDOM (cm )

1.5 x 1014

0.02 1.0 x 1014 0.01 0.5 x 1014

0.0

0.00 (B)

2x107

F (molecules cm

s

_3 _1

_

Solar irradiance _2 _1 _1 (photons cm s nm )

(A)

CDOM

nm 1)

421

1x10

7

0 107

(C)

6

10

log F

105 104 103 102 10 1 300

350 400 Wavelength (nm)

450

Figure 4 Spectral dependence of CO photoproduction rates with depth, plotted on a linear (B) and logarithmic (C) scale. Depths in (B) are (from top to bottom): surface, 0.5, 1, 1.5, and 2 m. Depths in (C) are (from top to bottom): surface, 0.5, 1, 1.5, 2, 4, 6, 8, and 10 m. These spectral dependencies were calculated using eqn [7], the wavelength dependence of the quantum yield for CO shown in Figure 3, and the CDOM absorption spectrum and surface solar irradiance shown in (A). The attenuation of irradiance down the water column in this spectral region was assumed to be only due to CDOM absorption, a reasonable assumption for coastal waters (see Figure 1). Note the rapid attenuation in production rates with depth in the UV-B, due to the greater light absorption by CDOM in this spectral region.

which upon substitution of eqn [4] becomes, Y ðl; zÞ ¼ ED0 ðlÞ: 1 eKd ðlÞz Fi ðlÞ P

ai ð lÞ P ai ðlÞ þ bbi ðlÞ

½9

In most, but not all seawaters, the total absorption will P greater than the total backscatter, P be much ai ðlÞb bbi ðlÞ, and thus the backscatter can be ignored; this approximation is not valid for most estuarine waters and some coastal waters, where a more sophisticated treatment would have to be applied. This approximation leads to the final


422


3.0 × 1010 CO2 2.5 × 1010

2.0 × 1010 1.5 × 1010 1.0 × 1010 0.5 × 1010

0

1.5 × 109

Y (molecules cm

s

_2 _1

_

nm 1)

CO

1.0 × 109

0.5 × 109

0 COS 6

6 × 10

4 × 106

2 × 106

0 300

400

350

450

Wavelength (nm) Figure 5 Spectral dependence of the photochemical flux with depth for CO2, CO, and COS. Fluxes with depth are from the surface to 0.25, 0.5, 1.0, 2.0, and 4 m, respectively (bottom spectrum to top spectrum). Below 4 m, increases in the flux are nominal. These spectral dependencies were calculated using eqn [10], the wavelength dependence of the quantum yields for CO2, CO and COS shown in Figure 3, and the surface solar irradiance shown in Figure 4A. CDOM is assumed to absorb all photons in this spectral region (see Figures 1 and 4).



expression for the variation of the spectral dependence of the flux with depth (Figure 5), ai ð lÞ Fðl; zÞ ¼ ED0 ðlÞ:ð1 eKd ðlÞz Þ Fi ðlÞ P ai ð l Þ

½11

with the total flux obtained by integrating over wavelength, ð ai ð lÞ F ED0 ðlÞ:Fi ðlÞ P dl ai ð lÞ

½12

l

To obtain global estimates of photochemical fluxes, many investigators assume that the absorption due to CDOM, aCDOM, dominates the absorption of all other seawater constituents in the ultraviolet, and P thus that aCDOM(l)/ ai(l)E1. While this approximation is reasonable for many coastal waters, it is not clear that this approximation is valid for all oligotrophic waters. This approximation leads to the final expression for flux, ð

Y ED0 ðlÞ:Fi ðlÞdl

couplings between atmospheric gas phase reactions and photochemical reactions in atmospheric aqueous phases.

½10

The spectral dependence of the total water column flux (z-N) is then given by, ai ð lÞ FðlÞ ¼ ED0 ðlÞ Fi ðlÞ P ai ð lÞ

423

½13

l

which relies only on the surface downwelling irradiance and the wavelength dependence of the quantum yield for the photoreaction of interest. Uncertainties in the use of this equation for estimating global photochemical fluxes include (1) the (usual) assumption that FðlÞ acquired for a limited number of samples is representative of all ocean waters, independent of locale or light history, and (2) differences in the spatially and temporally averaged values of ED0 ðlÞ utilized by different investigators.

Conclusions The absorption of solar radiation by abiotic sea water constituents initiates a cascade of reactions leading to the photo-oxidative degradation of organic matter and the concomitant production (or consumption) of a variety of trace gases and LMW organic compounds (Figure 2), as well as affecting trace metal speciation. The magnitude and impact of these processes on upper ocean biogeochemical cycles and their coupling with atmospheric processes are just beginning to be fully quantified and understood. There remains the need to examine possible

See also Air–Sea Transfer: Dimethyl Sulfide, COS, CS2, NH4, Non-Methane Hydrocarbons, Organo-Halogens. Air–Sea Transfer: N2O, NO, CH4, CO.

Further Reading Blough NV (1997) Photochemistry in the sea-surface microlayer. In: Liss PS and Duce R (eds.) The Sea Surface and Global Change, pp. 383--424. Cambridge: Cambrige University Press. Blough NV and Green SA (1995) Spectroscopic characterization and remote sensing of non-living organic matter. In: Zepp RG and Sonntag C (eds.) The role of Non-living Organic Matter in the Earth’s Carbon Cycle, pp. 23--45. New York: John Wiley. Blough NV and Zepp RG (1995) Reactive oxygen species in natural waters. In: Foote CS, Valentine JS, Greenberg A, and Liebman JF (eds.) Reactive Oxygen Species in Chemistry, pp. 280--333. New York: Chapman & Hall. de Mora S, Demers S, and Vernet M (eds.) (2000) The Effects of UV Radiation in the Marine Environment. Cambridge: Cambridge University Press. Ha¨der D-P, Kumar HD, Smith RC, and Worrest RC (1998) Effects of UV-B radiation on aquatic ecosystems. Journal of Photochemistry and Photobiology B 46: 53--68. Helz GR, Zepp RG, and Crosby DG (eds.) (1994) Aquatic and Surface Photochemistry. Ann Arbor, MI: Lewis Publishers. Huie RE (1995) Free radical chemistry of the atmospheric aqueous phase. In: Barker JR (ed.) Progress and Problems in Atmospheric Chemistry, pp. 374--419. Singapore: World Scientific Publishing Co. Kirk JTO (1994) Light and Photosynthesis in Aquatic Ecosystems. Cambridge: Cambridge University Press. Moran MA and Zepp RG (1997) Role of photoreactions in the formation of biologically labile compounds from dissolved organic matter. Limnology and Oceanography 42: 1307--1316. Thompson AM and Zafiriou OC (1983) Air–sea fluxes of transient atmospheric species. Journal of Geophysical Research 88: 6696--6708. Va¨ha¨talo AV, Salkinoja-Salonen M, Taalas P, and Salonen K (2000) Spectrum of the quantum yield for photochemical mineralization of dissolved organic carbon in a humic lake. Limnology and Oceanography 45: 664--676. Valentine RL and Zepp RG (1993) Formation of carbon monoxide from the photodegradation of terrestrial dissolved organic carbon in natural waters. Environmental Science Technology 27: 409--412.


424


Weiss EW, Andrews SS, Johnson JE, and Zafiriou OC (1995) Photoproduction of carbonyl sulfide in south Pacific Ocean waters as a function of irradiation wavelength. Geophysical Research Letters 22: 215--218. Zafiriou OC, Blough NV, Micinski E, et al. (1990) Molecular probe systems for reactive transients in natural waters. Marine Chemistry 30: 45--70.

Zepp RG, Callaghan TV, and Erickson DJ (1998) Effects of enhanced solar ultraviolet radiation on biogeochemical cycles. Journal of Photochemistry and Photobiology B 46: 69--82.


PHYTOBENTHOS M. Wilkinson, Heriot-Watt University, Edinburgh, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2172–2179, & 2001, Elsevier Ltd.

What is Phytobenthos? ‘Phytobenthos’ means plants of the seabed, both intertidal and subtidal, and both sedimentary and hard. Such plants belong almost entirely to the algae although seagrasses, which form meadows on some subtidal and intertidal areas, are flowering plants or angiosperms. Algae of sedimentary shores are usually microscopic, unicellular or filamentous, and are known as the microphytobenthos or benthic microalgae. Marine algae on hard surfaces can range from microscopic single-celled forms to large cartilaginous plants. Some use the term ‘seaweed’ for macroscopic forms whereas others also include the smaller algae of rocky seashores. This article is concerned with the nature, diversity, ecology, and exploitation of the marine benthic algae. Other plants of the shore are dealt with elsewhere in this encyclopedia as salt-marshes, mangroves, and seagrasses. (see Mangroves; Salt Marsh Vegetation).

What are Algae? Algae were regarded as the least highly evolved members of the plant kingdom. Nowadays most classifications either regard the microscopic algae as protists, while leaving the macroscopic ones in the plant kingdom, or regard all algae as protists. This distinction is not important for an understanding of the ecological role of these organisms so they will all be called plants in this article. The fundamental feature that algae share in common with the rest of the plant kingdom is photoautotrophic nutrition. In photosynthesis they convert inorganic carbon (as carbon dioxide, carbonate, or bicarbonate) into organic carbon using light energy. Thus they are primary producers, which act as the route of entry of carbon and energy into food chains. They are not the only autotrophs in the sea. Besides the other nonalgal plant communities mentioned earlier, there are chemoautotrophs in hydrothermal vent communities, which use inorganic reactions rather than light as the energy source, and some bacteria are photosynthetic. However, algae are responsible for at least 95% of marine primary production. Algal

photosynthesis uses chlorophyll a as the principal pigment that traps and converts light energy into chemical energy (although many accessory photosynthetic pigments may also be present) and water is the source of the hydrogen that is used to reduce inorganic carbon to carbohydrate. Oxygen is a byproduct so the process is called oxygenic photosynthesis. Broadly the same process occurs throughout the plant kingdom. Those true bacteria that are photosynthetic use alternative pathways, pigments, and hydrogen donors, e.g., hydrogen sulfide. One group of organisms falls between the algae and the bacteria – the cyanobacteria, until recently regarded as blue–green algae. These perform oxygenic photosynthesis and have similar photosynthetic pigments to algae, but their cell structure is fundamentally different. In common with the bacteria they have the more primitive, prokaryotic cell structure, lacking membrane-bound organelles and organized nuclei. Algae have the more advanced and efficient eukaryotic cell structure, with membrane-bound organelles and defined nuclei, in common with all other plants and animals. Blue–greens also have some physiological affinities with bacteria, particularly nitrogen fixation. This means that they can use elemental nitrogen as a source of nitrogen for biosynthesis of various organic nitrogen compounds, starting from the simple organic compounds formed in photosynthesis. Algae and other plants have lost this ability and require to absorb fixed nitrogen, combined inorganic nitrogen as nitrate and ammonium ions, from solution in sea water. Despite internal cellular differences, blue–greens have similar overall morphology to smaller algae and live indistinguishably in algal communities as primary producers. They will therefore be included with algae in this review. Algae are therefore similar to the ‘true’ plants in having oxygenic photosynthesis. They are distinguished from the rest of the plant kingdom only on a rather technical botanical point. Algae have simpler reproductive structures. In all the higher plant groups the reproductive organs are surrounded by walls of sterile cells (i.e., cells that are not gametes or spores) that are formed as a specific part of the reproductive organ. This does not occur in algae. Even in the highly complex large brown seaweeds with apparently complex reproductive structures, the gametangia, which produce eggs and sperm, enclosed within complex reproductive structures, have only membranous walls rather than cellular walls.


425

426

PHYTOBENTHOS

Diversity of Algae Algae as defined in the previous section include a large diversity of organisms from microscopic singlecelled ones, as little as about 2 mm in diameter, to the complex giant kelp nearly 70 m long, the largest plant on Earth. We can make sense of this diversity in three ways:

• • •

Structural diversity Habitat diversity Taxonomic classification

Structural Diversity of Algae

The simplest algae are single cells, which can vary in size from about 2 mm to 1 mm. They can be nonmotile, lacking flagella, or motile by means of flagella. An interesting intermediate situation is in the diatoms which lack flagella but are nonetheless motile by gliding over surfaces. Unicells can differ in the presence of external sculpturing and the number and orientation of flagella on each cell. Colonies are aggregations of single cells which can also be flagellate or nonflagellate. The simplest truly multicellular algae are filamentous, i.e., hair-like, chains of cells. These can be branched or unbranched and may be only one cell in thickness (uniseriate) or more than one cell in thickness (multiseriate). Heterotrichy is an advanced form of filamentous construction in which two separate branched systems of filaments may be present on one plant – a prostrate system which creeps along the substratum and an erect system which arises into the seawater medium from the prostrate system. The larger more advanced types of seaweeds can be traced in origin to modifications of the heterotrichous system. Reduction of one of the two filament systems and elaboration of the other can give rise either to encrusting forms (erect reduced, prostrate elaborated) or to erect plants in which the prostrate system only forms the attachment organ or holdfast. Three further modifications give rise to a wide diversity of large cartilaginous (leathery), foliose (leaf-like) and complex filamentous seaweeds. These three modifications are:

•

•

Presence of meristems, localized areas where cell division is concentrated, which may be apical, at the growing tips of branches, or may be intercalary, located along the length of the plant (in simpler algae growth is diffuse with cell division occurring anywhere in the plant, not localized to meristems). Pseudoparenchyma formation – the aggregation of many separate filaments together to make a

•

massive plant body, as opposed to true parenchyma formation, where massive tissues result only from multiplication of adjacent cells. This is a different use of the term parenchyma from that in higher plants where it means an unspecialized type of cell which acts as packing tissue. The occurrence of two or more phases of growth, which may or may not differ in pattern (pseudoparenchymatous or truly parenchymatous) and may involve formation of a secondary lateral meristem. Various phases of growth can give rise to the different tissues seen in cross-sections of seaweeds which may help in giving the ability to bend in response to water motion and wave action, without breaking.

Some seaweeds are able to secrete calcium carbonate so that they appear solid. Red calcareous species appear pink and are common throughout the world whereas green calcareous forms are commoner in the tropics and subtropics. Calcareous encrusting red algae form a pink calcareous coating on the rock surface which can be mistaken by the nonspecialist for a geological feature. The structure of a kelp plant illustrates the life of seaweeds. The plant is attached to the rock surface by a holdfast, which is branched and fits intimately to the microtopography of the rocks. From the holdfast arises the stipe, a stem-like structure which supports the frond in the water column. The frond is a wide flat area which gives a high surface area to volume ratio for light, carbon dioxide, and nutrient absorption. Superficially there is a resemblance to the roots, stem, and leaves of higher plants but there is no real equivalence because of the different life style. The holdfast is not an absorptive root system and does not penetrate the substratum, unlike roots penetrating the soil, since the seaweed can obtain all its requirements by direct absorption over its surface. The stipe is not a stem containing transport systems, as in higher plants, since these are not needed, again because of direct absorption. Similarly the frond does not have the complex structure of leaves with gas exchange and water retention organs such as stomata. Seaweeds do not have resistant phases such as seeds. Development is direct from spores or zygotes. A seaweed, such as a kelp, can be viewed as a chemical factory taking in light, nutrients, and inorganic carbon from the water and converting them into organic matter. This production can be going on even when the plant does not seem to be increasing in size. The formation of new organic matter is then balanced by the loss of decaying tissue from the tip of the plant and by organic secretions


PHYTOBENTHOS

from the frond. Both of these will be contributing to heterotrophic production in the kelp’s ecosystem. Habitat Diversity of Benthic Algae

Microphytobenthos in sedimentary shores can be distinguished according to whether they are epipsammic (attached to sand particles) or epipelic (between mud particles). They are mainly unicellular forms: diatoms, euglenoids, and blue–greens in estuarine muds; diatoms, dinoflagellates, and blue– greens in sand. Many show vertical migration within the top few millimeters of the sediment, photosynthesizing when the tide is out and burrowing before the return of the tide so that some escape being washed away. This is not 100% effective so that resuspended microphytobenthos can be a significant proportion of apparent phytoplankton in some estuaries. Diatoms migrate by gliding motility whereas euglenoids do so by alternate contraction and relaxation of the cell shape (metaboly). In estuaries, which may be turbid environments where photosynthesis by submerged plants may be reduced, and large expanses of intertidal mud flats may be available, microphytobenthos could be important primary producers which have been underestimated because they are not visually obvious. They may also help to stabilize sediments by the mucus secretions which keep them from desiccation when on the mud surface. Seaweeds do not just grow attached to hard surfaces such as bedrock, boulders, and artificial structures. In Britain, a habitat-diverse, open coast shore is likely to have 70–100 species of seaweed present out of a British total of about 630 species. Such a high total on a shore is only realized because of many habitat variations that harbor the more microscopic species. Most seaweeds have smaller species that grow attached to them as epiphytes. In turn they have even smaller species attached and this may continue for several orders down to very small microscopic plants. Endophytes are microscopic algae that grow between the cells within the tissues of larger ones. Epizoic algae grow attached to animals and endozoic forms grow inside animals, usually in skeletal parts. These include algae that penetrate calcareous substrata, i.e., shell-boring algae – red, green, and blue– green forms that bore through mollusk and barnacle shells and coral skeletons. They also include forms that inhabit proteinaceous animal skeletons – red and green filamentous species in skeletons of hydroids and bryozoans, and filamentous green seaweeds, mainly Tellamia, in the periostracum of periwinkles. Some of the shell-boring algae are also

427

endolithic, boring through chalk rocks and so possibly aiding coastal erosion. Finally, there are endozoic algae that live in soft parts of benthic animals. Zooxanthellae are nonmotile dinoflagellate unicells in coral polyps, where they contribute to the high productivity of coral reefs, and unicellular blue– greens live in the tissues of some sea-slugs.

Taxonomic Classification of Algae

Algae are classified into a number of divisions of the plant kingdom (equivalent to phylum), varying in number from about 8 to 16 depending on author. Distinction is based on fundamental cellular and biochemical features and so is independent of the form of the plant. Each division can contain a range of forms, from unicellular to complex multicellular, although in many divisions the unicellular and colonial forms predominate. General features used to distinguish the divisions are:

• • • • •

The range of accessory photosynthetic pigments present in addition to chlorophyll a (other chlorophylls, carotenoids, and biloproteins); The secondary more soluble components of the cell wall present in addition to the main fibrillar component; The chemical nature of the insoluble storage products resulting from excess photosynthesis; The presence, fine structure, number, and position of flagella on vegetative cells of flagellate organisms or on flagellate reproductive bodies of larger species; Specialized aspects of cell structure, peculiar to particular divisions, such as the silica frustules, which encase diatom cells.

The three biochemical features above are relatively uniform in the higher plants, compared with the algae, and similar to those of one algal division, Chlorophyta (green algae). This suggested origin of land plants occurred from only this one division, although this is now contested. At a fundamental level the algae are therefore much more diverse than the higher plants. Characteristics for each division can be found in the Further Reading list. The seaweeds are the macroscopic marine algae in the divisions Chlorophyta (green), Phaeophyta (brown) and Rhodophyta (red algae). Greens are mainly foliose and filamentous seaweeds. Browns have no unicellular forms and include the very large complex and leathery forms such as kelps and rockweeds. Reds include a wide range of heterotrichous forms forming a wide diversity of complex foliose and filamentous forms.


428

PHYTOBENTHOS

Seaweed Life Cycles Most seaweeds have more than one phase in their life cycle. They have generally the same pattern as higher plants where a sexually reproducing gametophyte generation gives rise to an asexually reproducing sporophyte generation and vice versa. There is not always an obligate alternation as in higher plants and there is much more diversity in the nature of the different phases in algae, with some red algae even having a third generation. Some only have one generation. In some cases this may reflect lack of experimental culture which is necessary for life cycle determination. The simplest life cycle with two phases is termed isomorphic where the gametophyte and sporophyte are morphologically identical. Many common green seaweeds are like this, e.g., Ulva and Enteromorpha. Life cycles with morphologically different phases are heteromorphic. An example is kelp plants which have a massive leathery sporophyte which alternates with a microscopic filamentous gametophyte. In many cases the two generations in a heteromorphic life cycle may have been known since the nineteenth century by separate names from before the life cycle was determined in culture. For example, the various species of the red seaweed, Porphyra, alternate with a filamentous shell-boring sporophyte formerly known as Conchocelis rosea. Life cycles can be under environmental control. In some Porphyra species the change between generations is controlled by daylength, bringing about an annual seasonal life cycle. In some simpler life cycles, individual generations may be able to propogate themselves so that the full life cycle is not seen. This can be environmentally controlled so that, for example, in Europe there is a change with latitude of the relative proportions of the sexual and asexual generations of the brown filamentous seaweed, Ectocarpus, connected with latitudinal variation of sea temperature.

The Validity of Laboratory Cultures of Phytobenthos Experimental culture is needed to determine life cycles but it is difficult to simulate all environmental conditions. Wave action and water flow are not usually simulated in the numerous batch culture dishes needed for replicated ecological experiments. Artificial light sources are unlike daylight in spectral composition and intensity, yet light quality may control photomorphogenesis in plants. It is therefore possible that laboratory culture could give false results. Two examples are given below.

The red seaweeds Asparagopsis armata and Falkenbergia rufulanosa, originally described as separate species, are phases in the same life cycle. During the twentieth century they have been spreading their geographical limit northwards in Europe from the Mediterranean to northern Scotland. Populations in Britain are rarely seen with reproductive organs but have a mode of attachment to the rock surface that suggests they were produced by vegetative reproduction. The two phases seem to have spread independently in Britain although in culture they could be made to participate in the same life cycle. This reflects a wider phenomenon in red seaweeds where by going north and south from the center of geographical distribution, reproductive potential declines. In ecological experiments aseptic conditions are often not used, to the surprise of microbiologists. This lack of sterility may be desirable. For example, the green foliose seaweed, Monostroma, develops abnormally as a filamentous form in aseptic culture because of the need for growth factors from contaminating bacteria. These examples are not meant to decry culture experiments but to counsel their critical interpretation.

Seaweed Ecology Seaweeds are present on all rocky shores but are more obvious where wave action is less. On temperate shores intertidal zones are generally dominated by brown fucoid seaweeds (wracks or rockweeds), while the shallow subtidal area is occupied by a kelp forest formed of large laminarian seaweeds. A variety of red, green, and brown seaweeds forms an understorey. Some general rules can be exemplified by consideration of shores in north-west Europe. Firstly, on intertidal rocky shores the number of species is greatest at the lower tidal levels and declines with increasing intertidal height. Shores sheltered from wave action show the greatest cover and biomass of seaweeds. Shores exposed to very strong wave action tend to be dominated by sessile animals rather than seaweeds. Most shores are intermediate in wave action and tend to have a mosaic distribution of organisms, at least on lower and mid-shore. This may include patches of grazing animals interspersed with patches of different seaweed communities. The mosaic may make it hard to see a zonation of organisms with height on the shore. On very sheltered shores there may be a very obvious zonation of large brown seaweeds, in order of descending height on the shore: Pelvetia canalicaulata, Fucus spiralis,


PHYTOBENTHOS

Fucus vesiculosus and Ascophyllum nodosum, Fucus serratus, Laminaria digitata. (Similar zonations, but with different species, may occur on temperate shores outside north-west Europe.) Clear-cut zonations with visually dominant species can give a false impression that discrete communities exist at different tidal heights. Understorey species also have zones but their boundaries do not coincide with the larger species so as not to give sharply delimited communities. With increase in wave action there is a change of species, e.g., in fucoids Fucus serratus is replaced by Himanthalia elongata and in laminarians Laminaria digitata is replaced by Alaria esculenta. Some species change form with wave action, for example, the bladder wrack, F. vesiculosus, loses its bladders (such morphological plasticity is a common feature confusing seaweed identification). With increased wave action, zones increase in breadth and height on shore. All these features can be incorporated in biologically defined exposure scales for the comparative description of rocky shores, which place shores on a numerical exposure scale. This facilitates the comparison of similar shores in pollution-monitoring studies to ensure that differences along a pollution gradient are due to human disturbance rather than to wave action. Rock pools interrupt the gradient of conditions with height on shore. They provide a constantly submerged environment, like the subtidal one, but which is of limited volume and so undergoes physicochemical fluctuations while the tide is out, unlike the open sea. There is a corresponding zonation of dominant seaweed types in rock pools. On the lower shore they are characterized by sublittoral species and on the mid-shore they include species restricted to pools such as Halidrys siliquosa. Upper shore pools, where salinity fluctuates, have few species and are characterized by euryhaline opportunists such as Enteromorpha spp. Intertidal zonation is only partly due to the desiccation and salinity tolerance of the seaweeds. Such factor tolerance is particularly important on the upper shore where conditions are most harsh for a marine organism and so few species are present, with few biotic interactions. On mid- and lower-shores biotic factors are important in determining species boundaries. Grazing by limpets and periwinkles on smaller algae, and on the microscopic germlings of larger ones, is important as is biological competition, which narrows down species occurrence to less than their tolerance range. Subtidally a zonation may also be seen with dense kelp forest, with a large variety of understorey

429

species in the shallowest water, below which is a kelp park with only scattered plants. Below the kelp depth limit may be a red algal zone to the photic depth limit where light becomes insufficient for positive net photosynthesis. Important factors are again both physical and biotic. Light tolerance plays a role but in the shallowest waters, where plant density is greatest, competition is important and zone limits may be set by grazers, this time by sea urchins rather than limpets. There is an old view that accessory pigment differences between green, brown, and red seaweeds equip them to dominate at different depths according to which spectral quality of light penetrates. This is an oversimplification. Deeper-growing plants are shade plants with lower overall light requirements and can be of any color group. Distribution into estuaries along a generally decreasing salinity gradient is another modifying factor like wave action. Colonization of hard surfaces by phytobenthos in estuaries is largely by marine species with species number declining going upstream. This occurs by selective attenuation firstly of red, then of brown species. Estuaries have broadly two algal zones: an outer one with fucoid dominated shores, which are a species-poor version of a sheltered open coast shore; and an inner zone dominated by filamentous mat-forming algae, principally greens and blue–greens. There can be a successional sequence in which a bare area of shore is successively colonized, starting with unicells, with increasingly larger and more complex algae. Patches in a mosaic distribution may be at different stages in such a sequence. The succession involves contrasting types of seaweed as shown in Table 1. Although opportunists are good at colonizing bare rock, in a stable environment they are eventually replaced by more precisely adapted late successional species. Various conditions may favor the unusual abundance of opportunists. Mistaken conclusions about effects of effluent discharges can be reached by environmentalists who do not realize that there are both natural and artificial causes. Opportunist domination of rocks is favored naturally by sand scour and they may have natural summer outbursts in temperate climates. Pollution, particularly by sewage-derived nutrients, may favor opportunist domination. On tidal flats, ‘green tides’ may occur where the mudflat is completely covered by thick opportunist mats. This can induce anoxia and ammonia release beneath the mat so inhibiting benthic invertebrate populations and interfering with bird-feeding on tidal flats.


430

PHYTOBENTHOS

Table 1

Types of seaweed involved in successional colonization

Morphology Size Growth rate Life span Reproduction Environmental tolerance

Opportunist species early in succession, e.g., foliose and filamantous green algae

Late successional species, e.g., large cartilaginous plants such as fucoids

Simple Smaller Faster Shorter All year round Very wide but not precise adaptation to a specific niche

Complex Larger Slower Longer Likely to have seasons Narrower but good adaptation to a specific niche

Uses of Seaweed Benthic macroalgae are an important biological resource. Their principal uses are human food, animal feeds, fertilizer, and industrial chemicals. In the West, human seaweed consumption is small compared with the east. One of the most valuable fisheries listed by the Food and Agriculture Organization (FAO) is a seaweed, Porphyra, used extensively for human consumption in Japan. In the West, human consumption is more a health food market with beneficial effects ascribed to high trace element content, because of the high bioaccumulation activity of many seaweeds, but this remains to be rigorously tested. Aqueous extracts of seaweeds such as fucoids and kelps are used as commercial and domestic fertilizers. Beneficial effects have been claimed in horticulture such as increased growth, faster ripening and increased fruit yield. Effects are usually ascribed to trace element or plant hormone (usually cytokinin) content. Various high value fine chemicals, such as pigments, can be obtained from seaweeds but the main seaweed chemical industry is extraction of phycocolloids. These are secondary cell wall components of red and brown seaweeds, which have gel-forming and emulsifying properties in aqueous solutions. This gives them hundreds of industrial applications ranging from textile printing to ice cream manufacture. The chemicals concerned are all macromolecular carbohydrates, principally alginates from brown seaweeds and agars and carrageenans from red seaweeds. Alginate is not a single substance but a biopolymer with considerable possibility for structural variation. It is a linear structure of repeating sugar units of two kinds, mannuronic acid (M units) and guluronic (G units). Different alginates vary in the ratio of M and G units and in chain length (total number of units). Structural differences confer different gel-forming and emulsifying properties making different alginates suitable for different industrial

applications. In turn, different alginate structures are found in different species so that different seaweeds are required for different applications.

Ensuring Supplies of Commercial Seaweeds Seaweeds can be harvested from natural populations or farmed in the sea. The approach may depend on the value of the product. In the West, harvesting is preferred for phycocolloids. Despite the wide industrial application they are low-value products and so will not support the high labor costs needed for cultivation. By contrast alginate is successfully produced from farmed kelps in China, where labor costs are lower, and Porphyra can be farmed for human consumption in Japan because of its high value. Sustainable harvesting should consider the ability of the seaweed resource to recover. Mechanized Macrocystis (giant kelp) harvesting in California is sustainable yet productive because of the growth pattern of the plant. It grows to almost 70 m long and cropping of the distal few meters by barges floating over the forest canopy allows numerous meristems to remain intact so growth continues. By contrast Laminaria hyperborea harvesting in Norway is more destructive because the desired alginate is in the stipe (the supporting ‘stem’) of the plant. Harvesting by kelp dredges, i.e., large bags with cutters at the front end towed over the seabed on skis just above the rock surface, cuts plants off just above the holdfast, so that the meristematic area is harvested. Sustainability of the Laminaria forest is based on the vegetation structure. There are different layers of plants. The dredge takes only the large canopy-forming plants with tough, nonyielding stipes. The smaller plants bend rather than breaking and so survive harvesting. With the absence of the canopy they receive more light and so grow quickly to replace the harvested plants. The forest biomass is regenerated in 3 years so the Norwegian


PHYTOBENTHOS

Government licenses areas of the seabed for harvesting not less than every 4 years. The regenerated kelp is better for alginate extraction as it is not contaminated by epiphytes. However, the sustainability of the kelp forest biodiversity is separate from commercial yield. It is a diverse ecosystem with many invertebrates in the kelp holdfasts, as well as diverse seabed fauna and flora, and pelagic species sheltered by the forest. The diversity of this had not recovered after much more than 4 years so the licensing system does not conserve the full ecosystem. There are biotic considerations in protecting the phytobenthic resource. In the early 1970s the decline of the sea otter population off California allowed its prey, a sea urchin which was a kelp grazer, to increase. This hindered natural regeneration of the forest, which was already declining due to sewage pollution from Los Angeles. Recovery required manual transplantation. Various species of sea urchin are among the most voracious subtidal grazers. Another one, Strongylocentrotus, was able to create barren areas of seabed off the Canadian coast where Laminaria longicruris had been harvested, thus preventing regeneration. Farming of seaweed can require knowledge from laboratory culture of environmental tolerances and the factors controlling life cycles. This allows manipulation of stocks in seawater tanks on land under controlled conditions to produce reproductive bodies. These can be used to seed ropes, nets, canes or other substrata which are then planted out into sheltered sea areas for growing on. This increases the habitat area for attachment in the sea, and so increases yield. In the case of Porphyra in Japan, such manipulation of environmental conditions allows up to five generations per year from an area of coast where natural seasonal changes would only give one. It is not practical to grow the plants entirely on land in environmentally controlled seawater tanks because of the high cost of the facilities, but it is feasible to retain a small stock of, for example, shells

431

infected with the Conchocelis-phase of Porphyra from which spores can be obtained on demand by manipulation of daylength and temperature. The technology exists to cultivate seaweeds entirely in land-based tanks should the economics be favorable in the future. For example, there are different genetic strains of Chondrus crispus, whose carrageenan yield and type can be controlled by nutrient and salinity conditions. Such tank culture may be useful in the future to produce high value fine chemicals for medical applications. Seaweeds are the most obvious type of plant in the sea and are the main component of the phytobenthos but other algae and other marine plants are described elsewhere in this Encyclopedia.

See also Coral Reefs. Eutrophication. Exotic Species, Introduction of. Phytoplankton Blooms. Primary Production Distribution. Primary Production Methods. Primary Production Processes. Rocky Shores.

Further Reading Guiry MD and Blunden G (eds.) (1991) Seaweed Resources in Europe: Uses and Potential. Chichester: Wiley. Hoek C, van den, Mann DG, and Jahns HM (1995) Algae: An Introduction to Phycology. Cambridge: Cambridge University Press. Lembi CA and Waaland JR (eds.) (1988) Algae and Human Affairs. Cambridge: Cambridge University Press. Lobban CS and Harrison PJ (1997) Seaweed Ecology and Physiology. Cambridge: Cambridge University Press. Luning K (1990) Seaweeds: Their Environment, Biogeography and Ecophysiology. New York: Wiley. South GR and Whittick A (1987) Introduction to Phycology. Oxford: Blackwell Scientific Publications.


PHYTOPLANKTON BLOOMS D. M. Anderson, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2179–2192, & 2001, Elsevier Ltd.

Introduction Among the thousands of species of microscopic algae at the base of the marine food chain are a few dozen that produce toxins. These species make their presence known in many ways, ranging from massive ‘red tides’ that discolor the water, to dilute, inconspicuous concentrations of cells noticed only because of the harm caused by their highly potent toxins. Impacts include mass mortalities of wild and farmed fish and shellfish, human illness and death, alterations of marine trophic structure, and death of marine mammals, sea birds, and other animals. ‘Blooms’ of these algae are commonly called red tides, since, in some cases, the tiny plants increase in abundance until they dominate the planktonic community and change the color of the water with their pigments (Figure 1). The term is misleading, however, since nontoxic species can bloom and harmlessly discolor the water, or can cause ecosystem damages as severe as those linked to toxic organisms. Adverse effects can also occur when toxic algal cell concentrations are low and the water is not discolored. Given the confusion surrounding the meaning of ‘red tide’, the scientific community now prefers the term ‘harmful algal bloom’ or HAB. This new descriptor includes algae that cause problems because of their toxicity, as well as nontoxic algae that cause problems in other ways. It also applies to macroalgae (seaweeds) which can cause major ecological impacts as well (Figure 2).

Impacts Toxic Algae

HAB phenomena take a variety of forms, with multiple impacts (Table 1). One major category of impact occurs when toxic phytoplankton are filtered from the water as food by shellfish which then accumulate the algal toxins to levels which can be lethal to humans or other consumers. The poisoning syndromes have been given the names paralytic, diarrhetic, neurotoxic, amnesic, and azaspiracid shellfish poisoning (PSP, DSP, NSP, ASP, and AZP, respectively). The symptomology and exposure route

432

for each of these are presented in Table 2. Except for ASP, all are caused by biotoxins synthesized by a class of marine algae called dinoflagellates. A fifth human illness, ciguatera fish poisoning (CFP) is caused by toxins produced by dinoflagellates that attach to surfaces in many coral reef communities. Ciguatoxins are transferred through the food chain from herbivorous reef fishes to larger carnivorous, commercially valuable finfish. The final human illness linked to toxic algae is called possible estuaryassociated syndrome (PEAS). This vague term reflects the poor state of knowledge of the human health effects of the dinoflagellate Pfiesteria piscicida and related organisms that have been linked to symptoms such as deficiencies in learning and memory, skin lesions, and acute respiratory and eye irritation – all after exposure to estuarine waters where Pfiesterialike organisms have been present. Another type of HAB impact occurs when marine fauna are killed by algal species that release toxins and other compounds into the water, or that kill without toxins by physically damaging gills or by creating low oxygen conditions as bloom biomass decays. Fish and shrimp mortalities from these types of HABs at aquaculture sites have increased considerably in recent years. HABs also cause mortalities of wild fish, sea birds, whales, dolphins, and other marine animals. To understand the breadth of these ecosystem impacts, think of the transfer of toxins through the food web as analogous to the flow of carbon. Just as phytoplankton are the source of fixed carbon that moves through the food web, they can also be the source of toxins which cause adverse effects either through toxin transmitted directly from the algae to the affected organism or indirectly through food web transfer. A prominent example of direct toxin transfer was the death of 19 whales in Massachusetts in 1987 due to saxitoxin that had accumulated in mackerel that the whales consumed. Similar events occurred in Monterey Bay, California in 1998 and again in 2000 when hundreds of sea lions died from domoic acid (the ASP toxin) vectored to them via anchovies. In both cases, the food fish accumulated algal toxins through the food web and passed those toxins to the marine mammals. Adult fish can be killed by the millions in a single outbreak, with obvious long- and short-term ecosystem impacts (Figure 3). Likewise, larval or juvenile stages of fish or other commercially important fisheries species can experience mortalities from algal


PHYTOPLANKTON BLOOMS

433

Figure 1 Dead fish and discolored water during a Florida red tide. (Photo credit: Florida Department of Environmental Protection.)

Figure 2 Seaweed washed up on a Florida beach. Macroalgal blooms, although not toxic, can be quite harmful to coastal ecosystems and cause major economic problems to the tourist industry. (Photo credit: B. Lapointe).

toxins. Impacts of this type are far more difficult to detect than the acute poisonings of humans or higher predators, since exposures and mortalities are subtle and often unnoticed. Impacts might not be apparent until years after a toxic outbreak, such as when a year class of commercial fish reaches harvesting age but is in low abundance. Chronic toxin exposure may therefore have long-term consequences that are critical with respect to the sustainability or recovery of natural populations at higher trophic levels. Many

believe that ecosystem-level effects from toxic algae are more pervasive than are realized, affecting multiple trophic levels, depending on the ecosystem and the toxin involved. Nontoxic Blooms

Nontoxic blooms of algae can cause harm in a variety of ways. One prominent mechanism relates to the high biomass that some blooms achieve. When


434


Table 1

Some harmful effects caused by algae in coastal and brackish waters

Effect

Causative organisms

Human health syndrome Paralytic shellfish poisoning (PSP)

Diarrhetic shellfish poisoning (DSP) Neurotoxic shellfish poisoning (NSP) Azaspiracid shellfish poisoning (AZP) Amnesic shellfish poisoning (ASP) Ciguatera fish poisoning (CFP) Respiratory problems and skin irritation Possible estuary-associated syndrome (PEAS) Hepatotoxicity Mortality of wild and cultured marine resources Hemolytic, hepatotoxic, osmoregulatory effects and other unspecified toxicity

Mechanical damage to gills Gill clogging and necrosis Inhibition of tourism and recreational activities Production of foams, mucilage, discoloration, repellent odors

Alexandrium spp., Pyrodinium bahamense var. compressum, Gymnodinium catenatum, Anabena circinalis, Aphanizomenon spp. Dinophysis spp., Prorocentrum spp. Gymnodinium breve Protoperidinium crassipes Pseudo-nitzschia spp. Gambierdiscus toxicus Gymnodinium breve, Pfiesteria piscicida, Nodularia spumigena Pfiesteria piscicida, P. shumwayae, and possibly other Pfiesteria-like organisms Microcystis aeruginosa, Nodularia spumigena

Gymnodinium spp., Cochlodinium polykrikoides, Pfiesteria piscicida, Heterosigma carterae, Chattonella spp., Fibrocapsa japonica, Chrysochromulina spp., Prymnesium spp., Aureococcus anophagefferens, Microcystis spp. Chaetoceros spp., Leptocylindrus spp. Phaeocystis spp., Thalassiosira spp.

Seaweed accumulation on beaches; coral reef overgrowth

Noctiluca scintillans, Phaeocystis spp., Cylindrotheca closterium, Aureococcus anophagefferens, Nodularia spumigena Cladophera, Codium, Ulva, and many other species

Adverse effects on marine ecosystems Hypoxia, anoxia Negative effects on feeding behavior, shading, reduction of water clarity Seaweed accumulation on beaches; coral reef overgrowth Toxicity to marine organisms, including invertebrates, fish, mammals, and birds

Noctiluca scintillans, Ceratium spp., many others Aureococcus anophagefferens, Aureoumbra lagunensis, Phaeocystis spp. Cladophera, Codium, Ulva, and many other species Gymnodinium breve, Alexandrium spp., Pseudo-nitzschia australis, Cocchlodinium polykrikoides

(Modified from Zingone and Enevoldsen, 2000.)

this biomass begins to decay as the bloom terminates, oxygen is consumed, leading to widespread mortalities of all plants and animals in the affected area. These ‘high biomass’ blooms are sometimes linked to excessive pollution inputs (see below), but can also occur in relatively pristine waters. Large, prolonged blooms of nontoxic algal species can reduce light penetration to the bottom, decreasing densities of submerged aquatic vegetation (SAV). Loss of SAV can have dramatic impacts on coastal ecosystems, as these grass beds serve as nurseries for the food and the young of commercially important fish and shellfish. Such indirect ecosystem effects from a nontoxic HAB were seen with a dense brown tide that lasted for over 7 years in the Laguna Madre section of southern Texas. The dense, coffeecolored bloom reduced light penetration and dramatically altered the abundance of seagrasses over a wide area.

Macroalgae (seaweeds) also cause problems. Over the past several decades, blooms of macroalgae have been increasing along many of the world’s developed coastlines. Macroalgal blooms occur in nutrient-enriched estuaries and nearshore areas that are shallow enough for light to penetrate to the seafloor. These blooms have a broad range of ecological effects, and often last longer than ‘typical’ phytoplankton HABs. Once established, macroalgal blooms can remain in an environment for years unless the nutrient supply decreases. They can be particularly harmful to coral reefs. Under high nutrient conditions, opportunistic macroalgal species outcompete, overgrow, and replace the coral. Nontoxic phytoplankton can also kill fish. The diatom genus Chaetoceros includes several species which have been associated with mortalities of farmed salmon, yet no toxin has ever been identified in that group. These species have long, barbed spines



Table 2

435

Human illnesses associated with harmful algal blooms

Syndrome

Causative organisms

Toxin produced

Route of acquisition

Clinical manifestations

Ciguatera fish poisoning (CFP)

Gambierdiscus toxicus and others

Ciguatoxin, maitotoxin

Acute gastroenteritis, paresthesias and other neurological symptoms

Paralytic shellfish poisoning (PSP)

Alexandrium species, Gymnodinium catenatum, Pyrodinium bahamense var. compressum, and others Gymnodinium breve and others

Saxitoxins

Toxin passed up marine food chain; illness results from eating large, carnivorous reef fish Eating shellfish harvested from affected areas

Neurotoxic shellfish poisoning (NSP)

Brevetoxins

Diarrhetic shellfish poisoning (DSP)

Dinophysis species

Okadaic acid and others

Azaspiracid shellfish poisoning (AZP)

Protoperidinium crassipes

Azaspiracids

Amnesic shellfish poisoning (ASP)

Pseudo-nitzchia species

Domoic acid

Possible estuaryassociated syndrome (PEAS)

Pfiesteria piscicida and others

Unidentified

Eating shellfish harvested from affected areas; toxins may be aerosolized by wave action Eating shellfish harvested from affected areas Eating shellfish harvested from affected areas Eating shellfish (or, possibly, fish) harvested from affected areas Exposure to water or aerosols containing toxins

Acute paresthesias and other neurological manifestations; may progress rapidly to respiratory paralysis and death

Gastrointestinal and neurological symptoms; respiratory and eye irritation with aerosols Acute gastroenteritis

Neurotoxic effects with severe damage to the intestine, spleen, and liver tissues in test animal Gastroenteritis, neurological manifestations, leading in severe cases to amnesia, coma, and death Deficiencies in learning and memory; acute respiratory and eye irritation, acute confusional syndrome

(Modified from Morris, 1999.)

Figure 3 Toxic blooms can kill wild and aquacultured fish by the millions. (Photo credits: G. Pitcher and M. Aramaki).


436


that lodge in gill tissues, leading to massive discharge of mucus that eventually results in lamellar degeneration and death due to the reduction in oxygen exchange. Why these diatoms have barbed spines is unknown. It is improbable that they have evolved specifically to kill fish, since the only mortalities from these species are caged fish that cannot escape the blooms. The problems now faced by the fish-farming industry are most likely the unfortunate result of an evolutionary strategy by certain Chaetoceros species to avoid predation or to stay afloat.

The Toxins The toxins produced by some HAB species are among the most potent natural poisons known. Saxitoxin, for example, is 1000 times more potent than cyanide, and 50 times stronger than curare. Many of the toxin classes are not single chemical entities, but instead represent families of compounds of similar chemical structure (Table 2). These toxins bind with high affinity to specific receptor sites such as ion channels of excitable cells (Table 3). Most binding is reversible, but dissociation times can be quite prolonged. Each derivative of the parent toxin is slightly different in structure, and thus binds to the target receptor with different affinity. Weaker toxins tend to dissociate more quickly, accounting for their reduced potency. The saxitoxin family binds to sodium channels, which are the protein ‘tunnels’ responsible for the flux of sodium in and out of nerve and muscle cells. The brevetoxins also bind to the sodium channel, but to a different site. Their effect is to cause continuous activation of the channel, rather than a blockage, as with saxitoxin. Domoic acid

Table 3

disrupts normal neurochemical transmission in the brain by binding to kainate receptors in the central nervous system. This results in sustained depolarization of the neurons, and eventually in cell degeneration and death. One unfortunate symptom in ASP victims is permanent, short-term memory loss due to lesions that form in portions of the brain such as the hippocampus where there is a high density of kainate receptors. Man is exposed to algal toxins principally by consumption of contaminated seafood products, although one type of toxin (brevetoxin) also causes respiratory asthma-like symptoms because of aerosol formation due to wave action. Acute single-dose lethality of seafood toxins has been extensively studied, but chronic and/or repeated exposure to low toxin concentrations, which undoubtedly occurs, has not been adequately examined. There is also a serious lack of knowledge as to how the toxins are distributed throughout the body and eliminated. Other important questions include how long the toxins circulate before elimination, and how they are metabolized by living organisms. These knowledge gaps prevent researchers from devising antidotes or effective treatments which may alleviate or lessen the symptoms. Therapeutic intervention is primarily limited to symptomatic treatment and life support. The mechanisms by which fish are killed by toxic HABs are not well understood. Neurotoxins can rapidly act on specific fish tissues, significantly reducing heart rate and resulting in reduced blood flow and death from lack of oxygen. Other classes of compounds released by HABs such as reactive oxygen species, polyunsaturated fatty acids, and galactolipids are not toxins per se, but still can kill fish and other animals rapidly. Some have hemolytic activity,

Properties of major algal toxins (values in parentheses (n) indicate the number of toxin derivatives)

Toxin family (n)

Syndrome

Solubility

Action on

Pharmacology

Saxitoxin (18) Brevetoxin (9)

PSP NSP

Water soluble Lipid soluble

Azaspiracid (3)

AZP

Lipid soluble

Blockage of sodium channel Selective sodium channel activators Not yet characterized

Okadaic acid (12)

DSP

Lipid soluble

Nerve, brain Nerve, muscle, lung, brain Nerve, liver, intestine, spleen Enzymes

Domoic acid (11)

ASP

Hot acid extraction

Brain

Maitotoxin/ ciguatoxin

CFP

Water/lipid soluble

Nerve, muscle, heart, brain


Diarrhea by stimulating phosphorylation of a protein that controls sodium excretion of intestinal cells Binds to glutamate receptors in the brain, causing continuous stimulation of brain cells; lesions formed Activation of calcium/sodium channels


and others cause gill rupture, excessive production of mucus, hypoxia, and edema. Scientists are searching for biochemical pathways within the algae to explain the metabolic role of the HAB toxins, but the search thus far has been fruitless. The toxins are not proteins, and all are synthesized in a series of chemical steps requiring multiple genes. Biosynthetic pathways have been proposed, but no chemical intermediates have been identified, nor have unique enzymes used only in toxin production yet been isolated. All we have are tantalizing clues that relate to toxin variability in individual cells, such as a 10-fold increase in saxitoxin production in cultured Alexandrium cells that are limited by phosphorus rather than nitrogen, or changes in the relative abundance of the different toxin derivatives produced by a particular dinoflagellate strain when growth conditions are varied. These observations indicate that the metabolism of the toxins is a dynamic process within the algae, but it is still not clear whether they have a specific biochemical role or are simply secondary metabolites. As with the nontoxic, spiny Chaetoceros species that kill fish, the illnesses and mortalities caused by algal ‘toxins’ may simply be the result of the ‘accidental’ affinity of those compounds for receptor sites on ion channels or tissues humans and in higher animals.

Ecology and Population Dynamics Growth Features, Bloom Mechanisms

Impacts from HABs are necessarily linked to the population size and distribution of the causative algae. The growth and accumulation of individual algal species in a mixed assemblage of marine organisms are exceedingly complex processes involving an array of chemical, physical, and biological interactions. Given the diverse array of algae from several different classes that produce toxins or cause problems in a variety of oceanographic systems, attempts to generalize the bloom dynamics of HAB species are doomed to failure. Some common mechanisms can nevertheless be highlighted. HABs are seldom caused by the explosive growth of a single species that rapidly dominates the water column. As stated by Ryther and co-workers many years ago, ‘ythere is no necessity to postulate obscure factors which would account for a prodigious growth of dinoflagellates to explain red water. It is necessary only to have conditions favoring the growth and dominance of a moderately large population of a given species, and the proper hydrographic and meteorological conditions to permit the accumulation of organisms at the surface and to

437

effect their future concentrations in localized areas.’ In other words, winds, tides, currents, fronts, and other features can create discrete patches of cells of streaks of red water at all scales. HABs in temperate and high latitudes are predominantly summer, coastal phenomena. They commonly occur during periods when heating or freshwater runoff create a stratified surface layer overlying colder, nutrient rich waters. Since the upper layer is quickly stripped of nutrients by other fast-growing algae, the onset of stratification often means that the only significant source of major plant nutrients lies below the interface between the layers – the pycnocline. This situation favors dinoflagellates and other motile algae, since nonmotile phytoplankton are unable to remain in suspension in the upper layer, and thus sink out of the zone where light permits photosynthesis. In contrast, motile algae are able to regulate their position and access the nutrients below the pycnocline. Many HAB species can swim at speeds in excess of 10 m d 1, and some undergo marked vertical migration, in which they reside in surface waters during the day to harvest the sunlight and then swim to the pycnocline and below to take up nutrients at night. This strategy can explain how dense accumulations of cells can appear in surface waters that are devoid of nutrients and which would seem to be incapable of supporting such prolific growth. Swimming can also allow a species to persist at some optimum depth in the presence of vertical currents. One striking observation about some HAB blooms is that highly concentrated subsurface layers of cells can sometimes be only tens of centimeters thick – so called ‘thin layers’ of cells. This is another example of the importance of physical processes in determining organism distributions in stratified coastal systems. Figure 4 shows a vertical profile of temperature and the corresponding thin layer of dinoflagellates that formed under those conditions along the coast of France. Similar thin layers are found throughout the world with scales as small as 10 cm in the vertical and 10 km in the horizontal. Horizontal transport of blooms is also an important feature of some HABs, often over large distances. Major toxic outbreaks can suddenly appear at a site due to the transport of established blooms from other areas by ocean currents. This is the case in the western Gulf of Maine in the USA, where blooms of toxic dinoflagellates are transported hundreds of kilometers within a buoyant coastal current formed by the outflow of a major river system. An NSP outbreak in North Carolina in 1987 is another example of such long distance transport. The toxic Gymnodinium breve population that contaminated


438


North Carolina shellfish that year and drove tourists and residents from beaches originated in a bloom off the south-west coast of Florida, nearly 1500 km away. That bloom was carried out of the Gulf of Mexico and up the south-east coast of the USA by a combination of current systems, culminating in the 0

Depth (m)

10 20 30 40 12 14 Temperature

16

0

50 Dinoflagellates

100

Figure 4 Subsurface accumulation of ‘thin layers’ of HAB species are common, as shown in this dataset from La Rochelle, France. (Adapted with permission from Gentien et al. (1995)).

Gulf Stream. After approximately 30 days of transport, a filament of water separated from the Gulf Stream and moved onto the narrow continental shelf of North Carolina, carrying toxic G. breve cells with it. The warm water mass remained in nearshore waters and was identifiable in satellite images for nearly 3 weeks (Figure 5). Another important aspect of physical/biological coupling relates to the enhanced phytoplankton biomass that can occur at hydrographic features such as fronts. This enhanced biomass is the result of the interaction between physical processes such as upwelling, shear, and turbulence, and physiological processes such as swimming and enhanced nutrient uptake. One example is the linkage between tidally generated fronts and the sites of massive blooms of the toxic dinoflagellate Gyrodinium aureolum (now called Gymnodinium mikimotoi) in the North Sea. The pattern generally seen is a high surface concentration of cells at the frontal convergence, contiguous with a subsurface chlorophyll maximum which

Figure 5 Satellite image of sea surface temperature, showing the warm Gulf Stream (blue) off the coast of North Carolina, A filament of the Gulf Stream can be seen extending towards shore. This carried a toxic population of Gymnodinium breve; that population originated in the Gulf of Mexico, >1500 km away. (Photo credit: P. Tester.).



Vertical section

439

Surface expression

Figure 6 Accumulation of HAB cells near a front. This schematic demonstrates how cells can accumulate at a frontal convergence, with a surface manifestation of the bloom at the frontal convergence, and a subsurface extension of the bloom that extends along the sloping pycnocline. (Adapted with permission from Franks (1992).)

follows the sloping interface between the two water masses beneath the stratified side of the front (Figure 6). The surface signature of the chlorophyll maximum (sometimes a visible red tide) may be 1– 30 km wide. Chlorophyll concentrations are generally lower and uniform on the well-mixed side of the front. Lateral transport of the front and its associated cells brings toxic G. aureolum populations into contact with nearshore fish and other susceptible resources, resulting in massive mortalities. This is a case where small-scale physical/biological coupling results in biomass accumulation, and larger-scale advective mechanisms cause the biomass to become harmful. Life Histories

An important aspect of many HABs is their reliance on life history transformations for bloom initiation and decline. A number of key HAB species have dormant, cyst stages in their life histories, including Alexandrium spp., Pyrodinium bahamense, Gymnodinium catenatum, Cocchlodinium polykrikoides, Gonyaulax monilata, Pfiesteria piscicida, Chattonella spp., and Heterosigma carterae. Resting cysts remain in bottom sediments (sometimes termed ‘seedbeds’) when conditions in the overlying waters are unsuitable for growth. When conditions improve, such as with seasonal warming, the cysts germinate, inoculating the water column with a population of cells that begins to divide asexually via binary fission to produce a bloom. At the end of the bloom, often in response to nutrient limitation, vegetative growth ceases and the cells undergo sexual reproduction, whereby gametes are formed that fuse to form the swimming zygotes that ultimately become dormant cysts. Figure 7 shows an example of the Alexandrium

tamarense life history. Clearly, the location of cyst seedbeds can be an important determinant of the location of resulting blooms, and the size of the cyst accumulations can affect the magnitude of the blooms as well. It is generally believed, however, that environmental regulation of cell division is more important to eventual bloom magnitude than the size of the germination inoculum from cysts. Cysts are also important in species dispersal. Natural or human-assisted transport of cysts from one location to another (e.g. via ballast water discharge or shellfish seeding) can allow a species to colonize a region and extend its geographic range. In 1972, a hurricane introduced Alexandrium fundyense into southern New England waters from established populations in the Bay of Fundy. Since that time, PSP has become an annually recurrent phenomenon in the region. Another example of human-assisted species introductions is the appearance of Gymnodinium catenatum in Tasmania in the 1970s, coincident with the development of a wood chip industry involving commercial vessels and frequent ballast water discharge. Grazing Interactions

HAB cells can be food for herbivorous marine organisms (grazers). The extent to which grazing can control HABs depends upon the abundance of zooplankton, their ability to ingest the harmful algal species, and the effects of the HAB species on the grazers. In this context, one of the least understood aspects of toxic algal physiology concerns the metabolic or ecological roles of the toxins. Some argue that toxins evolved as a defense mechanism against grazing. Indeed, experimental studies demonstrate that zooplankton are affected to some degree by the


440


1 2

9

8 3A

4A 7

5 6 30m Figure 7 Life cycle diagram of Alexandrium tamarense. Stages are identified as follows: (1) vegetative, motile cell; (2) temporary or pellicle cyst; (3) ‘female’ and ‘male’ gametes; (4) fusing gametes; (5) swimming zygote or planozygote; (6) resting cyst or hypnozygote; (7,8) motile, germinated cell or planomeiocyte; and (9) pair of vegetative cells following division. Redrawn from Anderson et al. 1998.).

algal toxins they ingest. Two different effects have been noted. The first is a gradual incapacitation during feeding, as if the zooplankton are gradually paralyzed or otherwise impaired. (One study even showed that a tintinnid ciliate could only swim backwards, away from its intended algal prey, after exposure to a toxic dinoflagellate culture.) The second response is an active rejection or regurgitation of the toxic algae by a grazing animal, as if it had an unpleasant taste. Either of these mechanisms would result in reduced grazing pressure on toxic forms, which would facilitate the formation of a bloom. However, this cannot be the sole explanation for the presence of the toxins, as nontoxic phytoplankton are abundant, and form massive blooms. An example of these mechanisms is seen in Figure 8, which depicts the feeding behavior of the copepod Acartia tonsa in mixtures containing equal concentrations of toxic and nontoxic Alexandrium spp. dinoflagellates. The copepods consumed nontoxic Alexandrium sp. cells at significantly higher

rates than toxic cells, indicating chemosensory discrimination of toxic cells by the grazers. Although there is clear evidence of the avoidance of certain HABs by grazers, or of incapacitation of the grazers themselves during feeding, it is also apparent that some toxins can be vectored through the food web via grazers. Fish kills of herring in the Bay of Fundy have been linked to PSP toxins from Alexandrium species that had been grazed by pteropods (small planktonic snails that are a favorite food of herring). Fortunately, in terms of human health, subsequent laboratory studies found that herring and other fish are very sensitive to these toxins and, unlike shellfish, die before they accumulate the toxins at dangerous levels in their flesh. Through similar food web transfer events, Alexandrium toxins have been implicated in kills of menhaden, sand lance, bluefish, dogfish, skates, monkfish, and even whales. Similar arguments can be made for other HAB toxins, especially the brevetoxins which have been linked to deaths of dolphins, and domoic acid (the ASP toxin),



_1

_1

Ingestion rate (ng C copepod h )

200

150

Alexandrium tamarense (nontoxic)

100

50

0

Alexandrium fundyense (toxic) Figure 8 Feeding behavior of the copepod Acartia tonsa in mixtures containing equal concentrations of toxic and non-toxic Alexandrium spp. dinoflagellates. Central bars denote mean rates, boxes indicate standard deviation. Redrawn with permission from Teegarden 1999.

which has caused extensive bird and sea lion mortalities along the eastern Pacific coast of the USA and Mexico.

Trends The nature of the global HAB problem has changed considerably over the last several decades (Figure 9). Virtually every coastal country is now threatened by harmful or toxic algal species, often in many locations and over broad areas. Thirty years ago, the problem was much more scattered and sporadic. The number of toxic blooms, the economic losses from them, the types of resources affected, and the number of toxins and toxic species have all increased dramatically in recent years. Disagreement only arises with respect to the reasons for this expansion, of which there are several possibilities. New bloom events may simply reflect indigenous populations that are discovered because of improved chemical detection methods and more observers. The discovery of ASP along the west coast of the USA in 1991 is a good example of this, as toxic diatom species were identified and their toxin detected as a direct result of communication with Canadian scientists who had discovered the same toxin 4 years earlier and developed new chemical detection methods for domoic acid. Several other ‘spreading events’ are most easily attributed to natural dispersal via currents, rather than

441

human activities. As described above, the first NSP event ever to occur in North Carolina was shown to be a Florida bloom transported over 1500 km by the Gulf Stream – a totally natural phenomenon with no linkage to human activities. Some believe that humans may have contributed to the global HAB expansion by transporting toxic species in ship ballast water (e.g. Gymnodinium catenatum in Tasmania) or by shellfish relays and seeding. Another factor underlying the global increase in HABs is that we have dramatically increased aquaculture activities worldwide (Figure 10), and this inevitably leads to increased monitoring of product quality and safety, revealing indigenous toxic algae that were probably always there. Nutrient enrichment is another cause for the increasing frequency of HAB events worldwide. Manipulation of coastal watersheds for agriculture, industry, housing, and recreation has drastically increased nutrient loadings to coastal waters. Just as the application of fertilizer to lawns can enhance grass growth, marine algae grow in response to nutrient inputs. Shallow and restricted coastal waters that are poorly flushed appear to be most susceptible to nutrient-related algal problems. Nutrient enrichment of such systems often leads to excessive production of organic matter, a process known as eutrophication, and increased frequencies and magnitudes of phytoplankton blooms, including HABs. There is no doubt that this is true in certain areas of the world where pollution has increased dramatically. It is perhaps real, but less evident in areas where coastal pollution is more gradual and unobtrusive. A frequently cited dataset from an area where pollution has been a significant factor in HAB incidence is from the Inland Sea of Japan, where visible red tides increased steadily from 44 per year in 1965 to over 300 a decade later, matching the pattern of increased nutrient loading from pollution (Figure 11). Effluent controls were instituted in the mid-1970s, resulting in a 70% reduction in the number of red tides that has persisted to this day. A related data set for the Black Sea documents a dramatic increase in red tides up to the mid-1990s, when the blooms began to decline. That reduction, which has also continued to this day, has been linked to reductions in fertilizer usage in upstream watersheds by former Soviet Union countries no longer able to afford large, state-subsidized fertilizer applications to agricultural land. Anthropogenic nutrients can stimulate or enhance the impact of toxic or harmful species in several ways. At the simplest level, toxic phytoplankton may increase in abundance due to nutrient enrichment, but remain as the same relative fraction of the total phytoplankton biomass (i.e. all phytoplankton


442


1970

2000

Figure 9 Expansion of the PSP problem over the past 30 years. Sites with proven records of PSP-causing organisms are noted in 1970, and again in 2000. Modified from Hallegraeff 1993. , PSP.

species are affected equally by the enrichment). Alternatively, some contend that there has been a selective stimulation of HAB species by pollution. This view is based on the ‘nutrient ratio hypothesis’ which argues that environmental selection of phytoplankton species is associated with the relative availability of specific nutrients in coastal waters, and that human activities have altered these nutrient supply ratios in ways that favor harmful forms. For example, diatoms, the vast majority of which are

harmless, require silicon in their cell walls, whereas most other phytoplankton do not. Since silicon is not abundant in sewage effluent but nitrogen and phosphorus are, the N:Si or P:Si ratios in coastal waters have increased through time over the last several decades. Diatom growth in these waters will cease when silicon supplies are depleted, but other phytoplankton classes (including toxic or harmful species) can continue to proliferate using the ‘excess’ nitrogen and phosphorus.


50

Molar total N/P ratio

45

60

40

50

35

40

30 30

25 20

20

Redfield ratio

10

15 Figure 10 One reason for the global expansion in HABs relates to the increase in aquaculture activities, as shown with the dense concentration of mussel rafts in Spain. Not only do these intense farming operations alter the nutrient loads and plankton composition of local waters, but the intense scrutiny of the product for quality control purposes can reveal cryptic toxins that may have always been present in the region. (Photograph courtesy of D.M. Anderson.).

12

Industrial products (10 yen)

Industrial production Seto Inland Sea Law (LSIS; 1973) 80

300

60 200 40

Red tide number

100

20

0 1950

Number of red tides

400

100

1960

1970

1980

1990

0 2000

Figure 11 Human pollution can directly affect the number of red tides and HABs. This figure shows how blooms increased dramatically during the 1960s and 1970s in the Seto Inland Sea of Japan, but declined after pollution control legislation was enacted. Redrawn from Okaichi 1998.

This concept is controversial, but is supported by a 23-year time series off the German coast which documents the general enrichment of coastal waters with nitrogen and phosphorus, as well as a fourfold increase in N:Si and P:Si ratios. This was accompanied by a striking change in the composition of the phytoplankton community, as diatoms decreased and flagellates increased more than 10-fold. Similar arguments hold for changes in N:P ratios, which have also changed due to the differences in removal of these nutrients with standard wastewater treatment practices. Some HAB species, such as Phaeocystis pouchetii, the species responsible for massive blooms that foul beaches with thick layers of smelly foam and slime, are not good competitors for phosphorus. As the N:P ratio of Dutch coastal waters decreased over time due to pollution effects, the size

0

10 1973

1978

1983

443

Phaeocystis summer bloom duration (in days)


1988

Year Figure 12 Changes in the duration of Phaeocystis blooms as the N:P ratio of Dutch coastal waters declined. Redrawn from Riegman et al. 1992.

and duration of Phaeocystis blooms increased (Figure 12). The global increase in HAB phenomena is thus a reflection of two factors – an actual expansion of the problem, and an increased awareness of its size or scale. It is expanding due to pollution or other global change issues, but improved methods and enhanced scientific inquiry have also led to a better appreciation of its size or scale. The fact that part of the expansion is a result of increased awareness should not negate our concern, nor should it alter the manner in which we mobilize resources to attack it. The fact that it is also growing due to human activities makes our concerns even more urgent.

Management Issues Management options for dealing with the impacts of HABs include reducing their incidence and extent (prevention), stopping or containing blooms (control), and minimizing impacts (mitigation). Where possible, it is preferable to prevent HABs rather than to treat their symptoms. Since increased pollution and nutrient loading may cause an increased incidence of outbreaks of some HAB species, these events may be prevented by reducing pollution inputs to coastal waters, particularly industrial, agricultural, and domestic effluents high in plant nutrients. This is especially important in shallow, poorly flushed coastal waters that are most susceptible to nutrient-related algal problems. Other strategies that may prevent HAB events include: regulating the siting of aquaculture facilities to avoid areas where HAB species are present, modifying water circulation for those HABs where restricted water exchange is a factor in bloom development, and restricting species introductions (e.g. through regulations on ballast


444


water discharges or shellfish and finfish transfers for aquaculture). Potential approaches to control HABs are similar to those used to control pests on land – e.g. biological, physical, or chemical treatments that directly target the bloom cells. However, more research is needed before these means are used to control HABs in natural waters. The most effective mitigation tools are monitoring programs that detect toxins in shellfish and/or monitor the environment for evidence of HAB events. These programs can provide advance warnings of outbreaks and/or indicate areas that should be closed to harvesting. Recent technological advances, such as remote-sensing and molecular techniques, have increased detection and characterization of HAB species and blooms, and are playing an increasing role in monitoring programs worldwide. A long-term goal of these HAB monitoring programs and tools is to develop the ability to forecast bloom development and movement.

Summary The HAB problem is significant and growing worldwide and poses a major threat to public health and ecosystem health, as well as to fisheries and economic development. The problems and impacts are diverse, as are the causes and underlying mechanisms controlling the blooms. The signs are clear that pollution and other alterations in the coastal zone have increased the abundance of algae, including harmful and toxic forms. All new outbreaks and new problems cannot be blamed on pollution, however, as there are numerous other explanations, some of which involve human activities, and some of which do not. As a growing world population increases its use of the coastal zone and demands more fisheries and recreational resources, there is a clear need to understand HAB phenomena and to develop scientifically sound management and mitigation policies.

Further Reading Anderson DM (1994) Red tides. Scientific American 271: 52--58. Anderson DM, Cembella AD, and Hallegraeff GM (eds.) (1998) The Physiological Ecology of Harmful Algal Blooms. Heidelberg: Springer-Verlag. Burkholder JM and Glasgow HB Jr (1997) Pfiesteria piscicida and other Pfiesteria-like dinoflagellates: Behavior, impacts, and environmental controls. Limnology and Oceanography 42: 1052--1075. Carmichael WW (1997) The cyanotoxins. Advances in Botanical Research 27: 211--256. Fogg GE (1991) The phytoplanktonic ways of life. New Phytologist 118: 191--232. Franks PJS (1992) Phytoplankton blooms at fronts: patterns, scales, and physical forcing mechanisms. Reviews in Aquatic Sciences 6(2): 121--137. Gentien P, Lunven M, Lehaitre M, and Duvent JL (1995) In situ depth profiling of particle sizes. Deep-Sea Research 42: 1297--1312. Hallegraeff GM (1993) A review of harmful algal blooms and their apparent global increase. Phycologia 32: 79--99. Morris JG (1999) Pfiesteria, ‘The Cell from Hell’ and other toxic algal nightmares. Clinical Infectious Diseases 28: 1191--1198. Smayda TJ (1989) Primary production and the global epidemic of phytoplankton blooms in the sea: a linkage In: Cosper EM, Carpenter EJ, and Bricelj VM (eds.) Novel Phytoplankton Blooms: Causes and Impacts of Recurrent Brown Tide and Other Unusual Blooms. Berlin, Heidelberg, New York: Springer-Verlag. Teegarden GJ (1999) Copepod grazing selection and particle discrimination on the basis of PSP toxin content. Marine Ecology Progress Series 181: 163--176. Yasumoto T and Murata M (1993) Marine toxins. Chemical Reviews 93: 1897--1909. Zingone A and Enevoldsen HO (2000) The diversity of harmful algal blooms: A challenge for science and management. Ocean and Coastal Management 43: 725--748.

See also Molluskan Fisheries. Network Analysis of Food Webs. Plankton. Primary Production Methods.


PHYTOPLANKTON SIZE STRUCTURE E. Maranõń, University of Vigo, Vigo, Spain & 2009 Elsevier Ltd. All rights reserved.

Introduction Phytoplankton are unicellular organisms that drift with the currents, carry out oxygenic photosynthesis, and live in the upper illuminated waters of all aquatic ecosystems. There are approximately 25 000 known species of phytoplankton, including eubacterial and eukaryotic species belonging to eight phyla. This phylogenetically diverse group of organisms constitutes the base of the food chain in most marine ecosystems and, since their origin more than 2.8 109 years ago, have exerted a profound influence on the biogeochemistry of Earth. Currently, phytoplankton are responsible for the photosynthetic fixation of around 50 1015 g C annually, which represents almost half of global net primary production on Earth. Some 20% of phytoplankton net primary production is exported toward the ocean’s interior, either in the form of sinking particles or as dissolved material. The mineralization of this organic matter gives way to an increase with depth in the concentration of dissolved inorganic carbon. The net effect of this phytoplankton-fueled, biological pump is the transport of CO2 from the atmosphere to the deep ocean, where it is sequestered over the timescales of deep-ocean circulation (102–103 years). A small fraction (o1%) of the organic matter transported toward the deep ocean escapes mineralization and is buried in the ocean sediments, where it is retained over timescales of 4106 years. It has been calculated that thanks to the biological pump the atmospheric concentration of CO2 is maintained 300–400 ppm below the levels that would occur in the absence of marine primary production. Thus, phytoplankton play a role in the regulation of the atmospheric content of CO2 and therefore affect climate variability. The cell size of phytoplankton ranges widely over at least 9 orders of magnitude, from a cell volume around 0.1 mm3 for the smallest cyanobacteria to more than 108 mm3 for the largest diatoms. Cell size affects many aspects of phytoplankton physiology and ecology over several levels of organization, including individuals, populations, and communities. Phytoplankton assemblages are locally diverse: typically, several hundreds of species can be found in just a liter of seawater. Therefore, a full determination of

the biological properties of all species living in a given water body is not possible. The study of cell size represents an integrative approach to describe the structure and function of the phytoplankton community and to understand its role within the pelagic ecosystem and the marine biogeochemical cycles. This article starts with a review of the relationship between cell size and phytoplankton metabolism and growth. It follows with a description of the general patterns of variability in the size structure of phytoplankton in the ocean. Next, the different mechanisms involved in bringing about these patterns are examined. The article ends with a consideration of the ecological and biogeochemical implications of phytoplankton size structure.

Phytoplankton Cell Size, Metabolism, and Growth Cell Size and Resource Acquisition

Like any other photoautotrophic organisms, phytoplankton must take up inorganic nutrients and absorb light in order to synthesize new organic matter. Both nutrient uptake and light absorption are heavily dependent on cell size. The supply of nutrients to the cell may become diffusion-limited when nutrient concentrations are low, if the rate of nutrient uptake exceeds the rate of molecular diffusion and a nutrient-depleted area develops around the cell. Assuming a spherical cell shape and applying Fick’s first law of diffusion, the uptake rate (U, mol s 1) can be expressed as U ¼ 4prDDC

½1

where r is the cell radius (mm), D is the diffusion coefficient (mm2 s 1), and DC (mol mm3) is the nutrient concentration gradient between the cell’s surface and the surrounding medium. The specific uptake rate (uptake per unit of cell volume) will then be U=V ¼ 4prDDCð4=3 pr3 Þ1 ¼ 3DDCr2

½2

Equation [2] indicates that specific uptake rate decreases with the square of cell radius. However, specific (i.e., normalized to mass or volume) metabolic rates in phytoplankton, and therefore specific resource requirements, decrease with cell size much more slowly. Typically, specific metabolic rates are proportional to cell mass or volume elevated to a power between 1/3 and 0 (see below), or to


445

446

PHYTOPLANKTON SIZE STRUCTURE

r elevated to a power between 1 and 0. Therefore, large cell size is a major handicap when nutrient concentrations are low. The nutrient concentration below which phytoplankton growth starts to be nutrient-limited can be calculated, for a range of cell sizes and growth rates, as follows. Nutrient limitation occurs when Uom Q, where m is the specific growth rate (s 1) and Q is the cell’s quota for the particular nutrient (mol cell 1). Q can be computed by applying an empirical relationship between cellular nitrogen content and V (pgN ¼ 0.017 2 V1.023) and D is assumed to be 1.5 cm2 s 1. The threshold for nutrient limitation increases exponentially with cell radius (Figure 1). In this particular example, if the ambient nitrogen concentration is 10 nM, a cell of r ¼ 1 mm could grow at growth rates well above 1 d 1 without suffering diffusion limitation, whereas a cell of r ¼ 6 mm would already experience diffusion limitation at m ¼ 0.1 d 1. Processes such as turbulence, sinking, and swimming, which enhance the advective transport of nutrients toward the cell surface, partially compensate for the negative impact of large cell size on diffusive nutrient fluxes. The effect, however, is small and does not alter the fact that larger cells are at a disadvantage over smaller cells for nutrient uptake. Light absorption in phytoplankton is a function of cell size and the composition and concentration of pigments. The amount of light absorbed per unit pigment decreases with cell size and with intracellular pigment concentration, because the degree of

self-shading among the different pigment molecules (the so-called package effect) increases. The optical absorption cross-section of a phytoplankton cell is given by a ¼ ð3=2Þðas QrÞ

½3

Q ¼ 1 þ 2ðer =rÞ þ 2ðer 1Þ=ðr2 Þ

½4

r ¼ as ci d

½5

where

and

Units of a and as are m2 (mg chl a) 1, Q and r are coefficients without dimensions, ci is the intracellular chlorophyll a concentration (mg chl a m 3), and d is cell diameter (m). The package effect can be quantified as the ratio between the actual absorption inside the cell (a ) and the maximum absorption possible, determined for the pigment in solution (as ). The higher the package effect, the lower the value of a =as . The data shown in Figure 2, calculated assuming as ¼ 0.04 m2 (mg chl a) 1 and a range of ci values between 106 and 108 mg chl a m 3, indicate that the package effect increases with cell size, which means that, other things being equal, larger phytoplankton are expected to be less efficient than smaller cells at absorbing light. This effect is much stronger under low

100 1 2

80

60

0.8 106

0.5 0.6 a* (a*s)−1

DIN (nM)

1

40

0.4

0.1

20

107

0.2 108 0 0

2

4

6

8

10

0 0

Cell radius (µm) Figure 1 Relationship between cell radius and the concentration of dissolved inorganic nitrogen (DIN) below which phytoplankton growing at rates of 0.1, 0.5, 1, and 2 d 1 begin to suffer diffusion limitation of growth.

10 20 Cell radius (µm)

30

Figure 2 Relationship between cell radius and the package effect for intracellular chlorophyll a concentrations of 106, 107, and 108 mg m 3.



R ¼ aW b

½6

which is equivalent to log R ¼ log a þ b log W

6 High light cultures slope = 0.73

4 log10 photosynthesis (pgC cell−1 d−1)

light conditions, when ci tends to be larger as a result of photoacclimation. The general allometric theory predicts a reduction in metabolic rates (R) with body size (W, in units of biomass or volume). From unicells to large mammals, individual metabolic rates (R) scale as:

2

0 Low light cultures slope = 0.59

½7

where b, the size-scaling exponent, usually takes a value of 3/4 and a, the intercept of the log–log relationship, is a taxon-related constant. When biomassspecific metabolism or growth rates are considered, b takes a value near 1/4. Equations [6] and [7] mean that a 10-fold increase in cell size is associated with only a 7.5-fold increase in individual metabolic rate, and therefore larger organisms have a slower metabolism. In the case of phytoplankton, one would expect that the geometrical constraints on resource acquisition and use would result in a reduction in specific metabolism with cell size. However, several experimental studies with laboratory cultures and natural phytoplankton assemblages suggest that the relationship between photosynthesis and cell size cannot be predicted by a single scaling model and that phytoplankton metabolism often departs from the 3/4-power rule. For instance, the size scaling of photosynthesis in cultured diatoms has been shown to depend on light availability (Figure 3). Light limitation leads not only to low photosynthetic rates (irrespective of cell size) but also to a reduction in the size scaling exponent b, indicating a faster decrease in specific photosynthesis with cell size as a result of a stronger package effect in larger cells. In addition, changes in taxonomic affiliation of phytoplankton species along the size spectrum may interfere with the effects of cell size per se. If we consider natural phytoplankton assemblages, which include a mixture of species from diverse taxonomic groups, the scaling between phytoplankton photosynthesis is approximately isometric (Figure 3). This means that, under natural conditions at sea, the expected slowdown of metabolism as cell size increases does not seem to occur. This pattern results from the fact that larger species possess strategies that allow them to sustain higher metabolic rates than expected for their size (see below). Thus, both resource availability and taxonomic variation along the size spectrum help explain why the scaling of phytoplankton growth rates and cell size is quite variable. However, before turning our attention to the size scaling of

447

−2 Natural phytoplankton slope = 1.03 −4 −2

0

2 4 log10 cell volume (µm3)

6

8

Figure 3 Scaling relationship between photosynthesis per cell and cell size in diatom cultures growing under high-light conditions, diatom cultures growing under light-limited conditions, and natural phytoplankton assemblages in the ocean.

phytoplankton growth rates, we need to look into the relationship between cell size and loss rates, because the net growth rate of any population ultimately depends on the balance between production and loss processes. Cell Size and Loss Processes

Respiration and exudation are the main metabolic loss processes for phytoplankton. In most organisms, individual respiration rates increase as the 3/4-power of body size (e.g., b ¼ 3/4). However, several studies of the size scaling of phytoplankton respiration in algal cultures show that, although taxon-related differences exist, b tends to be significantly higher than 3/4, indicating that respiration increases with cell size more steeply than predicted by the general allometric theory. This effect has been attributed to the fact that smaller algae seem to have lower respiration rates than expected for their size, perhaps as an energy-saving strategy in organisms with a comparatively small capacity for accumulation of reserves. As far as exudation is concerned, theory predicts that smaller cells, on account of their higher surface-to-volume ratios, should suffer a relatively greater loss of cellular compounds through the cell membrane. However, both laboratory work with cultures and experimental studies at sea have failed to provide conclusive evidence as to the existence of higher relative rates of exudation in smaller cells as compared to larger cells.


448


One of the most unquestionable effects of large cell size for phytoplankton is that it increases sinking velocity, which, for nonmotile cells, implies a reduction of their residence time in the euphotic layer. According to Stokes’ law, the sinking velocity of a spherical particle increases in proportion to the square of its radius. Assuming an excess cell density over medium density of 50 g l 1, a picophytoplankton cell of r ¼ 0.5 mm will have a sinking velocity of only 2–3 mm day 1, compared with 20–30 m day 1 for a microphytoplankton cell of r ¼ 50 mm. Although many large phytoplankton species have acquired different strategies to cope with sinking (such as motility, buoyancy control, departure from spherical shape, etc.), smaller species are clearly at an advantage over their larger counterparts to remain within the euphotic zone. This advantage is particularly relevant in strongly stratified water columns, where the absence of upward water motion makes it unlikely for cells sinking below the pycnocline to return to the upper, well-illuminated waters. While small cell size is superior in terms of resource acquisition and avoidance of sedimentation, large cell size can provide a major competitive advantage because it offers a refuge from predation. The main reason is that the generation time of predators increases with body size more rapidly than the generation time of phytoplankton. Small phytoplankton are typically consumed by unicellular protist herbivores, which have generation times similar to those of phytoplankton (in the order of hours to days). In contrast, larger phytoplankton are mostly grazed by metazoan herbivores, such as copepods and euphasiaceans, which have much longer generation times (in the order of weeks to months). When nutrients are injected into the euphotic layer, smaller phytoplankton are efficiently controlled by their predators and their abundance seldom increases substantially. On the contrary, larger phytoplankton, thanks to the time lag between their growth response and the numerical response of their predators, are able to form blooms and carry on growing until nutrients are exhausted. As discussed below, this trophic mechanism plays a role in determining the size structure of phytoplankton communities in contrasting marine environments. Cell Size and Growth Rates

Experiments with cultured and natural phytoplankton species have yielded quite variable values for the scaling exponent in the equation relating growth rate (units of time 1) to cell size. The most frequently reported values for b range between 0.1 and 0.3. It has been noted that the size scaling of

phytoplankton growth rates is relatively weak (that is, b tends to take a less negative value than 1/4). Furthermore, although the slope of the growth versus size relationship may be similar in different taxonomic groups, the intercept frequently is not: for instance, diatoms consistently have higher growth rates than other species of the same cell size. Another feature in the size scaling of phytoplankton growth is that very small cells (less than 5 mm in diameter) depart from the inverse relationship between cell size and growth rate, showing slower rates than expected for their size. This pattern probably results from the influence of nonscalable components (such as the genome and the membranes), which progressively take up more space as cell size decreases, thus leaving less cell volume available for rate-limiting catalysts and the accumulation of reserves. Although experimental determinations in natural conditions are still scarce, the available data suggest that phytoplankton populations in nature tend to exhibit less negative size scaling exponents in the power relation between growth rate and cell size. In fact, there is evidence to suggest that this size-scaling exponent may even become positive under favorable conditions for growth, such as high nutrient and light availability. This means that, when resources are plentiful, larger phytoplankton may grow faster than their smaller relatives. Several strategies allow larger species, and diatoms in particular, to achieve high growth rates (e.g., 41 day 1) in nature in spite of the geometrical constraints imposed by their size. These strategies include the increase in the effective surfaceto-volume ratio, due to changes in cell shape and the presence of the vacuole; the accumulation of nonlimiting substrates to increase cell size and optimize nutrient uptake; and the ability to sustain high specific uptake rates and store large amounts of reserves under conditions of discontinuous nutrient supply.

Patterns of Phytoplankton Size Structure in the Ocean Size-Fractionated Chlorophyll a

Given that chlorophyll a (chl a) serves as a proxy for phytoplankton biomass, a common approach to study the relative importance of phytoplankton with different cell sizes is to measure the amount of chl a in size classes, usually characterized in terms of equivalent spherical diameter (ESD) of particles. The most frequently considered classes are the picophytoplankton (cells smaller than 2 mm in ESD), the nanophytoplankton (cells with an ESD between 2 and 20 mm), and the microphytoplankton (cells with an ESD larger than 20 mm). When we plot together hundreds of



become more important during periods of prolonged stratification and low nutrient availability, as well as during conditions of intense vertical mixing that lead to light limitation of growth.

2 Microphytoplankton 1

449

Nanophytoplankton

log10 chl a (mg m−3)

Picophytoplankton

Size–Abundance Spectra 0

−1

−2

−3 −1

0 log10 total chl a (mg m−3)

1

Figure 4 Chlorophyll a concentration in picophytoplankton, nanophytoplankton, and microphytoplankton vs. total chlorophyll a concentration in samples obtained throughout the euphotic layer in coastal and oceanic waters of widely varying productivity.

measurements of size-fractionated chl a concentration, obtained throughout the euphotic layer in coastal and oceanic waters of widely varying productivity, several consistent patterns emerge (Figure 4). In relatively poor waters, where total chl a concentrations are below 0.8–1 mg m 3, picophytoplankton account for up to 80% of total chl a, while microphytoplankton typically contributes less than 10%. As total chl a increases, the concentration of picophytoplankton chl a reaches a plateau at around 0.5 mg m 3 and then decreases in very rich waters. Similarly, nanophytoplankton chl a rarely increases beyond 1 mg m 3. By contrast, microphytoplankton chl a continues to increase, so that at total chl a levels above 2 mg m 3 this size class accounts for more than 80% of total chl a, while picophytoplankton contribute less than 10%. Compared to the other size fractions, nanophytoplankton show smaller variability in their relative contribution to total chl a, which normally falls within the range 20–30%. The patterns shown in Figure 4 reflect both temporal and spatial variability in total phytoplankton biomass and the relative importance of each size class. Thus, the oligotrophic waters of the subtropical gyres are typically dominated by picophytoplankton, whereas in upwelling areas and coastal, well-mixed waters microphytoplankton usually account for most of the photosynthetic biomass. Similarly, in ecosystems that experience marked seasonal variability, microphytoplankton dominate the episodes of intense algal growth and biomass, such as the spring bloom. Small nano- and picophytoplankton

Although the partition into discrete classes is useful to describe broadly the size structure of the community, the different phytoplankton species are in fact characterized by a continuum of cell sizes that is best represented with a size–abundance spectrum. In this approach, the abundance (N) and cell size (W) of all species present in a sample are determined, using flow cytometry for picophytoplankton and small nanoplankton and optical microscopy for large nanoplankton and microphytoplankton. Size–abundance spectra are constructed by distributing the abundance data along an octave scale of cell volume. The abundance of all cells within each size interval is summed and the resulting abundance is plotted on a log-log scale against the nominal size of the interval. When these spectra are constructed at the local scale, it is common that the relationship between abundance and cell size shows irregularities, for example, departures from linearity. This is the case of large blooms, when one or a few species make up a major fraction of total phytoplankton abundance, which translates into a bump in the size–abundance spectrum. However, when numerous observations, collected over longer spatial and temporal scales, are put together, good linear relationships are usually obtained. In these cases, the slope of the linear relationship between log N and log W (the size-scaling exponent in the power relationship between N and W) is a general descriptor of the relative importance of small versus large cells in the ecosystem. Figure 5 illustrates the differences in the phytoplankton size–abundance spectrum between two contrasting ecosystems such as the oligotrophic subtropical gyres of the Atlantic Ocean (NpW 1.25) and the productive waters of the coastal upwelling region in the NW Iberian peninsula (NpW 0.90). Typically, the slopes of the size–abundance spectrum in oligotrophic, open ocean waters are in the range 1.1 to 1.4, while values between 0.6 and 0.9 are measured in coastal productive environments. The size spectrum extends further to the left in oligotrophic ecosystems, reflecting the presence of the prochlorophyte Prochlorococcus. At 0.1–0.2 mm3 in volume, this species is the smallest and most abundant photoautotrophic organism on Earth, and dominates picophytoplankton biomass in the oligotrophic open ocean. When combined with size-scaling relationships for biomass and metabolism, size–abundance spectra allow us to determine the variability in the flow of


450


6 Subtropical gyres slope = −1.25

4

log10 cell abundance (cell ml−1)

2 0 −2 (a) Coastal upwelling slope = −0.90

4 2 0 −2 −4

(b) −1

0

1 2 3 4 log10 cell volume (µm3 cell−1)

5

6

Figure 5 Size–abundance spectra in natural phytoplankton assemblages from (a) the Atlantic subtropical gyres and (b) coastal upwelling waters off NW Iberian peninsula. Spectra were constructed by combining measurements collected throughout the euphotic layer during different sampling surveys. Reprinted with permission from Maranõń E, Cermen+ o P, Rodrı´guez J, Zubkov MV, and Harris RP (2007) Scaling of phytoplankton photosynthesis and cellsize in the ocean. Limnology and oceanography 54(5): 2194. Copyright (2008) by the Americal Society of Limnology and Ocenography, Inc.

materials and energy along the size spectrum. For instance, if phytoplankton photosynthesis per cell (P) in the ocean scales isometrically with cell size (PpW1), the above-mentioned scaling exponents for phytoplankton abundance (between 0.6 and 0.9 for eutrophic waters and between 1.1 and 1.4 for oligotrophic ones) imply that total photosynthesis per unit volume (N P) must increase with cell size in productive ecosystems, while it decreases in unproductive ones.

Factors Controlling Phytoplankton Size Structure Given their superior ability to avoid diffusion limitation of nutrient uptake, it is no surprise that picophytoplankton dominate in the oligotrophic waters of the open ocean. However, the competitive advantage of being small, although reduced, is not eliminated under resource sufficient conditions. One can therefore ask why is it that picophytoplankton do not dominate also in nutrient-rich environments. This is tantamount to asking what mechanisms are

responsible for the increased importance of larger cells under conditions of high nutrient availability, such as those found within upwelling regions, frontal regions, and cyclonic eddies. Hydrodynamic processes have been suggested to play a role, since upwelling water motion causes a retention of larger cells, counteracting their tendency to sink out of the euphotic layer. However, upwelling of subsurface waters also contributes to the injection of nutrients into surface waters and therefore the physical transport mechanism cannot be easily distinguished from a direct nutrient effect. These two processes, however, have been separated experimentally during iron addition experiments in high-nutrient, lowchlorophyll regions. Invariably, iron addition brings about an increase in phytoplankton biomass and a marked shift toward a dominance by larger species, usually chain-forming diatoms. Since hydrodynamics remain unaltered during these experiments, these observations highlight the role of the nutrient field in determining phytoplankton size structure. Two main, nonexclusive processes contribute to the selective growth and accumulation of larger cells in high-light and nutrient-rich environments. First, larger phytoplankton are capable of sustaining higher biomassspecific metabolic rates and growth rates than smaller cells when resources are abundant. Second, larger phytoplankton are less efficiently controlled by grazing, as a result of the difference between their generation time and that of their predators.

Ecological and Biogeochemical Implications of Phytoplankton Size Structure Phytoplankton size structure affects significantly the trophic organization of the planktonic ecosystem and, therefore, the efficiency of the biological pump in transporting atmospheric CO2 toward deep waters (Table 1). In communities dominated by picophytoplankton, where resource limitation leads to low phytoplankton biomass and production, the dominant trophic pathway is the microbial food web. Given that the growth rates of picophytoplankton and their protist microbial grazers (dinoflagellates, ciliates, and heterotrophic nanoflagellates) are similar, trophic coupling between production and grazing is tight, most of phytoplankton daily primary production is consumed within the microbial community, and the standing stock of photosynthetic biomass is relatively constant. Phytoplankton exudation and microzooplankton excretion contribute to an important production of dissolved organic matter, which fuels bacterial production. In turn,



Table 1 General ecological and biogeochemical properties of plankton communities in which phytoplankton are dominated by small vs. large cells Phytoplankton dominated by

Small cells

Large cells

Total phytoplankton biomass Total primary production Dominant trophic pathway Main loss process for phytoplankton

Low

High

Low

High

Microbial food web Grazing by protists

Classic food chain

Photosynthesis-torespiration ratio f-ratio and e-ratio Main fate of primary production

B1 5–15% Recycling within the euphotic layer

Sedimentation and grazing by metazoans 41 440% Export toward deep waters

bacteria are efficiently controlled by protist microbial grazers. The resulting, complex food web is characterized by intense recycling of matter and low efficiency in the transfer of primary production toward larger organisms such as mesozooplankton or fish. Photosynthetic production of organic matter is balanced by the respiratory losses with the microbial community. In addition, the small size of microbial plankton implies that losses through sedimentation are unimportant. As a result, little newly produced organic matter escapes the euphotic later. In contrast, plankton communities dominated by large phytoplankton, such as chain-forming diatoms, are characterized by enhanced sinking rates and simpler trophic pathways, where phytoplankton are grazed directly by mesozooplankton (the so-called classic food chain). Phytoplankton photosynthesis exceeds community respiration, leaving an excess of organic matter available for export. Thus, a major fraction of phytoplankton production is eventually transported toward deep waters, either directly through sinking of ungrazed cells, or indirectly through sedimentation of packaged materials such as aggregates and zooplankton fecal pellets. It must be noted that the microbial trophic pathway is always present in all planktonic communities, but its relative importance decreases in productive waters because of the addition of the classic food chain. The ecological properties outlined above for phytoplankton assemblages dominated by small versus large cells dictate the biogeochemical functioning of the biological pump in contrasting marine environments. In stable, oligotrophic ecosystems, where small photoautotrophs dominate, primary production sustained

451

by nutrients coming from outside the euphotic layer (new production) is small, and so is the ratio between new production and total production (the f-ratio), as well as the ratio between exported production and total production (the e-ratio). Typical values of the f- and e-ratios in these systems are in the range 5–15%. Given that, in the long term, only new production has the potential to contribute to the transport of biogenic carbon toward the deep ocean, the biological pump in these systems has a low efficiency and the net effect of the biota on the ocean–atmosphere CO2 exchange is small. By contrast, phytoplankton assemblages dominated by larger cells are typical of dynamic environments that are subject to perturbations leading to enhanced resource supply. In these systems, production and consumption of organic matter are decoupled, new and export production are relatively high (e- and f-ratios 440%), and the biological pump effectively transports biogenic carbon toward the ocean’s interior, thus contributing to CO2 sequestration.

Nomenclature a as ci d D N P Q r R U V W m

absorption coefficient of pigments in vivo absorption coefficient of pigments in solution intracellular chlorophyll a concentration cell diameter nutrient diffusion coefficient cell abundance photosynthesis per cell cellular nutrient quota cell radius metabolic rate nutrient uptake rate cell volume cell size (volume or weight) growth rate

See also Carbon Cycle. Microbial Loops. Phytoplankton Blooms. Plankton. Primary Production Distribution. Primary Production Methods. Primary Production Processes.

Further Reading Agawin NSR, Duarte CM, and Agustı´ S (2000) Nutrient and temperature control of the contribution of picoplankton to phytoplankton biomass and production. Limnology and Oceanography 45: 591--600.


452


Banse K (1976) Rates of growth, respiration and photosynthesis of unicellular algae as related to cell size – a review. Journal of Phycology 12: 135--140. Blasco D, Packard TT, and Garfield PC (1982) Size dependence of growth rate, respiratory electron transport system activity, and chemical composition in marine diatoms in the laboratory. Journal of Phycology 18: 58--63. Brown JH, Gillooly JF, Allen AP, Savage VM, and West GB (2004) Toward a metabolic theory of ecology. Ecology 85: 1771--1789. Cermenõ P, Maranõń E, Rodrı´guez J, and Fernańdez E (2005) Large-sized phytoplankton sustain higher carbon-specific photosynthesis than smaller cells in a coastal eutrophic ecosystem. Marine Ecology Progress Series 297: 51--60. Chisholm SW (1992) Phytoplankton size. In: Falkowski PG and Woodhead AD (eds.) Primary Productivity and Biogeochemical Cycles in the Sea, pp. 213--237. New York: Plenum. Falkowski PG, Laws EA, Barber RT, and Murray JW (2003) Phytoplankton and their role in primary, new, and export production. In: Fasham MJR (ed.) Ocean Biogeochemistry – The Role of the Ocean Carbon Cycle on Global Change, pp. 99--122. Berlin: Springer. Finkel ZV (2001) Light absorption and size scaling of lightlimited metabolism in marine diatoms. Limnology and Oceanography 46: 86--94. Kiørboe T (1993) Turbulence, phytoplankton cell size, and the structure of pelagic food webs. Advances in Marine Biology 29: 1--72.

Legendre L and Rassoulzadegan F (1996) Food-web mediated export of biogenic carbon in oceans. Marine Ecology Progress Series 145: 179--193. Li WKW (2002) Macroecological patterns of phytoplankton in the north western North Atlantic Ocean. Nature 419: 154--157. Maranõń E, Holligan PM, Barciela R, et al. (2001) Patterns of phytoplankton size-structure and productivity in contrasting open ocean environments. Marine Ecology Progress Series 216: 43--56. Maranõ´ E, Cermen+ o P, Rodrı´guez J, Zubkov MV, and Harris RP (2007) Scaling of phytoplankton photosynthesis and cellsize in the ocean. Limnology and oceanography 54(5): 2194. Platt T, Lewis M, and Geider R (1984) Thermodynamics of the pelagic ecosystem: Elemental closure conditions for biological production in the open ocean. In: Fasham MJR (ed.) Flows of Energy and Material in Marine Ecosystems, pp. 49--84. New York: Plenum. Raven JA (1998) Small is beautiful: The picophytoplankton. Functional Ecology 12: 503--513. Rodrı´guez J, Tintore´ J, Allen JT, et al. (2001) Mesoscale vertical motion and the size structure of phytoplankton in the ocean. Nature 410: 360--363. Sheldon RW, Prakash A, and Sutcliffe WH (1972) The size distribution of particles in the ocean. Limnology and Oceanography 17: 327--341. Tang EPY (1995) The allometry of algal growth rates. Journal of Plankton Research 17: 1325--1335.


PLANKTON M. M. Mullinw, Scripps Institution of Oceanography, La Jolla, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2192–2194, & 2001, Elsevier Ltd.

The category of marine life known as plankton represents the first step in the food web of the ocean (and of large bodies of fresh water), and components of the plankton are food for many of the fish harvested by humans and for the baleen whales. The plankton play a major role in cycling of chemical elements in the ocean, and thereby also affect the chemical composition of sea water and air (through exchange of gases between the sea and the overlying atmosphere). In the parts of the ocean where planktonic life is abundant, the mineral remains of members of the plankton are major contributors to deep-sea sediments, both affecting the chemistry of the sediments and providing a micropaleontological record of great value in reconstructing the earth’s history. ‘Plankton’ refers to ‘drifting’, and describes organisms living in the water column (rather than on the bottom – the benthos) and too small and/or weak to move long distances independently of the ocean’s currents. However, the distinction between plankton and nekton (powerfully swimming animals) can be difficult to make, and is often based more on the traditional method of sampling than on the organisms themselves. Although horizontal movement of plankton at kilometer scales is passive, the metazoan zooplankton nearly all perform vertical migrations on scales of 10s to 100s of meters. This depth range can take them from the near surface lighted waters where the phytoplankton grow, to deeper, darker and usually colder environments. These migrations are generally diurnal, going deeper during the day, or seasonal, moving to deeper waters during the winter months to return to the surface around the time that phytoplankton production starts. The former pattern can serve various purposes: escaping visual predators and scanning the watercolumn for food. (It should be noted that predators such as pelagic fish also migrate diurnally.) Seasonal descent to greater depths is a common feature for several copepod species and may conserve energy at a time when food is scarce in the

w

Deceased.

upper layers. However, vertical migration has another role. Because of differences in current strength and direction between surface and deeper layers in the ocean, time spent in deeper water acts as a transport mechanism relative to the near surface layers. On a daily basis this process can take plankton into different food concentrations. Seasonally, this effective ‘migration’ can complete a spatial life cycle. The plankton can be subdivided along functional lines and in terms of size. The size category, picoplankton (0.2–2.0 mm), is approximately equivalent to the functional category, bacterioplankton; most phytoplankton (single-celled plants or colonies) and protozooplankton (single-celled animals) are nano- or microplankton (2.0–20 mm and 20–200 mm, respectively). The metazoan zooplankton (animals, the ‘insects of the sea’) includes large medusae and siphonophores several meters in length. Size is more important in oceanic than in terrestrial ecosystems because most of the plants are small (the floating seaweed, Sargassum, being the notable exception), predators generally ingest their prey whole (there is no hard surface on which to rest prey while dismembering it), and the early life stages of many types of zooplankton are approximately the same size as the larger types of phytoplankton. Therefore, while the dependence on light for photosynthesis is characteristic of the phytoplankton, the concepts of ‘herbivore’ and ‘carnivore’ can be ambiguous when applied to zooplankton, since potential plant and animal prey overlap in size and can be equivalent sources of food. Though rabbits do not eat baby foxes on land, analogous ontogenetic role-switching is very common in the plankton. Among the animals, holoplanktonic species are those that spend their entire life in the plankton, whereas many benthic invertebrates have meroplanktonic larvae that are temporarily part of the plankton. Larval fish are also a temporary part of the plankton, becoming part of the nekton as they grow. There are also terms or prefixes indicating special habitats, such as ‘neuston’ to describe zooplanktonic species whose distribution is restricted to within a few centimeters of the sea’s surface, or ‘abyssoplankton’ to describe animals living only in the deepest waters of the ocean. Groups of such species form communities (see below). Since the phytoplankton depend on sunlight for photosynthesis, this category of plankton occurs almost entirely from the surface to 50–200 m of the ocean – the euphotic depth (where light intensity is 0.1–1% of full surface sunlight). Nutrients such as


453

454

PLANKTON

nitrate and phosphate are incorporated into protoplasm in company with photosynthesis, and returned to dissolved form by excretion or remineralization of dead organic matter (particulate detritus). Since much of the latter process occurs after sinking of the detritus, uptake of nutrients and their regeneration are partially separated vertically. Where and when photosynthesis is proceeding actively and vertical mixing is not excessive, a near-surface layer of low nutrient concentrations is separated from a layer of abundant nutrients, some distance below the euphotic depth, by a nutricline (a layer in which nutrient concentrations increase rapidly with depth). Therefore, the spatial and temporal relations between the euphotic depth (dependent on light intensity at the surface and the turbidity of the water), the nutricline, and the pycnocline (a layer in which density increases rapidly with depth) are important determinants of the abundance and productivity of phytoplankton. Zooplankton is typically more concentrated within the euphotic zone than in deeper waters, but because of sinking of detritus and diel vertical migration of some species into and out of the euphotic zone, organic matter is supplied and various types of zooplankton (and bacterioplankton and nekton) can be found at all depths in the ocean. An exception is anoxic zones such as the deep waters of the Black Sea, although certainly types of bacterioplankton that use molecules other than oxygen for their metabolism are in fact concentrated there. Even though the distributions of planktonic species are dependent on currents, species are not uniformly distributed throughout the ocean. Species tend to be confined to particular large water masses, because of physiological constraints and inimical interactions with other species. Groups of species, from small invertebrates to active tuna, seem to ‘recognize’ the same boundaries in the oceans, in the sense that their patterns of distribution are similar. Such groups are called ‘assemblages’ (when emphasizing their statistical reality, occurring together more than expected by chance) or ‘communities’ (when emphasizing the functional relations between the members in food webs), though terms such as ‘biocoenoses’ can be found in older literature. Thus, one can identify ‘central water mass,’ ‘subantarctic,’ ‘equatorial,’ and ‘boreal’ assemblages associated with water masses defined by temperature and salinity; ‘neritic’ (i.e. nearshore) versus ‘oceanic’ assemblages with respect to depth of water over

which they occur, and ‘neustonic’ (i.e. air–sea interface), ‘epipelagic,’ ‘mesopelagic,’ ‘bathypelagic,’ and ‘abyssopelagic’ for assemblages distinguished by the depth at which they occur. Within many of these there may be seasonally distinguishable assemblages of organisms, especially those with life spans of less than one year. Regions which are boundaries between assemblages are sometimes called ecotones or transition zones; they generally contain a mixture of species from both sides, and (as in the transition zone between subpolar and central water mass assemblages) may also have an assemblage of species that occur only in the transition region. Despite the statistical association between assemblages and water masses or depth zones, it is far from clear that the factor that actually limits distribution is the temperature/salinity or depth that physically defines the water mass or zone. It is likely that a few important species have physiological limits confining them to a zone, and the other members of the assemblage are somehow linked to those species functionally, rather than being themselves physiologically constrained. Limits can be imposed on certain life stage, such as the epipelagic larvae of meso- or bathypelagic species, creating patterns that reflect the environment of the sensitive life stage rather than the adult. Conversely, meroplanktonic larvae, such as the phyllosome of spiny lobsters, can often be found far away from the shallow waters that are a suitable habitat for the adults.

See also Bacterioplankton. Continuous Plankton Recorders. Gelatinous Zooplankton. Phytoplankton Blooms. Protozoa, Planktonic Foraminifera.Small-Scale Physical Processes and Plankton Biology. Zooplankton Sampling with Nets and Trawls.

Further Reading Cushing DH (1995) Population Production and Regulation in the Sea. Cambridge: Cambridge University Press. Longhurst A (1998) Ecological Geography of the Sea. New York: Academic Press. Mullin MM (1993) Webs and Scales. Seattle: University of Washington Press.


PLANKTON AND CLIMATE A. J. Richardson, University of Queensland, St. Lucia, QLD, Australia & 2009 Elsevier Ltd. All rights reserved.

Introduction: The Global Importance of Plankton Unlike habitats on land that are dominated by massive immobile vegetation, the bulk of the ocean environment is far from the seafloor and replete with microscopic drifting primary producers. These are the phytoplankton, and they are grazed by microscopic animals known as zooplankton. The word ‘plankton’ derives from the Greek planktos meaning ‘to drift’ and although many of the phytoplankton (with the aid of flagella or cilia) and zooplankton swim, none can progress against currents. Most plankton are microscopic in size, but some such as jellyfish are up to 2 m in bell diameter and can weigh up to 200 kg. Plankton communities are highly diverse, containing organisms from almost all kingdoms and phyla. Similar to terrestrial plants, phytoplankton photosynthesize in the presence of sunlight, fixing CO2 and producing O2. This means that phytoplankton must live in the upper sunlit layer of the ocean and obtain sufficient nutrients in the form of nitrogen and phosphorus for growth. Each and every day, phytoplankton perform nearly half of the photosynthesis on Earth, fixing more than 100 million tons of carbon in the form of CO2 and producing half of the oxygen we breathe as a byproduct. Photosynthesis by phytoplankton directly and indirectly supports almost all marine life. Phytoplankton are a major food source for fish larvae, some small surface-dwelling fish such as sardine, and shoreline filter-feeders such as mussels and oysters. However, the major energy pathway to higher trophic levels is through zooplankton, the major grazers in the oceans. One zooplankton group, the copepods, is so numerous that they are the most abundant multicellular animals on Earth, outnumbering even insects by possibly 3 orders of magnitude. Zooplankton support the teeming multitudes higher up the food web: fish, seabirds, penguins, marine mammals, and turtles. Carcasses and fecal pellets of zooplankton and uneaten phytoplankton slowly yet consistently rain down on the cold dark seafloor, keeping alive the benthic (bottom-dwelling) communities of sponges, anemones, crabs, and fish.

Phytoplankton impact human health. Some species may become a problem for natural ecosystems and humans when they bloom in large numbers and produce toxins. Such blooms are known as harmful algal blooms (HABs) or red tides. Many species of zooplankton and shellfish that feed by filtering seawater to ingest phytoplankton may incorporate these toxins into their tissues during red-tide events. Fish, seabirds, and whales that consume affected zooplankton and shellfish can exhibit a variety of responses detrimental to survival. These toxins can also cause amnesic, diarrhetic, or paralytic shellfish poisoning in humans and may require the closure of aquaculture operations or even wild fisheries. Despite their generally small size, plankton even play a major role in the pace and extent of climate change itself through their contribution to the carbon cycle. The ability of the oceans to act as a sink for CO2 relies largely on plankton functioning as a ‘biological pump’. By reducing the concentration of CO2 at the ocean surface through photosynthetic uptake, phytoplankton allow more CO2 to diffuse into surface waters from the atmosphere. This process continually draws CO2 into the oceans and has helped to remove half of the CO2 produced by humans from the atmosphere and distributed it into the oceans. Plankton play a further role in the biological pump because much of the CO2 that is fixed by phytoplankton and then eaten by zooplankton sinks to the ocean floor in the bodies of uneaten and dead phytoplankton, and zooplankton fecal pellets. This carbon may then be locked up within sediments. Phytoplankton also help to shape climate by changing the amount of solar radiation reflected back to space (the Earth’s albedo). Some phytoplankton produce dimethylsulfonium propionate, a precursor of dimethyl sulfide (DMS). DMS evaporates from the ocean, is oxidized into sulfate in the atmosphere, and then forms cloud condensation nuclei. This leads to more clouds, increasing the Earth’s albedo and cooling the climate. Without these diverse roles performed by plankton, our oceans would be desolate, polluted, virtually lifeless, and the Earth would be far less resilient to the large quantities of CO2 produced by humans.

Beacons of Climate Change Plankton are ideal beacons of climate change for a host of reasons. First, plankton are ecthothermic (their body temperature varies with the surroundings),


455

456

PLANKTON AND CLIMATE

so their physiological processes such as nutrient uptake, photosynthesis, respiration, and reproductive development are highly sensitive to temperature, with their speed doubling or tripling with a 10 1C temperature rise. Global warming is thus likely to directly impact the pace of life in the plankton. Second, warming of surface waters lowers its density, making the water column more stable. This increases the stratification, so that more energy is required to mix deep nutrient-rich water into surface layers. It is these nutrients that drive surface biological production in the sunlit upper layers of the ocean. Thus global warming is likely to increase the stability of the ocean and diminish nutrient enrichment and reduce primary productivity in large areas of the tropical ocean. There is no such direct link between temperature and nutrient enrichment in terrestrial systems. Third, most plankton species are short-lived, so there is tight coupling between environmental change and plankton dynamics. Phytoplankton have lifespans of days to weeks, whereas land plants have lifespans of years. Plankton systems will therefore respond rapidly, whereas it takes longer before terrestrial plants exhibit changes in abundance attributable to climate change. Fourth, plankton integrate ocean climate, the physical oceanic and atmospheric conditions that drive plankton productivity. There is a direct link between climate and plankton abundance and timing. Fifth, plankton can show dramatic changes in distribution because they are free floating and most remain so their entire life. They thus respond rapidly to changes in temperature and oceanic currents by expanding and contracting their ranges. Further, as plankton are distributed by currents and not by vectors or pollinators, their dispersal is less dependent on other species and more dependent on physical processes. By contrast, terrestrial plants are rooted to their substrate and are often dependent upon vectors or pollinators for dispersal. Sixth, unlike other marine groups such as fish and many intertidal organisms, few plankton species are commercially exploited so any long-term changes can more easily be attributed to climate change. Last, almost all marine life has a planktonic stage in their life cycle because ocean currents provide an ideal mechanism for dispersal over large distances. Evidence suggests that these mobile life stages known as meroplankton are even more sensitive to climate change than the holoplankton, their neighbors that live permanently in the plankton. All of these attributes make plankton ideal beacons of climate change. Impacts of climate change on plankton are manifest as predictable changes in the distribution of individual species and communities, in the timing of important life cycle events or

phenology, in abundance and community structure, through the impacts of ocean acidification, and through their regulation by climate indices. Because of this sensitivity and their global importance, climate impacts on plankton are felt throughout the ecosystems they support.

Changes in Distribution Plankton have exhibited some of the fastest and largest range shifts in response to global warming of any marine or terrestrial group. The general trend, as on land, is for plants and animals to expand their ranges poleward as temperatures warm. Probably the clearest examples are from the Northeast Atlantic. Members of a warm temperate assemblage have moved more than 1000-km poleward over the last 50 years (Figure 1). Concurrently, species of a subarctic (cold-water) assemblage have retracted to higher latitudes. Although these translocations have been associated with warming in the region by up to 1 1C, they may also be a consequence of the stronger northward flowing currents on the European shelf edge. These shifts in distribution have had dramatic impacts on the food web of the North Sea. The cool water assemblage has high biomass and is dominated by large species such as Calanus finmarchicus. Because this cool water assemblage retracts north as waters warm, C. finmarchicus is replaced by Calanus helgolandicus, a dominant member of the warmwater assemblage. This assemblage typically has lower biomass and contains relatively small species. Despite these Calanus species being indistinguishable to all but the most trained eye, the two species contrast starkly in their seasonal cycles: C. finmarchicus peaks in spring whereas C. helgolandicus peaks in autumn. This is critical as cod, which are traditionally the most important fishery of the North Sea, spawn in spring. As cod eggs hatch into larvae and continue to grow, they require good food conditons, consisting of large copepods such as C. finmarchicus, otherwise mortality is high and recruitment is poor. In recent warm years, however, C. finmarchicus is rare, there is very low copepod biomass during spring, and cod recruitment has crashed.

Changes in Phenology Phenology, or the timing of repeated seasonal activities such as migrations or flowering, is very sensitive to global warming. On land, many events in spring are happening earlier in the year, such as the arrival of swallows in the UK, emergence of butterflies in the US, or blossoming of cherry trees in Japan. Recent



0.0

Warm temperate assemblage 1958−81

Subarctic assemblage 1958−81

1982−99

1982−99

2000−02

2000−02

0.02

0.04

0.06

0.08 0.1 0.0 0.2 0.4 Mean number of species per sample

0.6

0.8

457

1.0

Figure 1 The northerly shift of the warm temperate assemblage (including Calanus helgolandicus) into the North Sea and retraction of the subarctic assemblage (including Calanus finmarchicus) to higher latitudes. Reproduced by permission from Gregory Beaugrand.

evidence suggests that phenological changes in plankton are greater than those observed on land. Larvae of benthic echinoderms in the North Sea are now appearing in the plankton 6 weeks earlier than they did 50 years ago, and this is in response to warmer temperatures of less than 1 1C. In echinoderms,

temperature stimulates physiological developments and larval release. Other meroplankton such as larvae of fish, cirrepedes, and decapods have also responded similarly to warming (Figure 2). Timing of peak abundance of plankton can have effects that resonate to higher trophic levels. In the


458


11.5

6.0 SST

11.0

Phenology

Early seasonal cycles

6.5

10.5

7.0

10.0

7.5

9.5

8.0

9.0

8.5

Monthly phenology (inverted)

Sea surface temperature (°C)

Mean

9.0

8.5 Late seasonal cycles

1958 1963 1968 1973 1978 1983 1988 1993 1998 2003 Year Figure 2 Monthly phenology (timing) of decapod larval abundance and sea surface temperature in the central North Sea from 1958 to 2004. Reproduced from Edwards M, Johns DG, Licandro P, John AWG, and Stevens DP (2006) Ecological status report: Results from the CPR Survey 2004/2005. SAHFOS Technical Report 3: 1–8.

North Sea, the timing each year of plankton blooms in summer over the last 50 years has advanced, with phytoplankton appearing 23 days earlier and copepods 10 days earlier. The different magnitude of response between phytoplankton and zooplankton may lead to a mismatch between successive trophic levels and a change in the synchrony of timing between primary and secondary production. In temperate marine systems, efficient transfer of marine primary and secondary production to higher trophic levels, such as those occupied by commercial fish species, is largely dependent on the temporal synchrony between successive trophic production peaks. This type of mismatch, where warming has disturbed the temporal synchrony between herbivores and their plant food, has been noted in other biological systems, most notably between freshwater zooplankton and diatoms, great tits and caterpillar biomass, flycatchers and caterpillar biomass, winter moth and oak bud burst, and the red admiral butterfly and stinging nettle. Such mismatches compromise herbivore survival. Dramatic ecosystem repercussions of climatedriven changes in phenology are also evident in the subarctic North Pacific Ocean. Here a single copepod species, Neocalanus plumchrus, dominates the

zooplankton biomass. Its vertical distribution and development are both strongly seasonal and result in an ephemeral (2-month duration) annual peak in upper ocean zooplankton biomass in late spring. The timing of this annual maximum has shifted dramatically over the last 50 years, with peak biomass about 60 days earlier in warm than cold years. The change in timing is a consequence of faster growth and enhanced survivorship of early cohorts in warm years. The timing of the zooplankton biomass peak has dramatic consequences for the growth performance of chicks of the planktivorous seabird, Cassin’s auklet. Individuals from the world’s largest colony of this species, off British Columbia, prey heavily on Neocalanus. During cold years, there is synchrony between food availability and the timing of breeding. During warm years, however, spring is early and the duration of overlap of seabird breeding and Neocalanus availability in surface waters is small, causing a mismatch between prey and predator populations. This compromises the reproductive performance of Cassin’s auklet in warm years compared to cold years. If Cassin’s auklet does not adapt to the changing food conditions, then global warming will place severe strain on its long-term survival.



Changes in Abundance The most striking example of changes in abundance in response to long-term warming is from foraminifera in the California Current. This plankton group is valuable for long-term climate studies because it is more sensitive to hydrographic conditions than to predation from higher trophic levels. As a result, its temporal dynamics can be relatively easily linked to changes in climate. Foraminifera are also well preserved in sediments, so a consistent time series of observations can be extended back hundreds of years. Records in the California Current show increasing numbers of tropical/subtropical species throughout the twentieth century reflecting a warming trend, which is most dramatic after the 1960s (Figure 3). Changes in the foraminifera record echo not only increase in many other tropical and subtropical taxa in the California Current over the last few decades, but also decrease in temperate species of algae, zooplankton, fish, and seabirds. Changes in abundance through alteration of enrichment patterns in response to enhanced stratification is often more difficult to attribute to climate change than are shifts in distribution or phenology, but may have greater ecosystem consequences. An illustration from the Northeast Atlantic highlights the role that global warming can have on stratification and thus plankton abundances. In this region, phytoplankton become more abundant when cooler regions warm, probably because warmer temperatures boost metabolic rates and enhance stratification in these often windy, cold, and well-mixed regions.

Number per cm2 per year

25 0 3

459

But phytoplankton become less common when already warm regions get even warmer, probably because warm water blocks nutrient-rich deep water from rising to upper layers where phytoplankton live. This regional response of phytoplankton in the North Atlantic is transmitted up the plankton food web. When phytoplankton bloom, both herbivorous and carnivorous zooplankton become more abundant, indicating that the plankton food web is controlled from the ‘bottom up’ by primary producers, rather than from the ‘top down’ by predators. This regional response to climate change suggests that the distribution of fish biomass will change in the future, as the amount of plankton in a region is likely to influence its carrying capacity of fish. Climate change will thus have regional impacts on fisheries. There is some evidence that the frequency of HABs is increasing globally, although the causes are uncertain. The key suspect is eutrophication, particularly elevated concentrations of the nutrients nitrogen and phosphorus, which are of human origin and discharged into our oceans. However, recent evidence from the North Sea over the second half of the twentieth century suggests that global warming may also have a key role to play. Most areas of the North Sea have shown no increase in HABs, except off southern Norway where there have been more blooms. This is primarily a consequence of the enhanced stratification in the area caused by warmer temperatures and lower salinity from meltwater. In the southern North Sea, the abundance of two key HAB species over the last 45 years is positively related to warmer ocean temperatures. This work

Globigerina bulloides

Neogloboquadrina dutertrei

0 2

Globigerinoides ruber

0 0.6

Orbulina universa

0 0.1 0

Other subtropical species

1750

1775

1800

1825

1850

1875

1900

1925

1950

1975

2000

Year Figure 3 Fluxes of tropical/planktonic foraminifera in Santa Barbara Basin (California Current). Tropical/subtropical foraminifera showing increased abundance in the twentieth century. Reproduced from Field DB, Baumgartner TR, Charles CD, Ferreira-Bartrina V, and Ohman M (2006) Planktonic foraminifera of the California Current reflect 20th century warming. Science 311: 63–66.


460


supports the notion that the warmer temperatures and increased meltwater runoff anticipated under projected climate change scenarios are likely to increase the frequency of HABs. Although most evidence for changes in abundance in response to climate change are from the Northern Hemisphere because this is where most (plankton) science has concentrated, there is a striking example from waters around Antarctica. Over the last 30 years, there has been a decline in the biomass of krill Euphausia superba in the Southern Ocean that is a consequence of warmer sea and air temperatures. In many areas, krill has been replaced by small gelatinous filter-feeding sacs known as salps, which occupy the less-productive, warmer regions of the Southern Ocean. The decline in krill is likely to be a consequence of warmer ocean temperatures impacting sea ice. It is not only that sea ice protects krill from predation, but also the algae living beneath the sea ice and photosynthesizing from the dim light seeping through are a critical food source for krill. As waters have warmed, the extent of winter sea ice and its duration have declined, and this has led to a deterioration in krill density since the 1970s. As krill are major food items for baleen whales, penguins, seabirds, fish, and seals, their declining population may have severe ramifications for the Southern Ocean food web.

Impact of Acidification A direct consequence of enhanced CO2 levels in the ocean is a lowering of ocean pH. This is a consequence of elevated dissolved CO2 in seawater altering the carbonate balance in the ocean, releasing more hydrogen ions into the water and lowering pH. There has been a drop of 0.1 pH units since the Industrial Revolution, representing a 30% increase in hydrogen ions. Impacts of ocean acidification will be greatest for plankton species with calcified (containing calcium carbonate) shells, plates, or scales. For organisms to build these structures, seawater has to be supersaturated in calcium carbonate. Acidification reduces the carbonate saturation of the seawater, making calcification by organisms more difficult and promoting dissolution of structures already formed. Calcium carbonate structures are present in a variety of important plankton groups including coccolithophores, mollusks, echinoderms, and some crustaceans. But even among marine organisms with calcium carbonate shells, susceptibility to acidification varies depending on whether the crystalline form of their calcium carbonate is aragonite or

calcite. Aragonite is more soluble under acidic conditions than calcite, making it more susceptible to dissolution. As oceans absorb more CO2, undersaturation of aragonite and calcite in seawater will be initially most acute in the Southern Ocean and then move northward. Winged snails known as pteropods are probably the plankton group most vulnerable to ocean acidification because of their aragonite shell. In the Southern Ocean and subarctic Pacific Ocean, pteropods are prominent components of the food web, contributing to the diet of carnivorous zooplankton, myctophids, and other fish and baleen whales, besides forming the entire diet of gymnosome mollusks. Pteropods in the Southern Ocean also account for the majority of the annual flux of both carbonate and organic carbon exported to ocean depths. Because these animals are extremely delicate and difficult to keep alive experimentally, precise pH thresholds where deleterious effects commence are not known. However, even experiments over as little as 48 h show shell deterioration in the pteropod Clio pyrimidata at CO2 levels approximating those likely around 2100 under a business-as-usual emissions scenario. If pteropods cannot grow and maintain their protective shell, their populations are likely to decline and their range will contract toward lowerlatitude surface waters that remain supersaturated in aragonite, if they can adapt to the warmer temperature of the waters. This would have obvious repercussions throughout the food web of the Southern Ocean. Other plankton that produce calcite such as foraminifera (protist plankton), mollusks other than pteropods (e.g., squid and mussel larvae), coccolithophores, and some crustaceans are also vulnerable to ocean acidification, but less so than their cousins with aragonite shells. Particularly important are coccolithophorid phytoplankton, which are encased within calcite shells known as liths. Coccolithophores export substantial quantities of carbon to the seafloor when blooms decay. Calcification rates in these organisms diminish as water becomes more acidic (Figure 4). A myriad of other key processes in phytoplankton are also influenced by seawater pH. For example, pH is an important determinant of phytoplankton growth, with some species being catholic in their preferences, whereas growth of other species varies considerably between pH of 7.5 and 8.5. Changes in ocean pH also affect chemical reactions within organisms that underpin their intracellular physiological processes. pH will influence nutrient uptake kinetics of phytoplankton. These effects will have repercussions for phytoplankton community



Emiliania huxleyi

461

Gephyrocapsa oceanica

(a)

(b)

(c)

(d)

(e)

(f)

330 ppm

780−850 ppm

Figure 4 Scanning electron microscopy photographs of the coccolithophores Emiliania huxleyi and Gephyrocapsa oceanica collected from cultures incubated at CO2 levels of about 300 and 780–850 ppm. Note the difference in the coccolith structure (including distinct malformations) and in the degree of calcification of cells grown at normal and elevated CO2 levels. Scale bar = 1 mm. Repinted by permission from Macmillan Publishers Ltd., Nature, Riebesell U, Zondervan I, Rost B, Tortell PD, Zeebe RE, and Morel FMM, Reduced calcification of marine plankton in response to increased atmospheric CO2, 407: 364–376, Copyright (2000).

composition and productivity, with flow-on effects to higher trophic levels.

Climate Variability Many impacts of climate change are likely to act through existing modes of variability in the Earth’s climate system, including the well-known El Ninõ/ Southern Oscillation (ENSO) and the North Atlantic Oscillation (NAO). Such large synoptic pressure fields alter regional winds, currents, nutrient dynamics, and water temperatures. Relationships between integrative climate indices and plankton composition, abundance, or productivity provide an insight into how climate change may affect ocean biology in the future. ENSO is the strongest climate signal globally, and has its clearest impact on the biology of the tropical Pacific Ocean. Observations from satellite over the past decade have shown a dramatic global decline in primary productivity. This trend is caused by enhanced stratification in the low-latitude oceans in response to more frequent El Ninõ events. During an El Ninõ, upper ocean temperatures warm, thereby enhancing stratification and reducing the availability of nutrients for phytoplankton

growth. Severe El Ninõ events lead to alarming declines in phytoplankton, fisheries, marine birds and mammals in the tropical Pacific Ocean. Of concern is the potential transition to more frequent El Ninõ-like conditions predicted by some climate models. In such circumstances, enhanced stratification across vast areas of the tropical ocean may reduce primary productivity, decimating fish, mammal, and bird populations. Although it is unknown whether the recent decline in primary productivity is already a consequence of climate change, the findings and underlying understanding of climate variability are likely to provide a window to the future. Further north in the Pacific, the Pacific Decadal Oscillation (PDO) has a strong multi-decadal signal, longer than the ENSO period of a few years. When the PDO is negative, upwelling winds strengthen over the California Current, cool ocean conditions prevail in the Northeast Pacific, copepod biomass in the region is high and is dominated by large cool-water species, and fish stocks such as coho salmon are abundant (Figure 5). By contrast, when the PDO is positive, upwelling diminishes and warm conditions exist, the copepod biomass declines and is dominated by small less-nutritious species, and the abundance of coho salmon plunges.


462


Warm

Pacific Decadal Oscillation

15

Cool

Warm

Cool

0

−15 1940

1950

1960

1970

1980

1990

2000

1960

1970

1980

1990

2000

1970

1980

1990

2000

1970

1980

1990

2000

CALCOFI zooplankton

100

0

−100 1950

Coho survival

6

0

−4

Northern copepods

0.5

0

−0.5

Figure 5 Annual time series in the Northeast Pacific: the PDO index from May to September; anomalies of zooplankton biomass (displacement volumes) from the California Current region (CALCOFI zooplankton); anomalies of coho salmon survival; and biomass anomalies of cold-water copepod species (northern copepods). Positive (negative) PDO index indicates warmer (cooler) than normal temperatures in coastal waters off North America. Reproduced from Peterson WT and Schwing FB (2003) A new climate regime in Northeast Pacific ecosystems. Geophysical Research Letters 30(17): 1896 (doi:10.1029/2003GL017528).

These transitions between alternate states have been termed regime shifts. It is possible that if climate change exceeds some critical threshold, some marine systems will switch permanently to a new state that is less favorable than present.

In Hot Water: Consequences for the Future With plankton having relatively simple behavior, occurring in vast numbers, and amenable to


463


experimental manipulation and automated measurements, their dynamics are far more easily studied and modeled than higher trophic levels. These attributes make it easier to model potential impacts of climate change on plankton communities. Many of our insights gained from such models confirm those already observed from field studies. The basic dynamics of plankton communities have been captured by nutrient–phytoplankton– zooplankton (NPZ) models. Such models are based on a functional group representation of plankton communities, where species with similar ecological function are grouped into guilds to form the basic biological units in the model. Typical functional groups represented include diatoms, dinoflagellates, coccolithophores, microzooplankton, and mesozooplankton. There are many global NPZ models constructed by different research teams around the world. These are coupled to global climate models (GCMs) to provide future projections of the Earth’s climate system. In this way, alternative carbon dioxide emission scenarios can be used to investigate possible future states of the ocean and the impact on plankton communities.

One of the most striking and worrisome results from these models is that they agree with fieldwork that has shown general declines in lower trophic levels globally as a result of large areas of the surface tropical ocean becoming more stratified and nutrient-poor as the oceans heat up (see sections titled ‘Changes in abundance’ and ‘Climate variability’). One such NPZ model projects that under a middleof-the-road emissions scenario, global primary productivity will decline by 5–10% (Figure 6). This trend will not be uniform, with increases in productivity by 20–30% toward the Poles, and marked declines in the warm stratified tropical ocean basins. This and other models show that warmer, morestratified conditions in the Tropics will reduce nutrients in surface waters and lead to smaller phytoplankton cells dominating over larger diatoms. This will lengthen food webs and ultimately support fewer fish, marine mammals, and seabirds, as more trophic linkages are needed to transfer energy from small phytoplankton to higher trophic levels and 90% of the energy is lost within each trophic level through respiration. It also reduces the oceanic uptake of CO2 by lowering the efficiency of the Primary production (g C per m2 per year) 30

20 40° N 10

0°

0

−10 40° S −20

−30 100° E

160° W

60° W

40° E

Figure 6 Change in primary productivity of phytoplankton between 2100 and 1990 estimated from an NPZ model. There is a global decline in primary productivity by 5–10%, with an increase at the Poles of 20–30%. Reproduced from Bopp L, Monfray P, Aumont O, et al. (2001) Potential impact of climate change on marine export production. Global Biogeochemical Cycles 15: 81–99.


464


biological pump. This could cause a positive feedback between climate change and the ocean carbon cycle: more CO2 in the atmosphere leads to a warmer and more stratified ocean, which supports less and smaller plankton, and results in less carbon being drawn from surface ocean layers to deep waters. With less carbon removed from the surface ocean, less CO2 would diffuse into the ocean and more CO2 would accumulate in the atmosphere. It is clear that plankton are beacons of climate change, being extremely sensitive barometers of physical conditions. We also know that climate impacts on plankton reverberate throughout marine ecosystems. More than any other group, they also influence the pace and extent of climate change. The impact of climate change on plankton communities will not only determine the future trajectory of marine ecosystems, but the planet.

See also Marine Plankton Communities. Plankton.

Further Reading Atkinson A, Siegel V, Pakhomov E, and Rothery P (2004) Long-term decline in krill stock and increase in salps within the Southern Ocean. Nature 432: 100--103. Beaugrand G, Reid PC, Ibanez F, Lindley JA, and Edwards M (2002) Reorganisation of North Atlantic marine copepod biodiversity and climate. Science 296: 1692--1694. Behrenfield MJ, O’Malley RT, Siegel DA, et al. (2006) Climate-driven trends in contemporary ocean productivity. Nature 444: 752--755. Bertram DF, Mackas DL, and McKinnell SM (2001) The seasonal cycle revisited: Interannual variation and ecosystem consequences. Progress in Oceanography 49: 283--307.

Bopp L, Aumont O, Cadule P, Alvain S, and Gehlen M (2005) Response of diatoms distribution to global warming and potential implications: A global model study. Geophysical Research Letters 32: L19606 (doi:10.1029/2005GL023653). Bopp L, Monfray P, Aumont O, et al. (2001) Potential impact of climate change on marine export production. Global Biogeochemical Cycles 15: 81--99. Edwards M, Johns DG, Licandro P, John AWG, and Stevens DP (2006) Ecological status report: Results from the CPR Survey 2004/2005. SAHFOS Technical Report 3: 1--8. Edwards M and Richardson AJ (2004) The impact of climate change on the phenology of the plankton community and trophic mismatch. Nature 430: 881--884. Field DB, Baumgartner TR, Charles CD, Ferreira-Bartrina V, and Ohman M (2006) Planktonic forminifera of the California Current reflect 20th century warming. Science 311: 63--66. Hays GC, Richardson AJ, and Robinson C (2005) Climate change and plankton. Trends in Ecology and Evolution 20: 337--344. Peterson WT and Schwing FB (2003) A new climate regime in Northeast Pacific ecosystems. Geophysical Research Letters 30(17): 1896 (doi:10.1029/2003GL017528). Raven J, Caldeira K, Elderfield H, et al. (2005) Royal Society Special Report: Ocean Acidification Due to Increasing Atmospheric Carbon Dioxide. London: The Royal Society. Richardson AJ (2008) In hot water: Zooplankton and Climate change. ICES Journal of Marine Science 65: 279--295. Richardson AJ and Schoeman DS (2004) Climate impact on plankton ecosystems in the Northeast Atlantic. Science 305: 1609--1612. Riebesell U, Zondervan I, Rost B, Tortell PD, Zeebe RE, and Morel FMM (2000) Reduced calcification of marine plankton in response to increased atmospheric CO2. Nature 407: 364--367.


PLANKTON VIRUSES J. Fuhrman, University of Southern California, Los Angeles, CA, USA I. Hewson, University of California Santa Cruz, Santa Cruz, CA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Although they are the tiniest biological entities in the sea, typically 20–200 nm in diameter, viruses are integral components of marine planktonic systems. They are extremely abundant in the water column, typically 1010 per liter in the euphotic zone, and they play several roles in system function: (1) they are important agents in the mortality of prokaryotes and eukaryotes; (2) they act as catalysts of nutrient regeneration and recycling, through this mortality of host organisms; (3) because of their host specificity and density dependence, they tend to selectively attack the most abundant potential hosts, thus may ‘kill the winner’ of competition and thereby foster diversity; and (4) they may also act as agents in the exchange of genetic material between organisms, a critical factor in evolution and also in relation to the spread of human-engineered genes. Although these processes are only now becoming understood in any detail, there is little doubt that viruses are significant players in aquatic and marine plankton.

History It has only been in the past 25 years that microorganisms like bacteria and small protists have been considered ‘major players’ in planktonic food webs. The initial critical discovery, during the mid-1970s, was of high bacterial abundance as learned by epifluorescence microscopy of stained cells, with counts typically 109 l 1 in the plankton. These bacteria were thought to be heterotrophs (organisms that consume preformed organic carbon), because they apparently lacked photosynthetic pigments like chlorophyll (later it was learned that this was only partly right, as many in warm waters are in fact chlorophyllcontaining prochlorophytes). With such high abundance, it became important to learn how fast they were dividing, in order to quantify their function in the food web. Growth rates were estimated primarily by the development and application of methods measuring bacterial DNA synthesis. The results of

these studies showed that bacterial doubling times in typical coastal waters are about 1 day. When this doubling time was applied to the high abundance, to calculate how much carbon the bacteria are taking up each day, it became apparent that bacteria are consuming a significant amount of dissolved organic matter, typically at a carbon uptake rate equivalent to about half the total primary production. However, the bacterial abundance remains relatively constant over the long term, and they are too small to sink out of the water column. Therefore, there must be mechanisms within the water to remove bacteria at rates similar to the bacterial production rate. In the initial analysis, most scientists thought that grazing by protists was the only significant mechanism keeping the bacterial abundance in check. This was because heterotrophic protists that can eat bacteria are extremely common, and laboratory experiments suggested they are able to control bacteria at nearnatural-abundance levels. However, some results pointed to the possibility that protists are not the only things controlling bacteria. In the late 1980s, careful review of multiple studies showed that grazing by protists was often not enough to balance bacterial production, and this pointed to the existence of additional loss processes. About that same time, data began to accumulate that viruses may also be important as a mechanism of removing bacteria. The evidence is now fairly clear that this is the case, and it will be outlined below. This article briefly summarizes much of what is known about how viruses interact with marine microorganisms, including general properties, abundance, distribution, infection of bacteria, mortality rate comparisons with protists, biogeochemical effects, effects on species compositions, and roles in genetic transfer and evolution.

General Properties Viruses are small particles, usually about 20–200 nm long, and consist of genetic material (DNA or RNA, single or double stranded) surrounded by a protein coat (some have lipid as well). They have no metabolism of their own and function only via the cellular machinery of a host organism. As far as is known, all cellular organisms appear to be susceptible to infection by some kind of virus. Culture studies show that a given type of virus usually has a restricted host range, most often a single species or genus, although some viruses infect only certain


465

466

PLANKTON VIRUSES

subspecies and o0.5% may infect more than one genus. Viruses have no motility of their own, and contact the host cell by diffusion. They attach to the host usually via some normally exposed cellular component, such as a transport protein or flagellum. There are three basic kinds of virus reproduction (Figure 1). In lytic infection, the virus attaches to a host cell and injects its nucleic acid. This nucleic acid (sometimes accompanied by proteins carried by the virus) causes the host to produce numerous progeny viruses, the cell then bursts, progeny are released, and the cycle begins again. In chronic infection, the progeny virus release is not lethal and the host cell releases the viruses by extrusion or budding over many generations. In lysogeny after injection, the viral genome becomes part of the genome of the host cell and reproduces as genetic material in the host cell line unless an ‘induction’ event causes a switch to lytic infection. Induction is typically caused by DNA damage, such as from ultraviolet (UV) light or chemical mutagens such as mitomycin C. Viruses may also be involved in killing cells by mechanisms that do not result in virus reproduction.

Observation of Marine Viruses Viruses are so small that they are at or below the resolution limit of light microscopy (c. 0.1 mm). Therefore electron microscopy is the only way to observe any detail of viruses. Sample preparation requires concentrating the viruses from the water onto an electron microscopy grid (coated with a thin transparent organic film). Because viruses are denser than seawater, this can be done by ultracentrifugation, typically at forces of at least 100 000 g for a few hours. It should be noted that under ordinary gravity, forces like drag and Brownian motion prevent viruses from sinking. To be observable the viruses must be made electron-dense, typically by staining with uranium salts. The viruses are recognized by their size, shape, and staining properties (usually electron-dense hexagons or ovals, sometimes with a tail), and counted. Typical counts are on the order of 1010 viruses per liter in surface waters, with abundance patterns similar to those of heterotrophic bacteria (see below). Recently, it has been found that viruses can also be stained with nucleic acid stains like SYBR Green I, and observed and counted by

Chronic

Lytic

Lysogenic Virus attachment to host

Virus attachment to host 'Temperate' phage Host with viral nucleic acid integrated into genome or as extrachromosomal element

Viral nucleic acid injected

Nucleic acid (DNA or RNA) injected into host Normal division Induction Host makes copies of viral nucleic acid and coat proteins

Host makes copies of viral nucleic acid and coat proteins Normal division continues unless induction occurs

Host releases progeny viruses without lysing, by budding or extrusion Normal division

Viruses self-assemble inside host

Host lyses (bursts) to release progeny viruses

Host progeny continue to release viruses unless 'cured'

Figure 1 Virus life cycles. See text for explanation.


PLANKTON VIRUSES

Figure 2 Epifluorescence micrograph of prokaryotes and viruses from 16 km offshore of Los Angeles, stained with SYBR Green I. The viruses are the very numerous tiny bright particles, and the bacteria are the rarer larger particles. Bacterial size is approximately 0.4–1 mm in diameter.

epifluorescence microscopy. This is faster, easier, and less expensive than transmission electron microscopy (TEM). Epifluorescence viewing of viruses is shown in Figure 2, a micrograph of SYBR Green I-stained bacteria and viruses, which dramatically illustrates the high relative virus abundance. Epifluorescence microscopy of viruses is possible even though the viruses are below the resolution limit of light because the stained viruses are a source of light and appear as bright spots against a dark background (like stars visible at night). Epifluorescence counts are similar to or even slightly higher than TEM counts from seawater.

What Kinds of Viruses Occur in Plankton? Microscopic observation shows the total, recognizable, virus community, but what kinds of viruses make up this community, and what organisms are they infecting? Most of the total virus community is thought to be made up of bacteriophages (viruses that infect bacteria). This is because viruses lack metabolism and have no means of actively moving from host to host (they depend on random diffusion), so the most common viruses would be expected to infect the most common organism, and bacteria are by far the most abundant organisms in the plankton. Field studies show a strong correlation between viral and bacterial numbers, whereas the correlations between viruses and chlorophyll are weaker. This

467

suggests that most viruses are bacteriophages rather than those infecting phytoplankton or other eukaryotes. However, viruses infecting cyanobacteria (Synechococcus) are also quite common and sometimes particularly abundant, exceeding 108 per liter in some cases. Even though most of the viruses probably infect prokaryotes, viruses for eukaryotic plankton are also readily found. For example, those infecting the common eukaryotic picoplankter, Micromonas pusilla, are sometimes quite abundant, occasionally near 108 per liter in coastal waters. Overall, the data suggest that most viruses from seawater infect nonphotosynthetic bacteria or archaea, but viruses infecting prokaryotic and eukaryotic phytoplankton also can make up a significant fraction of the total. A great leap in our understanding of viral diversity has occurred with the application of molecular biological techniques. These involve the analysis of variability of DNA sequence of a single gene product (e.g., a capsid head protein, or an enzyme that makes DNA), or through a technique known as viriomics, whereby all genomes in a sample are cut up into small pieces, then all the pieces sequenced and reassembled. These studies have revealed that: (1) there is a lot of diversity of closely related viruses infecting the same or similar strains of microorganisms, a phenomenon known as ‘microdiversity’; (2) RNA-containing viruses comprise a small percentage of the viral mix in oceanic waters, and are similar to viruses that infect insects and mollusks; and (3) that the diversity of viruses in a liter of seawater is astonishingly high, estimated at 4104 different types for a sample from coastal California.

Virus Abundance Total direct virus counts have been made in many planktonic environments – coastal, offshore, temperate, polar, tropical, and deep sea. Typical virus abundance is 1–5 1010 l 1 in rich near-shore surface waters, decreasing to about 0.1–1 1010 l 1 in the euphotic zone of offshore low-nutrient areas, and also decreasing with depth, by about a factor of 10. A typical deep offshore profile is shown in Figure 3. Seasonal changes are also common, with viruses following general changes in phytoplankton, bacteria, etc. Virus:prokaryote ratios also provide an interesting comparison. In plankton, this ratio is typically 5–25, and commonly close to 10, even as abundance drops to low levels in the deep sea. Why this ratio stays in such a relatively narrow range is a mystery, but it does suggest a link and also tight regulatory mechanisms between prokaryotes and viruses.


468

PLANKTON VIRUSES

Viruses l−1

Prokaryotes l−1 108

109

108

109

1010

Virus:prokaryote ratio 1011

0

0

0

500

500

500

1000

1000

1000

1500

1500

1500

2000

Depth (m)

0

Depth (m)

Depth (m)

107

2000

5

10

15

20

25

30

2000

2500

2500

2500

3000

3000

3000

3500

3500

3500

4000

4000

4000

Figure 3 Depth profile of total prokaryote (bacteria–archaea) counts, total viral counts, and virus:prokaryote ratios from the Coral Sea (Apr. 1998), as determined by epifluorescence microscopy of SYBR Green-stained samples. Note the log scales of the counts.

Viral Activities Viruses have no physical activity of their own, so ‘viral activity’ usually refers to lytic infection. However, before discussing such infection, lysogeny and chronic infection are considered briefly. Lysogeny, where the viral genome resides in the host’s genome (Figure 1), is common. Lysogens (bacteria harboring integrated viral genomes) can easily be found and isolated from seawater, and lysogeny, which is linked to genetic transfer in a variety of bacteria, probably impacts microbial population dynamics and evolution. However, the induction rate appears to be low and seasonally variable under ordinary natural conditions, and lysogenic induction appears to be responsible for only a tiny fraction of total virus production in marine systems for most of the year. On the other hand, at this time we simply do not know if chronic infection is a significant process in natural systems. Release of filamentous (or other kinds of budding) viruses from native marine bacteria has not been noted in TEM studies of plankton, nor have significant numbers of free filamentous viruses (however, filamentous viruses have been observed in marine and freshwater sediments). But they could have been missed.

Regarding lytic infection, there are several studies with a variety of approaches that all generally conclude that viruses cause approximately 10–50% of total microbial mortality, depending on location, season, etc. These estimates are convincing, having been determined in several independent ways. These include: (1) TEM observation of assembled viruses within host cells, representing the last step before lysis; (2) measurement of viral decay rates; (3) measurement of viral DNA synthesis; (4) measurement of the disappearance rate of bacterial DNA in the absence of protists; (5) use of fluorescent virus tracers to measure viral production and removal rates simultaneously; and (6) direct observation of viral appearance in incubations where the viral abundance has been decreased several fold, yet host abundances remain undiluted.

Comparison to Mortality from Protists Because the earlier thinking was that protists are the main cause of bacterial mortality in marine planktonic systems, it is useful to ask how the contribution of viruses to bacterial mortality compares to that of protists. Multiple correlation analysis of abundances


PLANKTON VIRUSES

of bacteria, viruses, and flagellates showed virusinduced mortality of bacteria could occasionally prevail over flagellate grazing, especially at high bacterial abundances. In more direct comparisons, measuring virus and protist rates by multiple independent approaches, the total mortality typically balances production, and viruses are found to be responsible for anything ranging from a negligible proportion to the majority of total mortality. To sum up these studies, the consensus is that viruses are often responsible for a significant fraction of bacterial mortality in marine plankton, typically in the range 10–40%. Sometimes viruses may dominate bacterial mortality, and sometimes they may have little impact on it. It is unknown what controls this balance, but it probably includes variation in host abundance, because when hosts are less common, the viruses are more likely to be inactivated before diffusing to a suitable host, as well as the exact types of bacteria that are present and their palatability. Application of new molecular techniques, based on ribosomal RNA sequences, and variable regions within the host genomes have revealed that aquatic bacterial communities are typically dominated by a handful of bacterial taxa, while most taxa make up a tiny proportion of cell numbers. This would seem to support the notion that mostly dominant taxa are targets of viral attack.

Roles in Food Web and Geochemical Cycles The paradigm of marine food webs has been revised a great deal in response to the initial discovery of high bacterial abundance and productivity. It is now well established that a large fraction of the total carbon and nutrient flux in marine systems passes through the heterotrophic bacteria via the dissolved organic matter. How do viruses fit into this picture? Three features of viruses are particularly relevant: (1) small size; (2) composition; and (3) mode of causing cell death, which is to release cell contents and progeny viruses to the surrounding seawater. When a host cell lyses, the resultant viruses and cellular debris are made up of easily digested protein and nucleic acid, plus all other cellular components, in a nonsinking form that is practically defined as dissolved organic matter. This is composed of dissolved molecules (monomers, oligomers, and polymers), colloids, and cell fragments. This material is most probably utilized by bacteria as food. If it was a bacterium that was lysed in the first place, then uptake by other bacteria represents a partly closed loop, whereby bacterial biomass is consumed mostly by

469

other bacteria. Because of respiratory losses and inorganic nutrient regeneration connected with the use of dissolved organic substances, this loop has the net effect of oxidizing organic matter and regenerating inorganic nutrients (Figure 4). This bacterial–viral loop effectively ‘steals’ production from protists that would otherwise consume the bacteria, and segregates the biomass and activity into the dissolved and smallest particulate forms. The potentially large effect has been modeled mathematically, and such models show that significant mortality from viruses greatly increases bacterial community growth and respiration rates. Segregation of matter in viruses, bacteria, and dissolved substances leads to better retention of nutrients in the euphotic zone in virus-infected systems because more material remains in these small nonsinking forms. In contrast, reduced viral activity leads to more material in larger organisms that either sink themselves or as detritus, transporting carbon and inorganic nutrients to depth. The impact can be particularly great for potentially limiting nutrients like N, P, and Fe, which are relatively concentrated in bacteria compared to eukaryotes. At least one study has demonstrated that in cyanobacterial cultures that are starved for trace metals, the Fe in viral lysate of another culture of equal density is taken up within an hour of supplement, which suggests a major role of viral lysis in the availability of limiting nutrients. Therefore, the activity of viruses has the possible effect of helping to support higher levels of biomass and productivity in the planktonic system as a whole. There are other potential geochemical effects of viral infection and its resultant release of cell contents to the water, owing to the chemical and physical nature of the released materials and the location in the water column where the lysis occurs. For example, polymers released from lysed cells may facilitate aggregation and sinking of material from the euphotic zone. On the other hand, viral lysis of microorganisms within sinking aggregates may lead to the breakup of the particles, converting some sinking particulate matter into nonsinking dissolved material and colloids at whatever depth the lysis occurs. This contributes to the dissolution of sinking organic matter and its availability to free-living bacteria in the ocean’s interior. Viruses, particularly lysogens, have long been known to confer to hosts the ability to produce toxins – in fact cholera toxin is only produced by the cholera bacteria when lysogenized. Through study of viriomics in marine plankton, it is now known that viral genes may encode for several geochemically important enzymes including phosphorus uptake, and components of the photosynthetic apparatus. These appear to be used both under stable host–virus


470

PLANKTON VIRUSES

Higher trophic levels

Heterotrophic prokaryotes Viruses

'Futile cycle' Cellular debris Grazers

Dissolved organic matter

Inorganic nutrients C, N, P, Fe, ...

Phytoplankton Figure 4 Prokaryote–viral loop within the microbial food web. Arrows represent transfer of matter.

conditions, as well as during host replication, when the enhanced substrate utilization capabilities inferred to hosts is used to produce new virus particles.

Effects on Host Species Compositions and Control of Blooms Viruses mostly infect only one species or related species, and are also density dependent. Thus, the most common or dominant hosts in a mixed community are believed to be most susceptible to infection, and rare ones least so. Lytic viruses can increase only when the average time to diffuse from host to host is shorter than the average time that at least one member from each burst remains infectious. Therefore, when a species or strain becomes more abundant, it is more susceptible to infection. The end result is that viral infection works in opposition to competitive dominance. This may help to solve Hutchinson’s ‘paradox of plankton’, which asks us how so many different kinds of phytoplankton coexist on only a few potentially limiting resources, when competition theory predicts one or a few competitive winners. Although there have been several possible explanations for this paradox, viral

activity may also help solving it, because as stated above, competitive dominants become particularly susceptible to infection whereas rare species are relatively protected. Extending this argument, one might conclude that viruses have the potential to control algal blooms, such as those consisting of coccolithophorids, and so-called ‘red tides’ of dinoflagellates. There is now evidence that at least under some circumstances this may be true. Declining blooms have been found to contain numerous infected cells. Along similar lines, it is now commonly thought that viral infection can influence the species composition of diverse host communities even when they are responsible for only a small portion of the host mortality. This is again because of the near-species specificity of viruses in contrast to the relatively particular tastes of protists or metazoa as grazers. This conclusion is supported by mathematical models as well as limited experimental evidence.

Resistance The development of host resistance to viral infection is a common occurrence in laboratory and medical


PLANKTON VIRUSES

situations. Such resistance, where hosts mutate to resist the viral attack, is well known from nonmarine experiments with highly simplified laboratory systems. However, the existence of an apparently high infection rate in plankton suggests that the rapid development resistance is not a dominant factor in the plankton. How can the difference between laboratory and field situations be explained? Natural systems with many species and trophic levels have far more interactions than simple laboratory systems. One might expect that a species with a large fraction of mortality from one type of virus benefits from developing resistance. However, resistance is not always an overall advantage. It often leads to a competitive disadvantage from the loss of some important receptor, for example, involved in substrate uptake. Even resistance to viral attachment, without any receptor loss, if that were possible, would not necessarily be an advantage. For a bacterium in a low-nutrient environment whose growth may be limited by N, P, or organic carbon, unsuccessful infection by a virus (e.g., stopped intracellularly by a restriction enzyme, or with a genetic incompatibility) may be a useful nutritional benefit to the host organism, because the virus injection of DNA is a nutritious boost rich in C, N, and P. Even the viral protein coat, remaining outside the host cell, is probably digestible by bacterial proteases. From this point of view, one might even imagine bacteria using ‘decoy’ virus receptors to lure viral strains that cannot successfully infect them. With the proper virus and host distributions, the odds could be in favor of the bacteria, and if an infectious virus (i.e., with a protected restriction site) occasionally gets through, the cell line as a whole may still benefit from this strategy. There are other reasons why resistance might not be an overall advantage. As described earlier, model results show that the heterotrophic bacteria as a group benefit substantially from viral infection, raising their production by taking carbon and energy away from larger organisms. Viruses also raise the overall system biomass and production by helping to keep nutrients in the lighted surface waters. However, these arguments would require invoking some sort of group selection theory to explain how individuals would benefit from not developing resistance (i.e., why not ‘cheat’ by developing resistance and letting all the other organisms give the group benefits of infection?). In any case, evidence suggests that even if resistance of native communities to viral infection may be common, it is not a dominant force, because there is continued ubiquitous existence of viruses at roughly 10 times greater abundance than bacteria and with turnover times on the order of a

471

day (as discussed above). Basic mass balance calculations show that significant numbers of hosts must be infected and releasing viruses all the time. For example, with a typical lytic burst size of 50 and viral turnover time of 1 day, maintenance of a 10fold excess of viruses over bacteria requires 20% of the bacteria to lyse daily. The lack of comprehensive resistance might be due to frequent development of new virulent strains, rapid dynamics or patchiness in species compositions, or to a stable coexistence of viruses and their hosts. All these are possible, and they are not mutually exclusive. Lysogeny, which is common in marine bacterioplankton, also conveys resistance to superinfection (i.e., infection by the same, or similar virus). This may be an advantage so long as the prophage remains uninduced, in that viral genomes often also contain useful genes involved in membrane transport, photosynthesis, etc., that benefit the host. However, because the prophage is carried around in the host cell, this may also place an extra burden on the host cell machinery (i.e., it must also replicate the viral genome in addition to the host genome).

Genetic Transfer Viruses can also play central roles in genetic transfer between microorganisms, through two processes. In an indirect mechanism, viruses mediate genetic transfer by causing the release of DNA from lysed host cells that may be taken up and used as genetic material by another microorganism. This latter process is called transformation. A more direct process is known as transduction, where viruses package some of the host’s own DNA into the phage head and then inject it into another potential host. Transduction in aquatic environments has been shown to occur in a few experiments. Although transduction usually occurs within a restricted host range, recent data indicate that some marine bacteria and phages are capable of transfer across a wide host range. Although the extent of these mechanisms in natural systems is currently unknown, they could have important roles in population genetics, by homogenizing genes within a potential host population, and also on evolution at relatively long timescales. Gene transfer across species lines is an integral component of microbial evolution, as shown in the genomes of modern-day microbes that contain numerous genes that have obviously been transferred from other species. On shorter timescales, this process can be responsible for the dissemination of genes that may code for novel properties, whether introduced to native communities naturally or via genetic engineering.


472

PLANKTON VIRUSES

Summary It is now known that viruses can exert significant control of marine microbial systems. A major effect is on mortality of bacteria and phytoplankton, where viruses are thought to stimulate bacterial activity at the expense of larger organisms. This also stimulates the entire system via improved retention of nutrients in the euphotic zone. Other important roles include influence on species compositions and possibly also genetic transfer.

See also Bacterioplankton. Carbon Cycle. Nitrogen Cycle. Phosphorus Cycle. Phytoplankton Blooms. Primary Production Distribution. Primary Production Methods. Primary Production Processes.

Further Reading Ackermann HW and Du Bow MS (1987) Viruses of Prokaryotes, Vol. I: General Properties of Bacteriophages. Boca Raton, FL: CRC Press. Azam F, Fenchel T, Gray JG, Meyer-Reil LA, and Thingstad T (1983) The ecological role of water-column microbes in the sea. Marine Ecology Progress Series 10: 257--263. Bratbak G, Thingstad F, and Heldal M (1994) Viruses and the microbial loop. Microbial Ecology 28: 209--221. Breitbart M, Salamon P, Andresen B, et al. (2002) Genomic analysis of uncultured marine viral communities. Proceedings of the National Academy of Sciences of the United States of America 99: 14250--14255. Culley A, Lang AS, and Suttle CA (2003) High diversity of unknown picorna-like viruses in the sea. Nature 424: 1054--1057. Fuhrman JA (1992) Bacterioplankton roles in cycling of organic matter: The microbial food web. In: Falkowski

PG and Woodhead AS (eds.) Primary Productivity and Biogeochemical Cycles in the Sea, pp. 361--383. New York: Plenum. Fuhrman JA (1999) Marine viruses: Biogeochemical and ecological effects. Nature 399: 541--548. Fuhrman JA (2000) Impact of viruses on bacterial processes. In: Kirchman DL (ed.) Microbial Ecology of the Oceans, pp. 327--350. New York: Wiley-Liss. Fuhrman JA and Suttle CA (1993) Viruses in marine planktonic systems. Oceanography 6: 51--63. Lindell D, Sullivan MB, Johnson ZI, Tolonen AC, Rohwer F, and Chisholm SW (2004) Transfer of photosynthesis genes to and from Prochlorococcus viruses. Proceedings of the National Academy of Sciences of the United States of America 101: 11013--11018. Maranger R and Bird DF (1995) Viral abundance in aquatic systems – a comparison between marine and fresh waters. Marine Ecology Progress Series 121: 217--226. Noble RT and Fuhrman JA (1998) Use of SYBR Green I for rapid epifluorescence counts of marine viruses and bacteria. Aquatic Microbial Ecology 14(2): 113--118. Paul JH, Rose JB, and Jiang SC (1997) Coliphage and indigenous phage in Mamala Bay, Oahu, Hawaii. Applied and Environmental Microbiology 63(1): 133--138. Proctor LM (1997) Advances in the study of marine viruses. Microscopy Research and Technique 37(2): 136--161. Suttle CA (1994) The significance of viruses to mortality in aquatic microbial communities. Microbial Ecology 28: 237--243. Thingstad TF, Heldal M, Bratbak G, and Dundas I (1993) Are viruses important partners in pelagic food webs? Trends in Ecology and Evolution 8(6): 209--213. Wilhelm SW and Suttle CA (1999) Viruses and nutrient cycles in the sea – viruses play critical roles in the structure and function of aquatic food webs. Bioscience 49(10): 781--788. Wommack KE and Colwell RR (2000) Virioplankton: Viruses in aquatic ecosystems. Microbiology and Molecular Biology Reviews 64(1): 69--114.


PLATFORMS: AUTONOMOUS UNDERWATER VEHICLES J. G. Bellingham, Monterey Bay Aquarium Research Institute, Moss Landing, CA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction Autonomous underwater vehicles (AUVs) are untethered mobile platforms used for survey operations by ocean scientists, marine industry, and the military. AUVs are computer-controlled, and may have little or no interaction with a human operator while carrying out a mission. Being untethered, they must also store energy onboard, typically relying on batteries. Motivations for using AUVs include such factors as ability to access otherwise-inaccessible regions, lower cost of operations, improved data quality, and the ability to acquire nearly synoptic observations of processes in the water column. An example of the first is operations under Arctic and Antarctic ice, an environment in which operations of human-occupied vehicles and tethered platforms are either difficult or impossible. Illustrating the next two points, AUVs are becoming the platform of choice for deep-water bathymetric surveying in the offshore oil industry because they are less expensive than towed platforms as well as produce higherquality data (because they are decoupled from motion of the sea surface). Finally, the use of fleets of AUVs enables the rapid acquisition of distributed data sets over regions as large as 10 000 km2. AUVs are a new class of platform for the ocean sciences, and consequently are evolving rapidly. The Self Propelled Underwater Research Vehicle (SPURV) AUV, built at the University of Washington Applied Physics Laboratory, was first operated in 1967. However, adoption by the ocean sciences community lagged until the late 1990s. Adoption was spurred on by two developments: AUV development teams started supporting science field programs with AUV capabilities, and AUVs that nondevelopers could purchase and operate became available. The first served the purpose of building a user base and demonstrating AUV capabilities. The second enabled scientists to obtain and operate their own vehicles. Today, a wide variety of AUVs are available from commercial manufacturers. An even larger

number of companies develop subsystems and sensors for AUVs. The most common class of AUV in use today is a torpedo-like vehicle with a propeller at its stern, and steerable control surfaces to control turns and vertical motion (see Figure 1). These vehicles are used when speed or efficiency of motion is an important consideration. Such torpedo-like vehicles range in weight from a few tens of kilograms to thousands of kilograms. Most typical are vehicles weighing a few hundred kilograms, with an endurance of about a day at a speed of c. 1.5 m s1. Often they have parallel mid-bodies, which allow the vehicle length to be extended without large hydrodynamic consequences. This is useful when it is necessary to add new sensors or batteries to a vehicle. A disadvantage of torpedo-like vehicles is that, like an aircraft, they must maintain forward motion to generate lift over its control surfaces, and thus are not controllable at very low speed through the water. Gliders are a class of vehicles that use changes in buoyancy rather than a propeller for propulsion. Gliders use their ability to control buoyancy to generate vertical motion. Vertical motion is translated into horizontal motion with lifting surfaces, usually wings mounted in about the middle of the vehicle (see Figure 1). Several types of gliders weighing about 50 kg are in use today. These comparatively small vehicles are designed to move slowly, about 0.25 m s1, and operate sensors consuming a watt or less. By minimizing power consumption, these gliders can operate for periods of months using high-energydensity primary batteries. Disadvantages of this class of system are that they are limited to vertical profiling flight tracks, and can be overwhelmed by ocean currents, especially in the coastal environment or within boundary currents. However, larger gliders in development and testing will operate at higher speeds, and thus not suffer from this limitation. A final class of AUVs uses multiple thrusters to provide capabilities similar to that of a helicopter or a ship with dynamic positioning (see Figure 1). The additional thrusters enable maneuvers such as hovering, translating sideways, and moving vertically. These vehicles are used when maneuverability is needed, for example, when operation near a very rough bottom is a necessity. The disadvantage is that the additional thrusters reduce efficiency for moving large distances or at high speeds.


473

474

PLATFORMS: AUTONOMOUS UNDERWATER VEHICLES

Figure 1 From top-left corner clockwise: the Hugin AUV, the Spray glider (courtesy, MBARI), the ABE AUV (courtesy, Dana Yoerger, WHOI), and the Dorado AUV (courtesy, MBARI). Hugin and Dorado are examples of propeller-driven vehicles optimized for moving through the water efficiently, and have a torpedo-like configuration. Spray is an example of a glider, which is a buoyancy-driven vehicle that has no propeller. ABE is a highly maneuverable vehicle capable of hovering or pivoting in place, and of moving straight up or down. Also illustrated are different handling strategies. The large Hugin vehicle is launched and recovered from a ship using a stern ramp. Spray is hand-launched and recovered. ABE is launched and recovered with a crane. Dorado is shown being launched and recovered with a capture mechanism suspended from a J-frame.

While the vehicles described above are representative of the most commonly used systems, a wide range of other vehicles are in development or are in limited use. AUVs that come to the surface and use solar panels to recharge batteries have been demonstrated in seagoing operations. Gliders which extract energy from thermal differences in the ocean for

propulsion have also been tested. A hybrid vehicle is being developed for reaching the deepest portion of the ocean. The hybrid vehicle operates as a tethered platform via a disposable fiber optic link for tasks requiring human perception, in other words as a remotely operated vehicle (ROV), but operates as an AUV when that link is severed. These are just a few



of the diverse AUVs being developed to answer the needs of ocean science. Terminology

The military term for an AUV is unmanned underwater vehicle, or UUV. This phrase is ambiguous in that it can also refer to ROVs. While ROVs are unmanned, they are tethered and designed to be operated by human, and thus are not considered autonomous. However, common usage is to employ UUV as a synonym for AUV. Military terminology is relevant as many AUVs in use by the scientific community were developed under navy funding, and the military continues to be the largest single investor in AUV technology. Consequently, the technical literature on AUVs also employs military terminology.

Basics of AUV Performance Energy is a fundamental limitation for underwater vehicles. Thus, energy efficiency is a fundamental driver for vehicle design and operations. This section outlines the relationship between vehicle speed and endurance, and its dependence on factors such as power consumed by onboard systems. A simple but useful model for power consumption P of an AUV is as follows: P ¼ Pprop þ H

where Pprop ¼

1 CD Arv3 Z 2

½1

Here the total electrical power consumed by the vehicle, P, is equal to the sum of propulsion power, Pprop and hotel load, H. Hotel load is simply the power consumed by all subsystems other than propulsion. Propulsion power is a function of the drag coefficient of the vehicle, CD, the area of the vehicle, A, the density of water, r, the speed of the vehicle, v, and the efficiency of the propulsion system, Z. What are the typical values for the coefficients in eqn [1]? Consider an ‘example vehicle’ which is a 12 3/400 (0.32 m) diameter torpedo-like AUV. Note that this is a standard for a mid-size class AUV. For such a vehicle, parameter values might be CD ¼ 0.2 (based on frontal area), A ¼ 0.082 m2, and Z ¼ 0.5. Hotel load would depend on sensors, but an overall value of 30 W might be representative, although mapping sonars would consume much more power. We use r ¼ 1027 kg m–3. Note that these numbers, except seawater density of course, can differ greatly from vehicle to vehicle. For example, gliders optimized for low speed and long endurance

475

can operate with hotel loads on the order of a watt or less, but at the cost of operating a few very simple sensors. The power relationship provides insight into a variety of AUV design considerations. For example, the speed at which the vehicle will consume the least energy per unit distance traveled (the energetically optimum speed) can be computed from observing that power divided by vehicle speed equals energy per unit distance. Finding the minimum of P/v with respect to v yields the optimum speed from an energy conservation perspective: vopt ¼

ZH CD Ar

1=3 ½2

For the example vehicle values given above, the optimum vehicle speed is approximately 1 m s1. What can we say about vehicle performance at the energetically optimum speed? Substituting eqn [2] into eqn [1] we find that the power consumed at the optimum speed is (3/2)H which for our example vehicle is 45 W. If the total energy capacity of the battery system is Ecap, then the maximum range of the vehicle will be: dmax

1=3 2Ecap Z ¼ 3 CD ArH 2

½3

If our example vehicle carries 10 kg of high-energydensity primary batteries providing a total of 1.3 107 J, it will have an endurance of 80 h, and a range of 280 km. The same vehicle with 10 kg of rechargeable batteries, with one-third the energy capacity of the high-energy-density batteries, will have its endurance and range reduced proportionately. There are caveats to the above discussion. For example, propulsion efficiency, Z, is typically a strong function of speed as the electrical motors used tend to have comparatively narrow ranges of efficiency. In practice, a vehicle’s propulsion system is optimized for a particular speed and power. Also, vehicles are often operated at higher speeds than the energetically optimum speed given by eqn [2]. For example, operators of an AUV attended by a ship will be more sensitive to minimizing ship costs than to optimizing energy efficiency of the AUV.

AUV Systems and Technology AUVs are highly integrated devices, containing a variety of mechanical, electrical, and software subsystems. Figure 2 shows internal and external views of a deep-diving vehicle equipped with mapping


476


Doppler velocity logger and inertial navigation system Sub-bottom profiler

Fluid intake

Batteries

Sonar electronics pressure vessel

Computer enclosure

Water sensor

Acoustic transducers

Flotation

Antenna

Lifting eye

Sonar receive Fairing

Thruster and duct

Figure 2 A propeller-driven, modular AUV with labeled subsystems. The top view shows the interior in which an internal mechanical frame supports pressure vessels, internal components, and the propulsion system. This AUV, called a Dorado, is a ‘flooded’ vehicle because the fairings are not watertight, and thus the interior spaces fill with water. Consequently internal components must all be capable of withstanding ambient pressures. Joining rings are visible between the yellow fairing segments on the lower figure. These allow the vehicle to be separated along its axis, allowing replacement or addition of hull sections. This allows reconfiguration of the vehicle with new payloads, and if desired, with new batteries. The propulsion system on this vehicle is a ducted thruster capable of being tilted both vertically and horizontally, to steer the vehicle in the vertical and horizontal planes. Courtesy of Farley Shane, MBARI.

sonars. The anatomy of an AUV typically includes the following subsystems:

• • • • • •

software and computers capable of managing vehicle subsystems to accomplish specific tasks and even complete missions in the absence of human control; energy storage to provide power; propulsion system; a system for controlling vehicle orientation and velocity; sensors for measuring vehicle attitude, heading, and depth; pressure vessels for housing key electrical components;

• • • • •

navigation sensors to determine the vehicle position; communication devices to allow communication of human operators with the AUV; locating devices to allow operators to track the vehicle and locate it for recovery or in the case of emergencies; devices for monitoring vehicle health (e.g., leaks or battery failure); emergency systems for ensuring vehicle recovery in the event of failure of primary systems.

The mechanical design of an AUV must address issues such as drag, neutral buoyancy, the highly dynamic nature of launch and recovery, and the need



to protect many delicate electrical components from seawater. The desire to operate large numbers of sensors for long distances encourages the construction of larger vehicles to hold the necessary equipment and batteries. However, the need for ease of handling and minimizing logistical costs encourages the design of smaller vehicles. Operational demands will also create constraints on vehicle design; for example, launch and recovery factors will impose the need for lift points and discourage external appendages that will be easily broken. The need to service vehicle components imposes a requirement that internal components be easily accessible for servicing and testing, and when necessary, replacement. In addition, a host of supporting software and hardware are required to operate an AUV. Depending on the nature of the operations, supporting equipment will include:

• • • • • •

software and computers for configuring vehicle mission plans, and for reviewing vehicle data; systems for communicating with the vehicle both on deck and when deployed; systems for recharging and monitoring vehicle batteries; handling gear for transporting, deploying, and recovering AUVs; devices for detecting locating devices on the AUV; acoustic tracking systems for monitoring the location of the vehicle when in the vicinity of a support vessel.

AUV Mission Software

Functionally, AUV software must address a variety of needs, including: allowing human operators to specify objectives, managing vehicle subsystems to achieve mission objectives, logging data for subsequent review, and ensuring safety of the vehicle in the event of failures or unexpected circumstances. The software must be capable of managing vehicle sensors and control systems to maintain a set heading, speed, and depth. The software might also need to support interacting with a human operator during a mission. In addition to software on the vehicle itself, AUV operators rely on a suite of software applications to configure and validate missions, to maintain vehicle subsystems such as batteries, to review data generated by the vehicle, to prepare mission summaries, and when possible, to track and manage the vehicle while underway. The exponential growth of computational power available for both onboard and off-board computers, as well as the increasingly pervasive nature of the Internet, are

477

supporting a steady increase in software capabilities for AUVs. Most AUV missions involve sequential tasks such as descending from the surface to a set depth, then transiting to a survey location at a set speed, and then conducting a survey which might involve flying a lawn-mower pattern. The vehicle may be commanded to maintain constant altitude over the bottom if the vehicle is mapping the seafloor. A water-column mission might require the vehicle profile in the vertical plane, moving in a saw-tooth pattern called a yo-yo. The mission will likely include a transit from the end of the survey to a recovery location, with a final ascent to the surface for recovery. During the mission, the vehicle will monitor the performance of onboard subsystems, and in the event of detection of anomalies, like a low battery level, or a failed mission sensor, may abort the mission and return to the recovery point early. A more catastrophic failure might lead to the vehicle shutting down primary systems, and dropping a drop weight so as to float to the surface, and calling for help via satellite or direct radio frequency (RF) communications. More complex vehicle missions can involve capabilities such as adapting survey operations to obtain better measurements, or managing tasks such as AUV docking. An example of the first might be as simple as a yo-yo mission which cues its vertical inflections from water temperature in order to follow a thermocline. Also in the category of adaptive, but more demanding, surveys, is the capability of following a thermal plume to its source, for example, when an AUV is used to search for hydrothermal vents. The docking of an AUV with an underwater structure encompasses yet a different type of complexity, created by the large number of steps in the process, and the high likelihood that individual steps will fail. For example, docking involves homing on a docking structure, orienting for final approach, engaging the dock, and making physical connections to establish power and communication links. Any one of these steps might fail due to external perturbations; for example, currents or turbulence in the marine environment might cause the vehicle to miss the dock. The vehicle must be able to detect failures and execute a process to recover and try again. Docking is representative of the increasingly complex capabilities AUVs are expected to master with high reliability. Navigation

The ability for an AUV to determine its location on the Earth is essential for most scientific applications.


478


However, navigation in the subsea environment is complicated by the opacity of seawater to all but very low frequency electromagnetic radiation, rendering ineffective the use of commonly used technologies such as the Global Positioning System (GPS) and other radio-based navigation techniques. Consequently, navigation underwater relies primarily on various acoustic and dead-reckoning techniques and the occasional excursion to the surface where radiobased methods can be used. There is no single method of underwater navigation that satisfies all operational needs, rather a variety of methods are employed depending on the circumstances. Dead-reckoning methods integrate a vehicle’s velocity in time to obtain an updated location. In order to dead-reckon, the vehicle must know both the direction and speed of its travel. The simplest methods use a magnetic compass to determine direction, and use speed through the water as a proxy for Earth-referenced speed. However, the large number of error sources for magnetic compasses make measurement of heading to better than a degree accuracy technically challenging. Currents pose even more of a problem, as they may be comparable to the vehicle speed in amplitude, yet are not sensed by a water-relative measurement. Dead-reckoning is improved by measuring velocity relative to the seafloor, for example, using a Doppler velocity log (DVL) or a correlation velocity log. A DVL is commonly used by AUVs to measure velocity by measuring the Doppler shift of sound reflected off the seafloor. Correlation velocity logs are more complex in concept, involving measurement of the correlation of two pulses of sounds transmitted by the vehicle, reflected off the seafloor, and received by a hydrophone array. In practice, DVLs are used when a vehicle operates close to the seafloor, perhaps within 200 m, while correlation velocity logs are used when the vehicle is operating in mid-water columns or near the surface in deep water. Inertial navigation system (INS) technology is well developed, as it is widely used for platforms like aircraft and missiles. However, INS units appropriate for underwater use are expensive enough that they are used only when navigation requirements are stringent, for example, for producing high-accuracy maps. A modern INS includes an array of accelerometers for measuring acceleration on three axes and a laser or fiber optic gyroscope for measuring changes in orientation. Additionally, an INS will include a GPS for initializing the unit’s location and orientation, and a computer for acquiring and processing data from INS component sensors. The position reported by an INS will have an error which will grow in time, and thus it is important to constrain INS

error with ancillary measurements of velocity and position. For example, combining an INS with a DVL for constraining velocity can result in a system which provides navigation accuracies better than 0.05% of distance traveled. The two acoustic navigation methodologies most frequently used in AUV operations are ultrashort baseline navigation (USBL) and long baseline (LBL) navigation. A USBL system uses an array of hydrophone separated by a distance comparable to the wavelength of sound to measure the direction of propagation of an acoustic signal. Most often, a USBL system is mounted on a ship, and used to track a vehicle relative to the ship. With knowledge of the ship’s location and orientation, the location of the AUV can also be determined. In contrast, LBL navigation acoustically measures the range between the vehicle and an array of widely separated devices of known location. A common LBL approach is to place transponders on the seafloor, and let the vehicle range off the transponders. The process of determining location using ranges from known locations is called spherical navigation, as the vehicle should be located at the intersection of spheres with the measured radius, centered on the respective transponders. An alternative LBL navigation method is to track a vehicle which pings at a preset time to an array of hydrophones at known locations. If the time of the ping is not known, the problem of solving for the vehicle location is called hyperbolic navigation, as only the difference in time of arrival of the ping at the various hydrophones can be determined, and this knowledge constrains the vehicle to be on a hyperbola between the respective receivers. If the time of the ping is known, perhaps triggered at a preselected time by a carefully calibrated clock, then the problem reduces to spherical navigation. In practice, a wide variety of USBL and LBL systems have been implemented for underwater navigation. They must all address the challenges of acoustic propagation in the ocean, which include the absorption of sound by seawater, diffraction by speed of sound variations in the underwater environment, scattering by reflecting surfaces, and acoustic noise generated by physical, geological, biological, and anthropogenic processes. Other methods of navigation include using geophysical parameters, for example, water depth, to constrain the vehicle location in the context of known maps. These geophysically based navigation methods, similar to terrain contour mapping (TERCOM) navigation used by cruise missiles, depend on having good maps ahead of time. There are software approaches in development that simultaneously build maps and use those same maps for navigation.



These methods are called SLAM for simultaneous localization and mapping.

Using AUVs for Ocean Science Mapping the Seafloor

AUVs are becoming the platform of choice for highresolution seafloor maps. Obtaining high-resolution maps requires operating mapping sonars near the seafloor. Alternatives to an AUV include crewed submersibles and tethered platforms. Crewed submersibles are too valuable for routine mapping, and are reserved for other uses which require the presence of humans. Towed vehicles are used for sonar mapping, but have disadvantages as compared with AUVs, especially in deeper water. The principal problem is the high drag of the cable used for a tow sled, which in water depths of several thousand meters will limit speeds to approximately half a meter per second. Even at these slow speeds, a towed platform will stream behind the towing ship, creating several problems. Controlling the position of the towed vehicle over the bottom is very difficult, even when running on a constant heading. When surveying a defined area on the seafloor in a series of passes, the turns between passes may take longer than the actual survey passes themselves, as it is necessary to turn slowly to maintain control of the

479

towed body. If ultrashort baseline acoustic navigation techniques are used to determine the vehicle position, then layback of the towed body behind the ship introduces significant errors as compared with having the ship directly over the sonar platform. For this reason, some commercial use of towed sonar platforms use two ships, one to tow the sonar platform, and one positioned directly over the platform to determine its precise location. Finally, surface motion of the ship will be efficiently coupled to the tow body by the tow cable. Thus, even near the seafloor, the tow body will be subject to sea state experienced by the ship. Consequently, attraction of the use of AUVs includes more economical operations and high data quality. Figure 3 shows a cost comparison of a commercial deep-water towed survey and an equivalent AUV survey. Sonar systems used on AUVs for mapping include multibeam sonar, side scan sonar, and sub-bottom profilers. Multibeam sonars, operating at frequencies of hundreds of kilohertz in the case of AUV-mounted systems, allow measurement of range to the seafloor in multiple sonar beams and are used to build up three-dimensional maps such as that in Figure 4. Side scan sonars used by AUVs also typically operate at frequencies of hundreds of kilohertz, and are used to image seafloor features. Side scan sonars are particularly useful for finding objects, for example, looking for a shipwreck resting Report preparation

Cost (normalized to total towed survey cost)

100%

Tracking boat

90%

Time spent turning to return to survey site

80%

Survey

70%

Transit to work site

60%

Mobilization demob

50% 40% 30% 20% 10% 0% Towed survey

AUV

Figure 3 A comparison of the economics of deep survey taken from costs of a survey with a towed vehicle, and projected costs of the same survey with an AUV. The principal cost saving derives from the ability of the AUV to turn much faster than a deep-towed vehicle, reducing the total survey time. Also, the AUV can be acoustically tracked by its mother ship, while a towed vehicle requires a second ship for tracking because the towed vehicle will trail far behind the tow ship. Finally, mobilization and demobilization costs for the AUV can also be lower, although this depends on the size of the AUV employed.


480


122° 00' 50" W

122° 00' 40" W

122° 00' 30" W

122° 00' 20" W

km

36° 46' 50" N 0

0.1

36° 46' 50 0.2

36° 46' 45" N

36° 46' 45

36° 46' 40" N

36° 46' 40

36° 46' 35" N

36° 46' 35

36° 46' 30" N

36° 46' 30 122° 00' 50" W

−1012.5 −1000.0 −987.5

122° 00' 40" W

−975.0

122° 00' 30" W

−962.5 −950.0 −937.5 Topography (m)

−925.0

122° 00' 20" W

−912.5

−900.0

−887.5

Figure 4 A bathymetric survey produced by an AUV at a depth of about 1000 m. Note the very small size of the survey area and high resolution of the bathymetry. Courtesy of Dave Caress, MBARI.

on the seafloor. Sub-bottom profilers use lowerfrequency sound, ranging from 1 kHz to tens of kilohertz in the case of an AUV-mounted system, to penetrate into the seafloor. Depending on the bottom type (e.g., sandy, muddy, or rock), a subbottom system might penetrate tens of meters. Cumulatively sonar payloads will consume comparatively large amounts of energy, perhaps hundreds of watts. Mapping also requires high-fidelity navigation, and thus sonar-equipped AUVs will often also use more sophisticated navigation approaches, like inertial navigation. Consequently, mapping AUVs of today are larger, more sophisticated AUVs. Observing the Water Column

AUVs provide a relatively new tool for observing the physical, chemical, and biological properties of the

ocean. The smaller, buoyancy-driven gliders are unique in their combination of mobility and endurance, moving at about a quarter of a meter per second for periods of months. Larger vehicles carry more comprehensive payloads at higher speeds, but for shorter periods. Such vehicles might operate at 1.5 m s1 for a day. A common flight profile is to fly the vehicle on a constant heading, while moving between two depth extremes in a saw-tooth pattern. Often the upper depth extreme will be close to the surface. This strategy allows the production of vertical sections of ocean properties, such as those in Figure 5. Variations of this strategy might have the vehicle moving in a lawn-mower or zigzag pattern in the horizontal plane, to develop a full threedimensional map of ocean properties. Figure 6 shows a visualization of an internal wave interacting with a phytoplankton layer using such a three-dimensional mapping strategy.



481

Temperature 16

0

14

10

12 20

10

30

8 0

1000

2000

3000

4000

5000

6000

7000

Fluorometer 0 1500

10

1000 20 500 30 0

1000

2000

3000

4000

5000

6000

7000

0

Optical backscatter 0

1000 800

10

600 20 400 30

200 0

1000

2000

3000

4000 5000 Downtrack distance (m)

6000

7000

Figure 5 Vertical sections of water properties obtained by an Odyssey AUV operating in Massachusetts Bay. The y-axis of each figure is depth, in meters, and the x-axis is horizontal distance in meters. The top section shows temperature in degrees Celsius, the middle shows chlorophyll fluorescence in arbitrary units, and the bottom shows optical backscatter, also in arbitrary units. The path of the vehicle is shown as a white line, and the interpolated values of the measured property are plotted in color. The vehicle alternated between obtaining high-resolution observations of the thin layer of organisms at the thermocline with full water column profiles.

All AUVs are limited by the availability of sensors. Temperature, salinity, currents, dissolved oxygen, nitrate, optical backscatter properties, and chlorophyll fluorescence are examples of the growing in situ sensing capabilities available for AUVs. However, many important properties, for example, pH, dissolved carbon dioxide, and dissolved iron, cannot be measured reliably from a small moving platform. Furthermore, detection of marine organisms is usually accomplished by proxy; for example, chlorophyll fluorescence provides an indicator for phytoplankton abundance. In situ methods which directly detect, classify, and quantify marine organism abundance are not available, yet are increasingly important for understanding the structure and dynamics of ocean ecosystems. Operations in Ice-covered Oceans

AUVs offer unique operational capabilities for science in ice-covered oceans. Successful under-ice

operation has been carried out with AUVs in both the Arctic and Antarctic. Sea ice poses special operational challenges for seagoing ocean scientists. For example, ships with ice-breaking capability can operate in the ice pack, but will typically not be able to hold station, or even assure that tethers and cables deployed over the side will not be severed. AUVs are attractive in that they provide horizontal mobility under ice, and the ability to conduct operations near the seafloor without the complications intrinsic in tether management. Challenges of operating AUVs under ice revolve around the need to assure return of the AUV to the ship for recovery, the process of recovering the AUV through the ice onto the ship, the potential for having an AUV fail and become trapped under ice, and the difficulty of carrying out tasks that would normally be accomplished having an AUV surface (e.g., obtaining a GPS update). Most safety strategies for AUVs in ice-free oceans default to bring


482


Monterey Bay

Front

5m

10 m

Phytoplankton layer Internal wave

15 m

20 m 2 km 1 km 0 km

2 km

4 km

6 km

0 km

Figure 6 Interaction of layer of phytoplankton with an internal wave in Monterey Bay. Both physical and biological properties were measured by an Odyssey AUV, which moved in a horizontal zigzag pattern across the survey volume, while profiling constantly in the vertical plane. The phytoplankton layer, shown in green, was detected by a chlorophyll fluorescence sensor on the AUV. The cyan surface shows deflection of a level of constant density of seawater by a passing internal wave. Courtesy of John Ryan, MBARI.

the vehicle directly to the surface, for example, by dropping a weight. In the Arctic or Antarctic, this strategy could result in the vehicle becoming trapped under very thick ice, making the vehicle harder to find and potentially impossible to recover. The usual surface location devices such as RF beacons, strobes, and combinations of RF communication and satellite navigation will not work. Clearly AUV operations within the ice pack entail higher risk and a more sophisticated vehicle. Observation Systems, Observatories, and AUVs

An understanding of power consumption of AUVs provides insight to the attractiveness of employing multiple vehicles for certain ocean observation problems. In some circumstances a survey must be accomplished within a set period. In oceanography,

time-constrained surveys are most often encountered when surveying a dynamic process. For example, if the temporal decorrelation of ocean fields associated with upwelling off Monterey Bay is about 48 h, attempts to map the ocean fields need to be accomplished within that time frame. Scales of spatial variability will also determine acceptable separation of observations: for example, decorrelation lengths in Monterey Bay are on the order of 20 km, so observations need to be spaced significantly closer to minimize errors in reconstructing the ocean field. How does this relate to the number of vehicles required to accomplish such a survey? Consider a grid survey of a 100 km 100 km area with a resolution of 10 km. A single vehicle would have to travel c. 1000 km at a speed of nearly 6 m s1, traveling in a lawn-mower pattern. Using the example vehicle values from the ‘Basics of AUV



performance’ section, the AUV would consume about 3500 W if it were capable of operating at such a high speed. In contrast, six of the same vehicles operating at their optimum speed would consume a total of 270 W. In other words, the six vehicles would consume 12 times less energy for the complete 48-h survey. Autonomous mobile platforms are making observation of the interior of the ocean more affordable and more flexible, enabling the practical realization of coupled observation–prediction systems. For example, in late summer 2003, a diverse fleet of AUVs was deployed to observe and predict the evolution of episodic wind-driven upwelling in the environs of Monterey Bay. Over 21 different autonomous robotic systems, three ships, an aircraft, a coastal ocean dynamics application radar (CODAR), drifters, floats, and numerous fixed (moored) observation assets were deployed in the

483

Autonomous Ocean Sampling Network (AOSN) II field program (Figure 7). Gliding vehicles, with an endurance of weeks to months, provided a continuous presence with a minimal sensor suite. A few propeller-driven vehicles provided observations of chemical and biological ocean parameters, allowing tracking of ecosystem response to the upwelling process. Observations were fed to two oceanographic models, which provided synoptic realization of ocean fields and predicted future conditions. Among the many lessons are an improved knowledge of the scales of variability of upwelling processes, an understanding of how to scale observation systems to these processes, and insights to strategies for adaptive sampling of comparatively rapidly changing processes with comparatively slow vehicles. These lessons are particularly relevant today, given the present emphasis on developing oceanobserving systems.

Figure 7 Example of a distributed observing system using AUVs. This diagram depicts an AOSN deployment in Monterey Bay.


484


See also Gliders. Remotely Operated Vehicles (ROVs).

Further Reading Allmendinger EE (1990) Submersible Vehicle Systems Design. New York: SNAME. Bradley AM (1992) Low power navigation and control for long range autonomous underwater vehicles. Proceedings of the Second International Offshore and Polar Conference, pp. 473–478. Fossen T (1995) Guidance and Control of Ocean Vehicles. New York: Wiley. Griffiths G (ed.) (2003) Technology and Applications of Autonomous Underwater Vehicles. London: Taylor and Francis.

IEEE (2001) Special Issue: Autonomous Ocean Sampling Networks. IEEE Journal of Oceanic Engineering 26(4): 437--446. Jenkins SA, Humphreys DE, Sherman J, et al. (2003) Underwater glider system study. Scripps Institution of Oceanography Technical Report No. 53. Arlington, VA: Office of Naval Research. Rudnick DL and Perry MJ (eds.) (2003) ALPS: Autonomous and Lagrangian Platforms and Sensors, Workshop Report, 64pp. http://www.geo-prose.com/ ALPS (accessed Mar. 2008).

Relevant Website http://www.mbari.org – Monterey Bay 2003 Experiment, Autonomous Ocean Sampling Network, MBARI.


PLATFORMS: BENTHIC FLUX LANDERS R. A. Jahnke, Skidaway Institute of Oceanography, Savannah, GA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction The need to better understand chemical and biological processes and solute transport mechanisms across the deep-sea floor has fueled major engineering advances in seafloor instrumentation in the last few decades. In principle, the kinds of experiments that can be conducted on the seafloor are limited only by researchers’ imaginations. In reality, the majority of the instruments that have been developed seek to accomplish two basic types of operations: deploy, sample, and retrieve benthic flux chamber systems to directly measure seafloor exchange rates and deploy in situ sensor arrays capable of measuring the vertical distribution of solutes in near-surface pore waters from which exchange rates can be estimated based on pore water transport models. Recent advances have expanded the types of experiments and measurements that can be performed by these instruments. Generically, the instrument frames used to deploy these types of instruments have been called ‘bottom landers’. This term was coined in the early 1970s in recognition of the similarity between the approximate shape and function of these devices and the more famous ‘lunar lander’ that carried the first men to the moon. In the following, the design strategies and basic instrumentation for conducting benthic flux chamber incubations and sensor measurements at the deep-sea floor are discussed.

Benthic Flux Measurement Strategies In the 1960s and 1970s, it was increasingly recognized that the temperature and pressure changes that occur in bringing a sediment sample from the deepsea floor to the sea surface may alter the sample. Examples include changes in metabolic rates of benthic populations, changes in pore water concentration gradients from which diffusive fluxes are calculated, and changes in chemical concentrations in pore waters due to pressure- or temperaturedriven reactions. Additional artifacts continue to be identified to this day. Thus, analyzing deep-sea samples brought to the deck of a ship has been increasingly recognized as being not accurate enough

to examine important questions such as: What are the respiration rates of deep-sea-floor populations? What is the seafloor dissolution rate of calcium carbonate? Estimating benthic fluxes (i.e., the net exchange rate of solutes across the sediment surface) on the seafloor, at in situ temperature and pressures, is one strategy for avoiding sampling artifacts and improving the accuracy of deep-ocean measurements. There are two common techniques for estimating benthic fluxes. Benthic flux chamber incubations can be performed in which a known volume of bottom water is trapped above a known area of seafloor. Any solute transported out of the sediments covered by the chamber will be trapped within the chamber waters. Hence, the concentration of this constituent in the chamber waters will increase with incubation time. Conversely, the chamber water concentration of any chemical constituent that is being transported into the sediments will decrease with incubation time. The benthic flux is directly proportional to the rate of concentration increase or decrease, the volume of bottom water enclosed within the chamber, and area of the seafloor covered by the chamber. The main advantage of this method is that there is minimal disturbance to the sediment system, maximizing the probability of accurate estimates. In addition, the fluxes obtained will reflect the net exchange due to all transport processes occurring within the spatial dimensions of the chamber incubations. Thus, this technique will include nondiffusive exchange such as that driven by the active irrigation of organism burrows. A weakness of this approach is that other than inferences about total integrated reaction rates within the sediment column supporting the observed benthic flux, little information is gained about the reaction processes and distributions themselves. Additionally, if there is significant bottom currentdriven advective flow through the surface sediments, differences between hydrodynamic conditions within the chamber and the natural setting may result in significant inaccuracies in flux estimates. The other common strategy for estimating seafloor fluxes is to measure the concentration gradient of chemical constituents very near (preferably across) the sediment–water interface. Knowing the transport processes, the flux can be calculated based on the transport rates and measured concentration gradients. In principle, this method can be applied to all types of transport processes. In practice, however, the exact nature of nondiffusive transport processes is unknown and this calculation strategy is used almost


485

486

PLATFORMS: BENTHIC FLUX LANDERS

exclusively to estimate exchange due to molecular diffusion only. Thus, a limitation of this approach is that accurate benthic fluxes cannot be estimated in a location where pore water exchange due to processes other than molecular diffusion is significant. On the other hand, a strength of using concentration profiles is that the distribution of reactions can be assessed and inferences of reaction mechanisms can be made from variations in the fluxes estimated at different depths below the sediment surface.

Benthic Chamber Landers Design and Operations

The development of benthic chamber landers followed the experiences gained from conducting chamber or ‘bell jar’ incubations in shallow-water areas where they could be tended by scuba divers. The major challenge in developing this instrumentation was simply to automate the required operations to function reliably in the harsh conditions of the deep-sea floor. The earliest instrument that was routinely deployed in the deep sea and has provided a large data set is the free vehicle grab respirometer (FVGR) developed in the late 1970s by Dr. K.L. Smith, Jr., at Scripps Institution of Oceanography in San Diego, California, USA (Figure 1). The basic instrument consists of a structural frame upon which the critical components are mounted. To minimize weight, the frame is most often constructed of tubular aluminum. Instrument flotation is mounted on the upper portion of the instrument frame. The most common type of flotation is glass spheres (as shown in Figure 1), although syntactic foam flotation has also been used. The latter is much more expensive than glass spheres but is not susceptible to implosion, which is critical if manned submersibles are ever required to work in close proximity. Both types provide relatively constant buoyancy. At the top of the flotation section are mounted devices such as a flag, strobe light, radio transmitter, or satellite transmitter to help locate the instrument when it is at the sea surface. Expendable weights are mounted at the lower portion of the tubular frame, usually adjacent to the ‘feet’. As these instruments are free vehicles, these weights provide the negative buoyancy necessary to drive the instrument to the seafloor. A latching mechanism permits the weights to be released at the end of the experiment. Upon release, the flotation provides sufficient positive buoyancy to pull the instrument back to the sea surface where it may be recovered by a surface research vessel. The specific instrument packages, controlling electronics and other operational components, are

mounted in the central part of the frame. It is in these components where the greatest differences between individual designs occur. Chamber instruments have been designed to accommodate one to four chambers simultaneously. While increasing desirable replication, multiple chambers on a single instrument tend to increase variability in the data by decreasing chamber area and increase the complexity of the control, data storage, and sampling systems. Early chamber systems like the FVGR relied on oxygen electrodes to quantify the changes in oxygen within the chambers. Thus, the early instruments were only capable of estimating benthic oxygen demand. While electrodes continue to be widely used, many modern instruments also employ an electronically controlled sampling system so that benthic exchange of many types of solutes, such as nutrients, trace elements, and organic and inorganic carbon, can be assessed. The utility of these instruments is demonstrated by their proliferation. Today, there are more than 20 different research groups that have developed in situ benthic flux chamber instruments. These groups have designed and implemented numerous modifications and alterations to Smith’s initial design, yet the basic characteristics and capabilities of the chamber lander remain the same. Examples of the numerous important modifications that have been made to the early designs include several types of stirring mechanisms and time-series sampling systems so that the exchange of solutes other than oxygen can be addressed. In addition to the basic benthic flux chamber operations, complementary sampling devices have also been added. For example, one chamber instrument design has incorporated an in situ whole core squeezer. This device recovers a sediment core and then sequentially squeezes the surface pore waters from the sediments to provide high-resolution pore water samples. While not appropriate for all solutes due to surface exchange during squeezing, these samples provide important information concerning the pore water gradients of selected metabolites, such as oxygen and nitrate, which greatly enhances the interpretation of the flux chamber results. Examples of results from a benthic flux chamber deployment at c. 3000 m on the continental rise of the eastern US seaboard is displayed in Figure 2. On each plot, the concentration is on the vertical axis and the horizontal axis represents incubation time. These results demonstrate the range of possible responses. Solutes that are taken up by the sediments tend to decrease with incubation time. An example of this is oxygen which is consumed in the sediments through benthic respiration. Most other components are produced in the sediments through the



487

Transducer Radio beacon and strobe light

Glass sphere Flotation

Aluminum tubular frame

Acoustic release Remote release

Battery cylinder

Hydrostatic piston Electronic pressure case Camera Respirometer tray

Strobe

Cable release mechanism Grab tray Descent weight

Frame foot

Disturbance reflector Grab assembly Figure 1 Schematic of the free vehicle grab respirometer developed by Dr. Kenneth L. Smith, Jr. (Scripps Institution of Oceanography).

decomposition or dissolution of biogenic debris. This production supports a flux out of the sediments and the concentrations of these constituents increase with incubation time. Specific examples of these include phosphate, released from degrading organic tissue, total inorganic carbon (TIC) released through respiration and the dissolution of CaCO3, titration alkalinity (TA) produced by the dissolution of CaCO3, and silicate which is produced by the dissolution of opal. Some species, such as nitrate, may increase or decrease with incubation time depending on the relative rates of the competing sedimentary processes

that produce or consume them. For example, nitrate is produced by the oxidation of ammonium (nitrification) and is consumed by denitrifiying bacteria below the oxic zone. Whether there is a flux out of or into the sediments depends on the ratio of these rates. Direct estimates of benthic fluxes can be made from these results from the relationship Benthic flux ¼ ðSVÞ=A where S is the slope of the concentration versus time results, V is the volume of the bottom water trapped



300

19

275

18.5 Nitrate (μM)

Oxygen (μM)

488

250

17.5

225

17

200 20

40

60

80

2350

2450

2325

2425

TA (μEq l−1)

TIC (μM)

0

2300 2275

0

20

40

60

80

0

20

40

60

80

0

20

60

80

2400 2375

2250

2350 0

20

40

60

80

2

40

1.75 Silicate (μM)

Phosphate (μM)

18

1.5 1.25 1

30

20

10 0

20

40 Time (h)

60

80

40 Time (h)

Figure 2 Example benthic flux chamber results from 2927 m on the continental rise of the US eastern seaboard using the chamber system of Dr. Richard A. Jahnke (Skidaway Institute of Oceanography).

in the chamber, and A is the sediment surface enclosed by the chamber. Note that if the chamber has vertical sides, V/A ¼ height of the water column trapped within the chamber. Major Design Controversies and Differences

Despite the numerous advantages of benthic flux incubations for estimating seafloor exchange, there are a variety of limitations and concerns that need to be addressed in evaluating results from chamber incubations and in assessing the relative merits of the different instrument designs. As shown in the example results, concentration changes are required to evaluate the solute exchange rate. However, the concentration changes will also

alter the near-surface gradients, altering the flux. In the extreme case, where concentration changes are large, such as complete oxygen depletion within the chamber, changes in chamber water chemistry may alter near-surface chemical reactions and/or exchange processes, greatly altering the chemical flux. To minimize this source of uncertainty, chamber deployments must be designed to minimize the concentration changes and maintain the chamber sediments as near to their natural state as possible. Thus, it is important that the most precise analytical procedures be employed, so that accurate fluxes can be quantified without large concentration changes. Maintaining natural benthic fluxes requires that the sediment surface not be disturbed during chamber deployment. This need has resulted in several different types of deployment strategies and



instrument designs. Most instruments have been designed to settle to the seafloor and then, after some preset time period, slowly insert the chambers into the bottom, assuming that this waiting period will allow the sediments that are resuspended by the impact of the instrument with the bottom to be swept away, leaving a natural surface on which to conduct the incubation. Of course, if the pressure disturbance ahead of the descending instrument is sufficiently large, the entire sediment surface within the perimeter of the instrument frame will be impacted. Recognizing this potential artifact, some instrument designs positioned the chambers lower than the main frame, so that the chamber would encapsulate the sediments upon which the flux incubation will be performed before the ‘bow wave’ from the frame reaches the surface. Another approach is to suspend a descent weight below the instrument. This weight provides the negative buoyancy needed for the instrument to descend to the seafloor. Once the weight reaches the seafloor, the positive buoyancy of the instrument itself causes it to remain suspended above the bottom. A winch system is then activated and the instrument package is very slowly pulled to the seafloor. A recent innovative approach for minimizing bottom disturbance has been implemented on the ROVER lander, a benthic flux chamber device capable of ‘crawling’ around the seafloor. Thus, once the instrument has impacted the bottom, it simply crawls laterally to an undisturbed location prior to conducting benthic flux chamber incubations and pore water profiling operations. Perhaps the most controversial aspect of benthic flux incubations and instrument design is focused on mechanisms and requirements for simulating natural flow conditions within the chamber. Exchange across a homogeneous muddy sediment surface requires transport across the hydrodynamic boundary layer including the molecular diffusive sublayer. Hydrodynamic conditions within the chamber control the thickness of this layer. Changes in the hydrodynamic regime will alter the thickness of the layer and, at least temporarily, will alter the exchange rate. However, since the benthic flux is supported by metabolic and chemical reactions in the sediments that are not influenced by the diffusive sublayer thickness, the seafloor exchange rate will eventually revert to its natural rate. Thus, the need to accurately reproduce the natural hydrodynamic conditions with the benthic flux chamber depends critically upon the response time of the surface pore water gradients and the length of the chamber incubations. If controlled by molecular diffusion, the time required for pore water concentration gradients to recover after disturbance scales as the square of the thickness of the

489

disturbed layer. Layers of 1- and 3-mm thickness will recover on timescales of c. 10 and 80 min, respectively. If deployments are short relative to the gradient response times, accurate fluxes will require the maintenance of near-natural hydrodynamic conditions within the chamber. On the other hand, if the deployments are long relative to the pore water gradient response times, the fluxes will be insensitive to chamber hydrodynamics and a simple chamber water-stirring mechanism, such as rotating rods, a disk, or circulation through an external pump, is adequate. For most solutes in the deep sea, deployments greater than 10–20 h are sufficient to minimize artifacts due to changes in the diffusive sublayer thickness and thus only require a relatively simple mixing strategy. The situation is more complex if the sediments have many open burrows or exhibit elevated permeability, as would be characteristic of a well-sorted sandy sediment. In these environments, bottom flows may drive advective pore water exchange. Alternation of natural flow conditions by the presence of the chamber will likely influence the accuracy of the flux estimates and must be considered in interpreting results.

Sensor Landers Design and Operations

The basic instrument design and field deployment operations of the sensor landers are the same as that already discussed for the benthic flux chamber. The instrument consists of a frame, identical to that of a benthic flux chamber lander, with flotation mounted at the top and expendable descent weights attached near the feet. The major difference is that the benthic chamber is replaced with an instrument package capable of inserting microelectrodes or optode with high vertical resolution into the surface sediments. An example of the basic instrument package is presented in Figure 3. Each of these types of packages consists of three basic components: the sensing electrodes themselves; a mechanism for moving the electrodes vertically across the sediment–water interface; and the data storage and controlling electronics. In the design pictured in Figure 3, the sensing electrodes are attached directly to the main pressure case. Located within the pressure case are the power and electronics necessary to operate the motor and electrodes and to provide for data storage and retrieval. The motor drive system is positioned outside the case to move the case vertically, so that profiles of the measured components are obtained. The vertical resolution of the profiles is controlled by the size of


490


Threaded rod DC motor housing

Supporting AL-frame and tracks

Radial ball bearings Oil-filled bladder Pressure cylinder

Microelectrodes

Figure 3 Schematic of the microelectrode deployment apparatus developed by Dr. Clare Reimers (Rutgers University).

the electrode tip and the precision of the motor drive assembly. Early designs primarily employed oxygen and electrical resistivity electrodes. The latter is required to assess the porosity and tortuosity that influence the rate of diffusion in sediments. Porosity is a measure of the proportion of the sediment comprised of void space between sediment particles and is generally 70– 90% in muds and 30–60% in sands. Tortuosity is the actual distance a solute would have to travel around sediment particles relative to the straight-line distance between two points. In recent years, numerous other sensors have been developed and implemented for deep-sea lander use. These include pH, pCO2, total CO2, calcium, ammonium, and nitrate. It is anticipated that the development of other sensors will continue and the types of measurements possible in the future will continue to expand. An example of oxygen, resistivity (formation factor), and pH results obtained from a sensor lander

deployment is provided in Figure 4. Because oxygen is consumed within the sediments, primarily due to the respiration by benthic organisms but possibly also due to chemical consumption, oxygen concentrations decrease with increasing depth in the sediments. This downward concentration gradient implies a benthic flux into the sediment. Contrastingly, in the example provided in Figure 4, pH first increases and then decreases with sediment depth. This more complicated profile shape is due to competing reactions downcore. In this example, pH first increases with depth due to the dissolution of calcium carbonate and then decreases with depth due to continued production of carbon dioxide. Formation factor values increase with sediment depth due to the compaction of sediment particles. This decrease in resistivity implies a decrease in sediment porosity and increase in sediment tortuosity, both of which tend to decrease the effective diffusion coefficient with depth in the sediments.



491

Station G, 4685 m

0.5

1

1.5

2

2.5

3

Oxygen (μmol kg−1)

80 100 120 140 160 180 200 220 240 260

(b)

Formation factor

(c) −0.06

∇

(a)

pH

−0.02 0 0.02

0.06

−1 0 1

Depth (cm)

2 3 4 5 6 7 8 9 10 Figure 4 Examples of deep-sea-floor microlectrode oxygen, pH, and resistivity results obtained by Dr. Burke Hales (Oregon State University). Different symbols display simultaneous results from separate, individual electrodes installed a few centimeters apart on the instrument.

These results can be used to estimate the diffusive flux across the sediment–water interface and to evaluate the reaction rate within the measured depth interval. The benthic flux can be estimated from Fick’s first law corrected for porosity and the effects of sediment particles on diffusion rates as shown below: Benthic flux ¼ fDs dC=dz where f ¼ surface porosity, Ds ¼ effective sediment diffusion coefficient, and dC/dz ¼ vertical concentration gradient at the sediment–water interface. Porosity near the sediment surface is generally estimated from a regression of the measured resistivity and directly measured porosities at wider-spaced depth intervals. The latter are generally determined from the weight loss upon drying bulk sediments. The effective diffusion coefficient can be estimated by dividing the molecular diffusion coefficient by the porosity and tortuosity (i.e., Ds ¼ D/(fy), where D ¼ molecular diffusion in free solution, and y ¼ tortuosity). This expression requires independent measurements of porosity and tortuosity at the same vertical scale as the concentration profile. Such measurements are often not available. A common alternate but less-accurate strategy in fine-grained sediments is to approximate the effective diffusion coefficient as the molecular diffusion coefficient times the square of the porosity. This relationship has been empirically derived from numerous individual studies.

It is important to note that the flux equation shown above is not limited to the sediment surface but rather can be used to evaluate the diffusive flux at any depth horizon within the profile. Thus, it is often useful to define a sediment layer of a small thickness and calculate the diffusive fluxes at the top and bottom of the layer. Assuming that horizontal diffusive exchange can be neglected and that the profile is in steady state, the difference in these fluxes is a measure of the net production or consumption rate of the measured solute within the layer. Thus, by interpreting the vertical variations in the flux, one can evaluate the distributions of reactions in the sediments and potentially make inferences about the processes and mechanisms controlling solute diagenesis and benthic flux. Advantages, Limitations, and Design Concerns

Unlike benthic flux chambers that require a significant incubation time, sensor landers can obtain a profile relatively quickly, usually within 1–2 h. The exact required time is determined by the number of sampling depths and the response time of the sensors employed. Thus, sensor landers can be used to obtain in situ profiles relatively rapidly and estimate diffusive fluxes at the deep-sea floor. There are also several limitations to this approach. Because the measurements are generally made within several hours of the instrument reaching the seafloor,


492


the accuracy of the flux estimate can be severely compromised by physical disturbance of the sediment surface. Potential disturbance can be caused by the instrument itself or by the bow wave that precedes the instrument as it settles. Video recordings of these instruments settling onto the bottom reveal significant resuspension of surface sediments. While the resuspended sediments are generally allowed to settle back to the sea floor or be advected away prior to making measurements, the effect of this disturbance on the profiles is still a concern. The only instrument to completely avoid this potential problem is the ROVER lander (discussed in the next section). Maintaining natural hydrodynamic conditions within the benthic boundary layer is also critical to the accuracy of microelectrode flux estimates. Unlike benthic flux chamber incubations that extend over time intervals sufficient to return to initial conditions if diffusive boundary layer thicknesses are altered, electrode measurements are rapid and would record transient conditions if made directly after altering bottom hydrodynamic conditions. Because the instrument frame and electrodes themselves may alter bottom flow, such changes are a concern. Since the electrode sensing tips are very small (generally 5–20 mm in diameter) as required to achieve fine vertical resolution, they also respond to horizontal

variations that may be caused by burrowing organisms or physical inhomogeneities. Because the geochemical questions being asked often require knowledge of the mean benthic flux for a known area or region, numerous replicate profiles are often required to estimate the average profile and benthic flux.

Special Landers and Lander-based Research Strategies In addition to the basic landers discussed above, a variety of instruments have been developed in the last decade for special purposes and it is likely that new types will continue to be developed in the future. For example, simple benthic flux chamber landers have been developed to measure advective pore water flows around hydrothermal vent and midocean ridge systems. In these types of chambers, osmotic pumps are used to continuously add a tracer and remove a sample. Pore water advection rates as low as 0.1 mm yr 1 can be measured with this system. For seafloor microbial studies, a lander has been developed capable of injecting radiolabeled tracers continuously throughout the upper 70 cm of the sediment column. The sediments surrounding the line of injection are cored and recovered at the end of a preset incubation period and returned to the ship

Location flag Time-lapse camera

Flotation sphere

Strobe

Microprofiler

Bumper

Benthic chamber

Roller Treads

Figure 5 The ROVER lander developed by Dr. Kenneth L. Smith, Jr. (Scripps Institution of Oceanography).



for analysis. Other lander instruments are being developed to address specific research questions and to operate in particularly demanding environments. For example, the benthic exchange of certain trace metals is sensitive to the redox conditions of the sediment–water interface. To accurately measure benthic fluxes of these metals, the ‘oxystat’ lander has been developed in which the chamber waters are circulated through gaspermeable tubing in contact with bottom waters. Exchange of oxygen across the tubing wall keeps the oxygen concentrations in the chambers at near-natural levels, minimizing potential changes to the metal fluxes. Another lander device is the ‘integrated sediment disturbing lander’, in which a set of small plow blades is rotated through the surface sediments within the chamber to evaluate the impact of sediment disturbance, such as would occur following a bottom trawl or large storm, on benthic solute exchange. Perhaps the most innovative lander recently developed is the ROVER (Figure 5). ROVER is capable of performing repeated benthic flux chamber incubations and microelectrode profiling measurements. Most importantly, after a measurement cycle is complete, the instrument uses a tractor-tread propulsion system to move c. 5 m, so that the next measurement is performed on a natural, undisturbed surface. Since this instrument is crawling laterally on the sediment surface, it eliminates the potential disturbances discussed for free-fall instruments. ROVER was constructed to examine temporal variations in seafloor fluxes and is capable of performing duplicate benthic flux chamber and microelectrode profiling measurements at 30 individual sites over a 6-month period on a single deployment. There is currently a significant effort underway to develop and install seafloor cabled observatories. It is likely that the ROVER-type of instrumented systems will in the future be attached to seafloor cable nodes and observatories and provide long-term, near-real-time observations. Finally, an eddy correlation instrument is being developed in which the benthic oxygen flux is assessed by simultaneously directly measuring the concentration and velocity normal to the seafloor of a small volume of water above the sediment surface by using an acoustic Doppler velocimeter and an oxygen microelectrode in concert. By integrating, for a sufficient length of time, the instantaneous fluxes toward and away from the sediments, the net benthic flux can be estimated.

experiments on the seafloor. These capabilities have greatly improved our understanding of the benthic processes, especially at abyssal depths where pressure- and temperature-driven artifacts have hindered earlier studies.

See also Deep-Sea Sediment Drifts. Electrical Properties of Sea Water. Grabs for Shelf Benthic Sampling. Mineral Extraction, Authigenic Minerals. Pore Water Chemistry. Sensors for Micrometeorological and Flux Measurements. Temporal Variability of Particle Flux.

Further Reading Berelson WM, Hammond DE, Smith KL, Jr., et al. (1987) In situ benthic flux measurement devices: Bottom lander technology. MTS Journal 21: 26--32. Berg P, Roy H, Janssen F, et al. (2003) Oxygen uptake by aquatic sediments with a novel non-invasive eddycorrelation technique. Marine Ecology Progress Series 261: 75--83. Greeff O, Glud RN, Gundersen J, Holby O, and Jorgensen BB (1998) A benthic lander for tracer studies in the sea bed: In situ measurements of sulfate reduction. Continental Shelf Research 18: 1581--1594. Jahnke RA and Christiansen MB (1989) A free-vehicle benthic chamber instrument for sea floor studies. DeepSea Research 36: 625--637. Reimers CE (1987) An in situ microprofiling instrument for measuring interfacial pore water gradients: Methods and oxygen profiles from the North Pacific Ocean. Deep-Sea Research 34: 2019--2035. Sayles FL and Dickinson WH (1991) The ROLAI2D lander: A benthic lander for the study of exchange across the sediment–water interface. Deep-Sea Research 38: 505--529. Smith KL, Jr., Glatts RC, Baldwin RJ, et al. (1997) An autonomous bottom-transecting vehicle for making long time-series measurements of sediment community oxygen consumption to abyssal depths. Limnology and Oceanography 42: 1601--1612. Tengberg A, De Bovee F, Hall P, et al. (1995) Benthic chamber and profiling landers in oceanography – a review of design, technical solutions and functioning. Progress in Oceanography 35: 253--294. Tengberg A, Stahl H, Gust G, et al. (2004) Intercalibration of benthic flux chambers. Part I: Accuracy of flux measurements and influence of chamber hydrodynamics. Progress in Oceanography 60: 1--28.

Conclusions In conclusion, bottom lander technology developed in the last several decades now permits a wide variety of remote sampling procedures and incubation

493

Relevant Website http://www.cobo.org.uk – Coastal Ocean Benthic Observatory.


PLATINUM GROUP ELEMENTS AND THEIR ISOTOPES IN THE OCEAN G. E. Ravizza, Woods Hole Oceanographic Institution, Woods Hole, MA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2208–2217, & 2001, Elsevier Ltd.

investigation. These areas are (1) sea water Os isotope geochemistry; (2) Ir and the other PGEs as tracers of extraterrestrial material in marine sediments; and (3) anthropogenic release of PGEs to the marine environment.

The Apparent Rarity of the PGEs in Sea Water The platinum group elements (PGEs) include three Introduction

second-series transition metals, ruthenium (Ru), rhodium (Rh) palladium (Pd), and three third-series transition metals, osmium (Os), iridium (Ir) and platinum (Pt). The marine chemistry of this group of elements is the least understood and most poorly documented among the many elements in the periodic table. During the 1970s and 1980s when the attention of the marine chemistry community was focused on characterizing the distribution of all the elements in the ocean, direct measurement of the PGEs in sea water was, for the most part, beyond the reach of available analytical methods. Indeed, the most important attribute that links the marine chemistry of all the PGEs is their very low concentration in sea water. This group of metals accounts for 6 of the 10 least abundant elements in sea water. This article has three objectives. First, to explain how the dissolved inventories of PGEs in the ocean are maintained at such low concentrations relative to most other elements. Second, to review the current status of our knowledge regarding the vertical distribution of these metals in the oceanic water column. Third, to present a brief overview of areas of marine PGE research that are the focus present research activity, and are likely to motivate future Table 1

The underlying reason for the very low concentrations of all the PGEs in sea water has little to do with the aqueous chemistry of these elements. Rather it is the chemical partitioning of these elements within the deep earth that explains their relative scarcity in the sea water (Table 1). Comparison of the PGE content of meteoritic material, believed to represent the primordial material that constituted the undifferentiated earth, to ultramafic rocks, believed to be representative of the deep silicate earth, reveals a nearly uniform 100-fold depletion of the PGEs in the deep silicate earth. This depletion indicates that roughly 99% of the whole earth PGE inventory is sequestered within the earth’s metallic core. Further comparison of the PGE concentrations of ultramafic rocks to estimated PGE concentrations of upper crustal rocks shows an additional more variable depletion of the PGEs in rocks exposed at the earth’s surface relative to the deep silicate earth. This concentration contrast results from the fact that the PGEs are retained in the solid residue when the deep earth is melted to form the earth’s crust. The net effect of the strong affinity of the PGEs for phases that reside in the deep earth is a strong depletion of the PGEs in the rocks typically exposed at the earth’s

Representative PGE concentrations in important earth reservoirs and their ratios

Chondrites (ppb)a Silicate earth (ppb)b Upper crust (ppb)b Sea water (pg kg1) Sea water (fmol kg1) Sea water/crustc Log (Seawater/crust)

Ru

Rh

Pd

Os

Ir

Pt

710 5 1.1 2 20 2.00 106 5.7

130 0.9 0.38 100 100 0.00026 3.5

550 3.9 2 60 550 3.00 105 4.5

490 3.4 0.04 10 50 0.00025 3.6

455 3.2 0.04 0.1 0.5 3.00 106 5.5

1010 7.1 1.5 50 260 3.30 105 4.5

a

Values from McDonough and Sun (1995) Chem. Geol. 120 p. 223. Values from Schmidt et al. (1997) Geochim. Cosmochim. Act. 61 p. 2977. c Seawater/crust ¼ (row 4)/(row 3) as a dimensionless ratio. b

494


PLATINUM GROUP ELEMENTS AND THEIR ISOTOPES IN THE OCEAN

Very insoluble

Very soluble

Sn Pb Pr Zr Y La Dy

Si

Nd Ho

Au Pd

Sm Er

Cu Pt

Eu Tm

Ir Tl Ag

Rb Rh U I

Na

Sc Co Yb Ru Ge Ba W

Os F Li Ca

S

Ce Gd Lu Ga Cr V P

Cd As Mo K

Mg Cl

Fe Th Al Be Tb Mn In Zn Ni Cs Bi Sb Se Re Sr

B Br

Ti

_9

_8

_7

_6

_5

_4

_3

_2

_1

0

Log (average deep water concentration/average upper crustal concentration)

Figure 1 Histogram of seawater/upper crust partition coefficients. This parameter qualitatively represents gross patterns in elemental solubility across the periodic table. Note that the PGEs are neither particularly insoluble compared to other elements, nor particularly similar to one another. Data are derived from compilations by Taylor and McLennan (1985) The continental crust, its composition and evolution. Blackwell Scientific; Nozaki (1997) EOS v. 78, p. 221 and the PGE compilation in Table 1.

surface. In simplest terms the concentrations of the PGEs in sea water are low compared to other elements because there is a relatively smaller inventory of these elements in the earth’s surficial environment. In the context of the marine chemistry of the PGEs it is significant that the conceptual basis for considering these elements as a coherent group is more closely linked to their behavior in the deep earth than to their behavior in the ocean. Once the very low average crustal concentrations of the PGEs is taken into account it is clear that as a group these elements are not extremely insoluble in sea water (Figure 1). More importantly it also becomes apparent that there are obvious first order differences in the solubility of the PGEs. The striking contrast between the solid–solution partitioning of Os and Ir, two elements that exhibit very similar behavior in the deep earth, illustrates this clearly. The important general point here is that although the PGEs are often referred collectively, in a manner analogous to the rare earth elements, the PGEs do not exhibit systematic variations in charge or ionic radius that give rise to systematic similarities or differences in their marine chemistry. Although the different PGEs do not share a coherent set of chemical affinities in the marine environment, the fact that all these elements occur in sea water at very low concentrations has two important implications that extend to all six elements in

495

this group. The first and most obvious is that quantifying PGE distributions in sea water is analytically very challenging. Consequently, the cumulative set of published data constraining the water column distribution of these elements is quite small. Moreover, when different methodologies have been employed to measure the same element the results frequently disagree. The second implication of the low concentrations of the PGEs in sea water relates to understanding of the speciation of these metals in sea water. It is now generally accepted that many of the first series transition metals are strongly complexed in surface sea water by organic ligands that occur at nanomolar concentrations. Given that concentrations of the PGEs in sea water are 103–106 times lower than ligand concentrations it seems likely that the marine chemistry of the PGEs will also be strongly influenced by these ligands. Although this is largely a matter of speculation, the simple conceptual point is that a very complete description of all the more abundant species and their affinities for the various PGEs would be required to approach PGE speciation theoretically. Such a detailed description of sea water chemistry is unavailable, and as a result the true speciation of the PGEs in sea water is largely unconstrained. Finally it is noteworthy that the theoretical uncertainties regarding the speciation of the PGEs and the practical problems associated with the analysis of these elements in sea water are interrelated. A sound knowledge of the chemical form(s) of an element in sea water greatly facilitates the development of reliable methods for its separation and quantification.

Overview of Water Column PGE Data Ruthenium (Ru)

There is little that can be said about the water column distribution of Ru because there are so few data available. The data that are available are limited to isolated analyses of surface sea water. The most recent work reports 20 fmol kg1 for analysis of surface waters from the SIO (Scripps Institute of Oceanography) pier in southern California. Although this value seems reasonable given the relatively low concentration of Ru in upper crustal rocks (Table 1), there is no means of further evaluating the accuracy of this particular analysis, or of assessing how representative this single analysis is of the ocean in general. The likely valence of Ru in sea water is as Ru(IV) however, this assessment is based on very scant data for the stability constants for Ru in water. Ru enrichment in ferromanganese crusts has been interpreted as evidence that Ru is redox active in the


496


marine environment, being subject to oxidation to an insoluble form that is coprecipitated with ferromanganese oxides. However, additional work is required to establish with any degree of certainty that this element is in fact redox active in the marine environment. Rhodium (Rh)

The water column distribution of Rh has been investigated in only a single study. In analytical terms Rh is perhaps the most difficult of the PGEs to quantify because Rh has only one stable isotope, 103 Rh. Therefore, the efficiency of Rh preconcentration from sea water must be monitored using a short-lived radiotracer. This methodology was used to generate a full vertical profile from the eastern North Pacific (Figure 2). The significant features of this profile are the clear surface water depletion of Rh and the relatively large concentrations (approximately 100 fmol kg1) of Rh in deep waters. Eastern Pacific (34˚N 122˚W) 0

1

Depth (km)

2

Palladium (Pd)

Our knowledge of the distribution of Pd in the water column is based on a study by Lee in 1983, the first to report a full vertical profile of any PGE in sea water. Vertical profiles of filtered and unfiltered samples from two different stations in the Pacific both show a systematic increase in concentration with increasing depth in the water column. The pattern of depth variation closely mimics that of Ni in the same samples. This similarity in the vertical distribution of these two metals and their similar upper crustal partition co-efficients (Figure 1) have been rationalized in terms of similar outer electron configuration. Both metals are believed to be stable in their divalent form in sea water. Subsequent more detailed study of the marine chemistry of Ni demonstrates that organic complexation plays an important role in Ni speciation, and laboratory experiments show the same can be true for Pd. Thus it seems likely that complexation by organic ligands plays an important role in Pd speciation in sea water. Osmium (Os)

3

4 August November 5

Distributions of this type are traditionally interpreted as the result of particulate scavenging in surface waters followed by remineralization at depth. Type of distribution contrasts strongly with that of Co, the first series transition metal which is located directly above Rh in the periodic table, illustrates that elements from the same group in the periodic table can exhibit very different chemical behavior. The contrasting behavior of Co and Rh is potentially related to the fact the Co has an active redox chemistry in the marine environment whereas Rh is believed to be stable only as Rh(III) complexes. It is unclear why the upper crustal partition coefficient calculated for Rh is so large (Figure 1); by analogy to other trivalent metals a much lower value would be expected.

0

0.5 Rh (pmol kg

1

1.5

_1)

Figure 2 Profile of dissolved Rh in sea water. Data are from Bertine et al. (1993) Marine Chemistry 42: 199.

Until very recently there were no data available reporting the concentration of Os in sea water. Since 1996, however, there have been several independent studies that focused on this problem making Os the PGE whose marine chemistry has been most extensively studied. Although the recent studies agree that deep water Os concentration is roughly 50–60 fmol kg1 (Table 2), the vertical distribution of Os in the water column is still open to debate (Figure 3). Results from analyses of samples from the Indian Ocean led to the conclusion that Os behaves conservatively in sea water. A separate study in the Eastern Tropical North Pacific reported a 30% depletion in Os concentration within the core of the oxygen minimum



Table 2 Comparison of Os concentrations of fully oxic deep water from different studies Os (fmol kg 1) Indian Oceana Eastern Pacifica North Pacificb

55–59 44–52 53–55

a

See Figure 3. b Sharma et al. (2000) Earth Planet. Sci. Lett. 179, p. 139.

zone, and interpreted this as evidence that Os was subject to removal from sea water under reducing conditions. Although it is widely believed that Os is redox active is sea water, as first indicated by the strong enrichment of Os in anoxic marine sediments, detailed knowledge of Os speciation in sea water does not exist. Inorganic speciation calculations considering the major ions in sea water indicate that in fully oxic sea water Os should be stable in it highest valence, Os(VIII), and exist as an oxyanion. As mentioned above in the case of Ru, the paucity of data constraining the stability constants for potential Os ligands in sea water precludes any rigorous assessment of the likely redox state of Os in sea water. Some working on separation of Os from sea water have suggested that Os is strongly complexed by organic ligands in sea water. The fact that Os(VIII) is highly reactive toward many organic compounds, and is subject to reduction by them in the laboratory, lends some credibility to this inference. Iridium (Ir)

Among the stable elements that have been measured in sea water Ir is the least abundant, with concentrations on the order of 1 fmol kg1. Although a full vertical profile from the open ocean is not available, a vertical profile from the Baltic Sea has been reported (Figure 4). These data provide compelling evidence that Ir, unlike Os, is not subject to enhanced removal from solution under reducing conditions. Rather, these data are suggestive of Ir scavenging in Baltic surface waters, likely by Fe- and Mn-oxyhydroxides, and subsequent release in deeper anoxic waters. Ir(III) is likely to be the stable valence of Ir in sea water. Given that Ir and Rh are believed to exist in the þ 3 valence, have a d6 and electron configuration, and reside in the same group in the periodic table, it is surprising that the apparent crustal partition coefficients for these two elements differ so dramatically (Table 1). This inconsistency suggests either that this simplistic view of the speciation of these metals is incorrect, or that there is a large

497

systematic error in the available concentration data for Rh or Ir. The former seems more likely than the later, and Ir removal from sea water via oxidation to an insoluble form of Ir(IV) has been proposed in previous discussions of the marine chemistry of Ir. Platinum (Pt)

Though relatively little work has been done on the water column distribution of Pt in recent years, several studies were conducted from the mid-1980s to the early 1990s. As is the case for Os, only a general consensus regarding deep-water concentrations was achieved, constraining values to fall between 1 and 0.3 pmol kg1. The depth variations reported in each of the three separate studies differed (Figure 5). Because each of these three studies employed different analytical methodologies, it is unclear to what extent the contrasting vertical profiles reflect true variability among the various ocean basins. Its seems unlikely that the strong near-surface Pt enrichment present in the Indian Ocean profile would be restricted to this ocean basin, or that deep water Pt concentrations would exhibit a fourfold difference between eastern and western Pacific. Though the differing vertical profiles suggest that some of the available sea water Pt data are subject to analytical artifact, it is uncertain which data are most reliable. The uncertainties regarding the vertical distribution of Pt are mirrored in our understanding of the chemical form of Pt in sea water. There is agreement among different workers that the two relevant valances of Pt are Pt(II) and Pt(IV). Some workers argue that Pt(II), stabilized by strong chloro-complexes, is the primary form of Pt in sea water, whereas others argue that Pt(II) is only significant in surface waters and that Pt(IV) dominates in oxic deep water. This author believes these types of inferences must be regarded as largely speculative because the relevant complexing ligands are unknown and consequently appropriate redox potentials cannot be prescribed. Moreover marine chemistry is replete with examples of persistent disequilibrium and the slow kinetics of ligand exchange are a persistent theme in discussions of the aqueous chemistry of Pt. Consequently even if the required thermodynamic data were available for Pt and the other PGEs, they would not necessarily inform us of the true speciation of these metals in sea water.

Topics of Special Interest in Marine PGE Research Os Isotope Geochemistry

The isotopic composition of Os in natural materials varies as a result of the decay of two long-lived



0

150 _ Oxygen ( µmol kg 1)

100 200

250

35

45 _ Os (fmol kg 1)

E. Pacific

S. Indian (37°S 57°E)

50

S. Indian

E. Pacific (9°N 104°W)

.

55

0.85

0.95

187Os

1.05

1.15 / 188Os

1.25

Figure 3 Vertical profiles of Os concentration and 187Os/188Os ratio in sea water from the Indian Ocean and the eastern tropical North Pacific. The Pacific data (Woodhouse et al. 1999 EPSL 173: 223) include analyses of both filtered and unfiltered samples; no systematic difference between the two is apparent. Note that the low Os concentrations in the Pacific profile coincide with the core of a very strong oxygen minimum zone. Indian Ocean data (Levassuer et al. 1998 Science 282: 272) indicate that Os behaves conservatively.

5

4

3

2

1

0

Depth (km)

498 PLATINUM GROUP ELEMENTS AND THEIR ISOTOPES IN THE OCEAN


50

Depth (m)

187

Range of measured concentration in the open ocean (n= 5)

0

100

150

200

250 0

499

1

2

3

4

5

6

7

Ir (fmol kg_1) Figure 4 Dissolved Ir profile from the Baltic Sea plotted with the range of Ir concentrations reported for analyses of open ocean samples. Baltic Sea samples were filtered prior to acidification, open-ocean data were acidified and unfiltered. The abrupt increase in dissolved Ir at 150 m depth in the Baltic Sea profile coincides with complete depletion of dissolved oxygen. Anbar et al. (1996) Science 273: 1524.

naturally occurring radionuclides. The decay of 187 Re produces 187Os and the decay of 190Pt produces 186Os; changes in Os isotopic composition that arise from these decay schemes are commonly reported as variations in 187Os/188Os and 186Os/188Os, respectively (Table 3). The long half-life and low isotopic abundance of 190Pt restricts the range of 186 Os/188Os variations in most natural materials, making these isotopic analyses extremely challenging. Significant 186Os/188Os variability in marine deposits has yet to be documented. However, available Pt and Os concentration data from metalliferous sediments and marine manganese nodules demonstrate that these deposits have Pt/Os ratios among the highest measured in terrestrial materials. These data suggest that the Pt-Os decay scheme may be exploited in the future as a tool for dating these deposits, and provide the impetus for further investigating the geochemical processes that are responsible for producing the large Pt/Os ratio variation observed in marine deposits. In contrast to the Pt-Os decay scheme, the Re-Os system gives rise to large variations in 187Os/188Os of marine deposits and is currently the subject of vigorous investigation. This work is motivated by two fundamentally important attributes of Os geochemistry; the relatively short marine residence time of Os, and the record of past variations in the 187 Os/188Os of sea water preserved in marine sediments. Direct analyses of sea water do not yield evidence of any resolvable difference in the

Os/188Os ratio between different ocean basins, but higher precision analyses of Mn crust surfaces do suggest that the 187Os/188Os of the Atlantic ocean may be slightly larger than in the Indian or Pacific basins. This isotopic contrast is extremely small compared to the large range in 187Os/188Os of sources of Os supplied to the ocean (Table 4). The nearly homogeneous character of modern sea water relative to oceanic inputs implies that the marine residence time of Os is poised close to the mixing time of the oceans. Spatial variations in 187Os/188Os of modern sea water are also small compared to the record of temporal variations in sea water 187 Os/188Os preserved in marine sediments. Past variations in the 187Os/188Os of sea water provide a globally integrated record of Os input to ocean that can be exploited to make inferences about the geologic history of chemical weathering, and to identify extraterrestrial impacts in the sedimentary record. Detailed discussion of the marine Os isotope record is beyond the scope of this review. The PGEs as Tracers of Extraterrestrial Material in Marine Sediments

Very large concentrations of the PGEs in extraterrestrial material relative to the average upper crustal material (Table 1) make the PGEs valuable indicators of the presence of particulate extraterrestrial material in marine sediments. For example, addition of 0.01% by weight chondritic material to a sediment with average crustal Ir and Os concentrations would roughly double the concentrations of these elements in the mixture relative to that of the starting material. Ir is more widely exploited than Os as a tracer of particulate extraterrestrial material because methods for low level Ir analysis were established earlier and are more widely available. The global Ir enrichment at the Cretaceous–Tertiary boundary, and the subsequent identification of a major extraterrestrial impact crater, provide the best known example of this type of research. Ir data have been applied in a similar manner to study numerous other event horizons in the geologic record. Other PGE analyses can be integrated into these studies to provide additional constraints on PGE source. For example, the Pt/Ir ratios typical of upper crustal material are roughly 10 times larger than in chondrites (Table 1). This type of contrast in element ratios can be used to help evaluate whether elevated Ir concentrations are truly related to an extraterrestrial PGE source, or are the result of natural enrichment of PGEs from the ambient environment. Many studies motivated by the controversy surrounding the interpretation of the Ir anomaly at the



5

4

3

2

1

0

0

0.4

1.6

35° N 122° W

35° N 138° W

33° N 139° W

1.2 0.8 Pt (pmol kg_1)

Eastern Pacific

0

0.5

1.5

06° S 51° E

27° S 56° E

1 Pt (pmol kg_1)

Indian Ocean

0

0.4

0.8 1.2 _ Pt (pmol kg 1)

1.6

24° N 165° E

26° N 33° W

32° N 64° W

Atlantic and Western Pacific

Figure 5 Dissolved Pt profiles from the Pacific, Indian and Atlantic Oceans. Although deep-water concentrations are similar to one another the vertical distribution of Pt differs dramatically among the various profiles. Data are from the following sources. Eastern Pacific: Hodge et al. (1986) Analytical Chemistry 58, p. 616; Indian: Jacinto and van den Berg (1992) Nature 338, p. 332; Atlantic and western Pacific: Colodner et al. (1993) Analytical Chemistry 65, p. 419 and Colodner (1991) The marine geochemistry of rhenium, platinum and iridium. Ph. D. thesis. MIT/ WHOI Joint Program in Oceanography.

Depth (km)

500 PLATINUM GROUP ELEMENTS AND THEIR ISOTOPES IN THE OCEAN

501


Abundance

Half-life

Daughter

187

62.6% 0.0124%

42 billion years 449 billion years

187

Re 190 Pt

Os 186 Os

400 Co (μg g_1)

Parent

3

Co Ir

2.5 2

300 1.5 200

1

100 Compiled from walker et al. 1997 Geochim. Cosmochim. Act. 61, p. 4799.

0.5

0

0 0

5

10 15 Depth (m)

(A)

12.0 Ir (ng g_1)

1.04–1.07 1.00–1.04 1.00–1.04 0.64–2.94 0.11–0.39 0.12–0.14 0.2– 1.06

_ Average Upper Crustal Co _ 10 μg g _1 Average Upper Crustal Ir _ 0.05 ng g 1

10.0

187

Atlantic Mn crust surfacesa Indian Mn crust surfacesa Pacific Mn crust surfacesa Riversb Hydrothermal fluidsc Meteoritic material Cenozoic sea waterd

20

25

(Ir/Co)sw = 0.11 ng μg_1

Table 4 Comparison of 187Os/188Os ranges among ocean basins, sources of Os to sea water and Cenozoic sea water Os/188Os

Ir (ng g_1)

500

Table 3 Radioactive decay schemes that influence the isotopic composition of naturally occurring Os

12.0 10.0

8.0

8.0

6.0

6.0

4.0

Ir/Co = 0.0035 ng μg_1

2.0

4.0 2.0

0 0

100

(B)

200

300 Co (μg g_1)

400

0 500

a

Burton et al. (1999) Earth Planet. Sci. Lett. 171, p. 185. b Levasseur et al. (1999) Earth Planet. Sci. Lett. 174, p. 7. c Sharma et al. (2000) Earth Planet. Sci. Lett. 179, p. 139. d Pegram and Turekian (1999) Geochim cosmochim. Act. 63, p. 4053.

Cretaceous–Tertiary Boundary have demonstrated that a wide variety of natural processing can give rise to PGE enrichments unrelated to any extraterrestrial input of these elements. Therefore, it is important to emphasize that the PGEs are only one of several possible lines of evidence used to test impact hypotheses in the geologic record. PGEs in marine sediments are important not only in the context of identifying specific extraterrestrial impact events, but also in quantifying the background flux of cosmic dust to the earth’s surface. This flux is critically important to determining what level of Ir enrichment is likely to constitute evidence of an extraterrestrial impact. Slowly accumulating pelagic clays from the abyssal North Pacific are the best available records of the average long-term flux of extraterrestrial material to the earth’s surface. This is because of their very slow accumulation rates, on the order of a few millimeters per thousand years. These slow accumulation rates reflect the fact that this region of the ocean is far removed from terrestrial sources of particulate material. Thus a few meters of sediment can provide a nearly continuous record of accumulation that spans several million years and maximizes the contribution of the background flux of extraterrestrial Ir relative to total Ir burial flux. The best example of such a record is the

Figure 6 Concentration variations of Co and Ir vs. depth in LL44-GPC3 (A) and the same data plot as Ir vs. Co (B). The large Ir concentrations at 20 m depth (A) and 10 ng g1 (B) correspond to the Cretaceous–Tertiary boundary Ir spike in this core. The slope of the Ir-Co trend represented by the bulk of the data is close to the Ir/Co ratio of average upper crust. This similarity suggests much of the Ir in this core may be derived from terrestrial rather than extraterrestrial sources. The steep line on the lower plot corresponds to the Ir/Co ratio of deep-water. In order for a significant fraction of the total Ir to occur as particulate extraterrestrial material Ir must be significantly more insoluble than Co, consistent with data from Figure 1. Data are from Kyte et al. (1993) Geochimica et Cosmochimica Acta 57: 1719.

red clay sequence from LL44-GPC3, a core recovered from the North Pacific (Figure 6). However in such sediment records the influence of Ir that is scavenged from sea water and is not directly associated with extraterrestrial particles complicates interpretations. In the case of LL44-GPC3, lower than chondritic Os/Ir ratios and 187Os/188Os ratios much higher than those that characterize meteoritic material indicate that more than 50% of the average total Ir flux is derived from sea water. Determining the proportion of the seawater-derived Ir that originated from dissolution of cosmic dust and that which originated from terrestrial sources is very difficult. Analyses of dissolved Ir in rivers that accompany recent analyses of dissolved Ir in sea water suggest that riverine supply of Ir may account for more than half of the seawater-derived Ir that accumulates in deep-sea sediments. Uncertainties associated with these types of interpretations ultimately limit the precision and accuracy of estimates of the


502


long-term background flux of extraterrestrial Ir to the earth.

Anthropogenic Release of PGEs to the Marine Environment Since the mid-1970s commercial demand for the PGEs, particularly Pt and Pd has been increasing rapidly (Figure 7). Although considerable effort is invested in recovering and recycling these metals, due in large part to their high cost, there is increasing evidence that release of these metals to the environment is giving rise to higher environmental concentrations in portions of the environment subject to anthropogenic perturbation. Among the many uses of PGEs, the utilization of Pt, Pd and more recently Rh, in automobile catalytic converters is the one pathway for anthropogenic PGE release to the environment that is best documented and most likely to lead to widespread dispersal of these metals. Although the most immediate impact of this mode of release is on land, anthropogenic PGEs also find their way to the marine environment. Direct release from storm sewers draining roadways, and indirect release from municipal sewage plants that treat road run-off with other wastewater streams are the two most likely modes of transport of autocatalyst PGEs to the marine environment. It is important to stress that the municipal waste streams may also carry PGEs associated with medical, dental, chemical and electronic applications. Although marine chemists have been aware of the potential importance of anthropogenic PGE release

for many years, the same analytical challenges that limit the amount of PGE data from pristine marine environments are also responsible for the paucity of information constraining the distribution and behavior of anthropogenic PGEs in the marine environment. The long-standing interest of marine chemists in anthropogenic PGEs is clearly illustrated by the fact that the report of dissolved Pd profiles in sea water (see above) in the mid 1980s was accompanied by data demonstrating elevated Pd concentrations in contaminated sediments from Japan. Release from autocatalysts was suggested as a possible source. In the intervening years there have been very few studies that have addressed this matter. Recent work in contaminated sediments from Boston Harbor in the north-eastern part of the USA shows that human activity has resulted in greater than fivefold increases in bulk sediment Pt and Pd concentrations, relative to background levels of approximately 1 ng g1 (Figure 8). However, as enrichment of these metals predated the introduction of catalytic converters, there must be important sources of these metals to the marine environment other than autocatalysts. Temporal trends in the data show that although the concentrations of Ag and Pb in Boston Harbor sediments are decreasing, likely due to the cessation of sewage release, concentrations of Pt and Pd are either stable or increasing with time. This trend is consistent with a significant input of these metals from nonpoint sources such as

250

Pd/Al (ng g_1)

200 6000 Pd Pt Rh

′ Demand ( 000 oz)

5000 4000

100 50

3000 2000

150

Autocatalysts introduced in the US

0 0

1000 0 1970

20

40

60

80

100

120

_ Pt/Al (ng g 1) 1975

1980

1985 Year

1990

1995 2000

Figure 7 Representation of global demand for Pt, Pd and Rh by the automobile industry. Note that although no data are shown for Pd prior to 1980, Pd was used in some autocatalyst formulations prior to this time. These increasing demand trends suggest that PGEs release from autocatalysts is likely to become an increasingly important source of anthropogenic PGEs to the environment. Data are from the Johnson Matthey Platinum/2000 publication.

Figure 8 Plot of Pd/Al vs. Pt/Al in bulk sediment samples from Boston Harbor ( ) and Massachusetts Bay (&) illustrating the influence of anthropogenic PGE release in this area. Pd and Pd concentrations are normalized to Al to eliminate grain size and dilution effects. All sediments from Massachusetts Bay are uninfluenced by human activity, based on depositional age estimates and Ag analyses. Contaminated sediments from Boston Harbor are enriched in Pt and Pd relative to pristine sediment. This is true for absolute Pt and Pd concentrations as well the Pt/Al and Pd/Al data shown above. Data are from Tuit et al. (2000) Environmental Science Technology 34: 927.



the release of untreated road run-off. The impact of human activity on levels of dissolved Pt and Pd in coastal waters is not well documented. Similarly, in the marine environment, the chemical form of anthropogenic PGEs, and the extent to which these metals are subject to biological uptake are also poorly known. These gaps in our knowledge of the marine chemistry of the PGEs will likely influence the future direction of marine PGE research.

Summary The PGEs are among the least abundant elements in sea water. The low concentrations of these metals in sea water reflect their generally low concentration in earth surface material rather than uniformly low solubility. Although there is a general consensus regarding the approximate concentrations of these metals in sea water, their vertical distribution in the water column remains controversial and poorly documented. Improving our understanding of the marine chemistry of the PGEs both in the water column and in marine sediments is important to interpreting the marine Os isotope record, exploiting PGEs as tracers of extraterrestrial material in marine sediments, and understanding the consequences of anthropogenic release of PGEs to the marine environment.

503

See also Glacial Crustal Rebound, Sea Levels, and Shorelines. Satellite Altimetry. Sea Level Variations Over Geologic Time.

Further Reading Donat JR and Bruland KW (1995) Trace elements in the oceans. In: Steinnes E and Salbu B (eds.) Trace Elements in Natural Waters, ch. 11. Boca Raton: CRC Press. Goldberg ED and Koide M (1990) Understanding the marine chemistries of the platinum group metals. Marine Chemistry 30: 249--257. Helmers E and Kummerer K (eds) (1997) Platinum group elements in the environment – anthropogenic impact. Environtal Science and Pollution Research 4: 99. Kyte FT (1988) The extraterrestrial component in marine sediments: description and interpretation. Paleoceanography 3: 235--247. Lee DS (1983) Palladium and nickel in north-east Pacific waters. Nature 313: 782--785. Peucker-Ehrenbrink B and Ravizza G (2001) The marine Os isotope record: a review. Terra Nova, in press. Peucker-Ehrenbrink B (1996) Accretion of extraterrestrial matter during the last 80 million years and its effect on the marine osmium isotope record. Geochimica et Cosmochimica Acta 60: 3187--3196.


PLIO-PLEISTOCENE GLACIAL CYCLES AND MILANKOVITCH VARIABILITY K. H. Nisancioglu, Bjerknes Centre for Climate Research, University of Bergen, Bergen, Norway & 2009 Elsevier Ltd. All rights reserved.

Introduction A tremendous amount of data on past climate has been collected from deep-sea sediment cores, ice cores, and terrestrial archives such as lake sediments. However, several of the most fundamental questions posed by this data remain unanswered. In particular, the Plio-Pleistocene glacial cycles which dominated climate during the past B2.8 My have puzzled scientists. More often than not during this period large parts of North America and northern Europe were covered by massive ice sheets up to 3 km thick, which at regular intervals rapidly retreated, giving a sea level rise of as much as 120 m. The prevalent theory is that these major fluctuations in global climate, associated with the glacial cycles, were caused by variations in insolation at critical latitudes and seasons. In particular, ice sheet growth and retreat is thought to be sensitive to high northern-latitude summer insolation as proposed by Milankovitch in his original astronomical theory.

Brief History of the Astronomical Theory Long before the first astronomical theory of the ice ages, the people of northern Europe had been puzzled by the large erratic boulders scattered a long way from the Alpine mountains where they originated. Based on these observations the Swiss geologist and zoologist Louis Agassiz presented his ice age theory at a meeting of the Swiss Society of Natural Sciences in Neuchatel in 1837, where he claimed that the large boulders had been transported by Alpine glaciers covering most of Switzerland in a past ice age. A few years later, the French mathematician Joseph Alphonse Adhemar was the first to suggest that the observed ice ages were controlled by variations in the Earth’s orbit around the Sun. At this point it was known that there had been multiple glaciations, and Adhemar proposed that there had been alternating ice ages between the North and the South Pole following the precession of the equinoxes.

504

Indeed, the winter is warmer when the Earth is at the point on its orbit closest to the Sun, and colder when the Earth is furthest from the Sun. Adhemar correctly deduced that the precession of the equinoxes had a period of approximately 21 000 years, giving alternating cold and warm winters in the two hemispheres every 10 500 years. In 1864, James Croll expanded on the work by Adhemar and described the influence of changing eccentricity on the precession of the equinoxes. He assumed that winter insolation controlled glacial advances and retreats, and determined that the precession of the equinoxes played an important role in regulating the amount of insolation received during winter. Based on this, he estimated that the last ice age lasted from about 240 000 to 80 000 years ago. Croll was aware of the fact that the amplitude of the variations in insolation was relatively small, and introduced the concept of positive feedbacks due to changing surface snow and ice cover as well as changes in atmosphere and ocean circulation. In parallel to the work of Croll, geologists in Europe and America found evidence of multiple glacial phases separated by interglacial periods with milder climate similar to that of the present day, or even warmer. These periodic glaciations were consistent with Croll’s astronomical theory. However, most geologists abandoned his theory after mounting evidence from varved lake sediments in Scandinavia and North America showed that the last glacial period ended as late as 15 000 years ago, and not 80 000 years ago as suggested by Croll. Milankovitch’s Astronomical Theory of the Ice Ages

Following Croll’s astronomical theory there was a period where scientists such as Chamberlin and Arrhenius tried to explain the ice ages by natural variations in the atmospheric content of carbon dioxide. The focus of the scientific community on an astronomical cause of the ice ages was not renewed until the publication of Milankovitch’s theory in a textbook on climate by the well-known geologists Wladimir Ko¨ppen and Alfred Wegener in 1924. This was the first comprehensive astronomical theory of the Pleistocene glacial cycles, including detailed calculations of the orbitally induced changes in insolation. Milutin Milankovitch was of Serbian origin, born in 1879. He obtained his PhD in Vienna in 1904 and


PLIO-PLEISTOCENE GLACIAL CYCLES AND MILANKOVITCH VARIABILITY

was later appointed Professor of Applied Mathematics at the University of Belgrade. He was captured during World War I, but allowed to work at the Hungarian Academy of Sciences, where he completed his calculation of the variations of the orbital parameters of the Earth and their impact on insolation and climate. Milankovitch’s basic idea was that at times of reduced summer insolation, snow and ice could persist at high latitudes through the summer melt season. At the same time, the cool summer seasons were accompanied by mild winter seasons leading to enhanced winter accumulation of snow. When combined, reduced summer melt and a slight increase in winter accumulation, enhanced by a positive snow albedo feedback, could eventually lead to full glacial conditions. During World War II, Milankovitch worked on a complete revision of his astronomical theory which was published as the Kanon der Erdbestrahlung in 1941. However, the scientific establishment was critical of Milankovitch, and his theory was largely rejected until the early 1970s. By this time, great advances in sediment coring, deep-sea drilling, and dating

(a)

505

techniques had made it possible to recover climate records covering the last 500 000 years. By studying the variations in oxygen isotopes of foraminifera in deepsea sediment cores as well as by reconstructing past sea level from terraces of fossil coral reefs, new support was emerging for an astronomical phasing of the glacial cycles. The oxygen isotope data from the long deep-sea cores presented in a paper in 1976 by Hays et al. were considered as proof of the Milankovitch theory, as they showed cycles with lengths of roughly 20 000 and 40 000 years as well as 100 000 years in agreement with Milankovitch’s original calculations.

Orbital Parameters and Insolation The Earth’s orbit around the Sun is an ellipse where the degree to which the orbit departs from a circle is measured by its eccentricity (e). The point on the orbit closest to the Sun is called the perihelion, and the point most distant from the Sun the aphelion (Figure 1). If the distance from the Earth to the Sun is rp at perihelion, and ra at aphelion, then the eccentricity is defined as e ¼ ðra rp Þ=ðra þ rp Þ.

Spring equinox

Today Spring

Win

ter

Summer solstice

Winter solstice Perihelion

Aphelion

Sum

mer

Fall

Fall equinox

(b)

11 000 years ago

Fall equinox

Winter solstice Sun

Perihelion

Summer solstice

Aphelion Earth

Spring equinox Figure 1 Sketch of the Earth’s orbit around the Sun today and at the end of the last glacial cycle (11 000 years ago), showing the positions of the solstices and equinoxes relative to perihelion. The longitude of perihelion (o) is measured as the angle between the line to the Earth from the Sun at spring equinox and the line to the Earth at perihelion.


506


Variations in the eccentricity of the Earth’s orbit follow cycles of 100 000 and 400 000 years giving a change in annual mean insolation on the order of 0.2% or less. This change in insolation is believed to be too small to produce any notable effect on climate. A more significant change in insolation is caused by variations in the seasonal and latitudinal distribution of insolation due to obliquity. Obliquity (e) is the angle between Earth’s axis of rotation and the normal to the Earth’s plane around the Sun (Figure 1). This angle is 23.51 today, but varies between values of 22.11 and 24.51 with a period of 41 000 years. A decrease in obliquity decreases the seasonal insolation contrast, with the largest impact at high latitudes. At the same time, annual mean insolation at high latitudes is decreased compared to low latitudes. An example of the effect of obliquity variations on seasonal insolation is shown in Figure 2(a). During times when obliquity is small, high-latitude summertime insolation decreases, whereas midlatitude wintertime insolation increases. The magnitude of the change in high-latitude summer insolation due to obliquity variations can be as large as 10%. The third and last variable affecting insolation is the longitude of perihelion (o). This parameter is

defined as the angle between the line to the Earth from the Sun at spring equinox and the line to the Earth at perihelion (Figure 1). It determines the direction of the Earth’s rotational axis relative to the orientation of the Earth’s orbit around the Sun, thereby giving the position of the seasons on the orbit relative to perihelion. Changes in the longitude of perihelion result in the Earth being closest to the Sun at different times of the year. Today, the Earth is closest to the Sun in early January, or very near winter solstice in the Northern Hemisphere. All other things being equal, this will result in relatively warm winter and cool summer seasons in the Northern Hemisphere, whereas the opposite is the case in the Southern Hemisphere. At the time of the last deglaciation, 11 000 years ago the Earth was closest to the Sun at summer solstice, resulting in extra warm summers and cool winters in the Northern Hemisphere. An example of the effect of changes in precession on seasonal insolation is shown in Figure 2(b). If the Earth’s orbit were a circle, the distance to the Sun would remain constant at all times of the year and it would not make any difference where on the orbit the seasons were positioned. Therefore, the impact of variations in the longitude of perihelion depends on the eccentricity of the Earth’s orbit and is described by

(a) 90 −40

Latitude (deg)

60 30

0

0

0

0

0

−30 −60

−40

−90 SE

SS

FE Time of year

SE −60 0 60 (W m−2)

WS

(b)

0

40

80

40

0

30 0 −30

0

−60 −90 SE

SS

40 −80

FE Time of year

0

0

40

−4

40

Latitude (deg)

0 −4

0

60

−40

90

WS

SE −60 0 60 W m−2

Figure 2 Insolation difference in units of W m 2 as a function of latitude and season: (a) when decreasing obliquity from 24.51 to 221 in the case of a perfectly circular orbit (e ¼ 0); and (b) for a change in precession going from summer solstice at perihelion to summer solstice at aphelion while keeping obliquity at today’s value (e ¼ 23.51) and using a mean value for eccentricity (e ¼ 0.03). The annual mean insolation difference is shown to the right of each figure and the seasons are defined as follows: FE, fall equinox; SE, spring equinox; SS, summer solstice; WS, winter solstice.



the precession parameter (e sin o). The combined effect of eccentricity and longitude of perihelion can give changes in high-latitude summer insolation on the order of 15% and varies with periods of 19 000 and 23 000 years, but is modulated by the longerperiod variations in eccentricity. Figure 3 shows the variations in obliquity (e), eccentricity (e), and the precession parameter (e sin o).

507

(a) 25 24 23 22

0

200

400

600

800

1000

Time before present (millenia)

Plio-Pleistocene Glacial Cycles Some of the longest continuous records of past climate come from deep-sea sediment cores. Ocean sediments are laid down over time, and by drilling into the seafloor, layered sediment cores can be extracted containing valuable information about the conditions at the time when the layers were formed. By studying the relative abundance of oxygen isotopes in shells of tiny marine organisms (foraminifera) found in the sediments, it is possible to estimate the amount of water tied up in the continental ice sheets and glaciers. This is because water molecules containing the lighter isotope of oxygen (16O) are more readily evaporated and transported from the oceans to be deposited as ice on land. Thus, leaving the ocean water enriched with the heavy oxygen isotope (18O) during glacial periods. However, the fractionation of the oxygen isotopes when forming the shells of the foraminifera also depends on the surrounding water temperature: low water temperature gives higher d18O values (the ratio of 18O and 16O relative to a standard). Therefore records of d18O are a combination of ice volume and temperature. By analyzing benthic foraminifera living on the seafloor where the ocean is very cold, and could not have been much colder during glacial times, the contribution of temperature variations to the d18O value is reduced. The benthic d18O ice volume record of Hays et al. from 1976 was one of the very first continuous records of the late Pleistocene extending back to the Brunhes–Matuyama magnetic reversal event (780 000 years ago), making it possible to construct a timescale by assuming linear accumulation rates. Analysis of the data showed cycles in ice volume with periods of about 20 000 years and 40 000 years, with a particularly strong cycle with a period of roughly 100 000 years. Later studies extended the record past the Brunhes–Matuyama reversal, showing that the late Pliocene (3.6–1.8 Ma) and early Pleistocene records (1.8–0.8 Ma) were dominated by smalleramplitude cycles with a period of 41 000 years, rather than the large 100 000 years cycles of the late Pleistocene (0.8–0 Ma). Many records generated since this time have confirmed these early observations, namely:

(b) 0.06 0.04 0.02 0

0

200

400

600

800

1000


Summer

Winter

(c) 0.05

0

-0.05 0

200

400

600

800

1000


Summer

Winter

Figure 3 The three most important cycles regulating insolation on Earth are obliquity, eccentricity, and precession: (a) obliquity, or tilt of the Earth’s axis varies with a period of 41 000 years; (b) eccentricity of the Earth’s orbit varies with periods of 100 000 years and 400 000 years; and (c) precession of the equinoxes has a dominant period of 21 000 years and is modulated by eccentricity.

1. from about 3 to 0.8 Ma, the main period of ice volume change was 41 000 years, which is the dominant period of orbital obliquity; 2. after about 0.8 Ma, ice sheets varied with a period of roughly 100 000 years and the amplitude of


508


2.0 41 ka

~100 ka

2.5

18 O (per mil)

3.0 3.5 4.0 4.5 5.0 5.5 −3000

−2500

−2000

−1500

−1000

−500

0

Time before present (millenia) Figure 4 Benthic foraminiferal d18O ice volume record from the North Atlantic plotted to a paleomagnetic timescale covering the last 3 My. The transition from a dominant 41 000 to a 100 000-year periodicity in ice volume occurs close to the Brunhes–Matuyama magnetic reversal event (B780 000 years ago).

oscillations in d18O increases, implying growth of larger ice sheets. The long benthic d18O record from a deep-sea sediment core extracted from the North Atlantic shown in Figure 4 illustrates both of these points. This isotope record is plotted with a paleomagnetic timescale determined by the depth of magnetic field reversals recorded by ferromagnetic grains in the sediment core. Using this simple timescale, which is not biased by orbital tuning, one can clearly observe the 41 000-year periodicity of the late Pliocene and early Pleistocene (3.0–0.8 My), as well as the dominance of the stronger B100 000-year periodicity of the late Pleistocene (last 800 000 years). Note that the main periods of orbital precession (19 000 and 23 000 years) are of less importance in the benthic ice volume record, whereas it is known that they increase in strength after about 800 000 years (the mid-Pleistocene transition). The lack of an imprint from orbital precession in the early part of the record and the reason for the dominance of roughly 100 000 years periodicity in the recent part of the record are some of the major unanswered questions in the field. Only eccentricity varies with periods matching the roughly 100 000 years periods observed in the late Pleistocene. Although eccentricity is the only orbital parameter which changes the annual mean global insolation received on Earth, it has a very small impact. This was known to Croll and Milankovitch, who saw little direct importance in variations in eccentricity and assumed that changes in precession and obliquity would dominate climate by varying the amount of seasonal, rather than annual mean insolation received at high latitudes. Milankovitch postulated that the total amount of energy received from

the Sun during the summer at high northern latitudes is most important for controlling the growth and melt of ice. To calculate this insolation energy, he divided the year into two time periods of equal duration, where each day of the summer season received more insolation than any day of the winter season. The seasons following these requirements were defined as the caloric summer and caloric winter half-years. These caloric half-years are of equal duration through time and the amount of insolation energy received in each can be compared from year to year. For Milankovitch’s caloric summer half-year insolation (Figure 5),obliquity (e) dominates at high latitudes (4651 N), whereas climatic precession (e sin o) dominates at low latitudes (o551 N). In the mid-latitudes (B55 651 N), the contribution by obliquity and climatic precession are of similar magnitude. In the Southern Hemisphere, variations in caloric halfyear insolation due to obliquity are in phase with the Northern Hemisphere and could potentially amplify the global signal, whereas variations due to climatic precession are out of phase. By taking into account the positive snow albedo feedback, Milankovitch used his caloric insolation curves to reconstruct the maximum extent of the glacial ice sheets back in time (Figure 6). Milankovitch’s predicted cold periods occurred roughly every 40 000–80 000 years, which fit reasonably well with the glacial advances known to geologists at that time. However, as the marine sediment core data improved, it became clear that the last several glacial periods were longer and had a preferred period of roughly 100 000 years (Figure 4), which was not consistent with Milankovitch’s original predictions. Based on these observations, and without knowledge of the 41 000 years cycles of the



509

6 80° N 60° N

4 (J m−2) × 106

40° N 2

0

2

4 −600

−500

−400

−300

−200

−100

0

Time before present (millenia) Figure 5 Caloric summer half-year insolation following Milankovitch’s definition plotted for the latitudes of 401 N, 601 N, and 801 N. Caloric half-years are periods of equal duration where each day of the summer half-year receives more insolation than any day of the winter half-year.

early Pliocene and late Pleistocene, scientists reasoned that climatic precession and its modulation by eccentricity must play the leading role in past climates. Following Andre Berger and others, who recalculated and improved the records of orbital insolation, most researchers replaced the caloric halfyear insolation as a driver of glacial climate by midmonth, or monthly mean insolation, for example, June or July at 651 N (Figure 7). As can be seen from Figure 7, monthly mean insolation is dominated by precession. As insolation time series at a given time of the year (e.g., June or July) are in phase across all latitudes of the same hemisphere, the proxy records could be compared equally well with insolation from other latitudes than the typical choice of 651 N shown here. This means that any direct response of climate at high latitudes to monthly or daily insolation requires a strong presence of precession in the geologic record. Although both the frequencies of precession and obliquity are clearly found in the proxy records, a simple linear relationship between summer insolation and glacial cycles is not possible. This is particularly true for the main terminations spaced at roughly 100 000 years, which must involve strongly nonlinear mechanisms. The strong positive feedback on global climate caused by greenhouse gases, such as CO2, was pointed out as early as 1896 by the Swedish physical chemist Svante Arrhenius. Shortly thereafter, the American geologist Thomas Chamberlin suggested a possible link between changing levels of CO2 and glacial cycles. From the long ice cores extracted from Antarctica, covering the last 740 000 years, it is now known that atmospheric levels of CO2 closely follow the glacial temperature record (Figure 8). Although

greenhouse gases, such as CO2, cannot explain the timing and rapidity of glacial terminations, the changing levels of atmospheric greenhouse gases clearly contributed by amplifying the temperature changes observed during the glacial cycles.

Modeling the Glacial Cycles Following the discovery of the orbital periods in the proxy records, a considerable effort has gone into modeling and understanding the physical mechanisms involved in the climate system’s response to variations in insolation and changes in the orbital parameters. In this work, which requires modeling climate on orbital timescales (410 000 years), the typical general circulation models (GCMs) used for studying modern climate and the impact of future changes in greenhouse gases require too much computing power. These GCMs can be used for simulations covering a few thousand years at most, but provide valuable equilibrium simulations of the past climates, such as the Last Glacial Maximum (LGM). Instead of the GCMs, it has been common to use Energy Balance Models (EBMs) to study changes in climate on orbital timescales. These types of models can be grouped into four categories: (1) annual mean atmospheric models; (2) seasonal atmospheric models with a mixed layer ocean; (3) Northern Hemisphere ice sheet models; and (4) coupled climate–ice sheet models, which in some cases include a representation of the deep ocean. Studies with the first type of simple climate models were pioneered by the early work of Budyko in the 1960s, who investigated the sensitivity of climate to changes in global annual mean insolation. However, changes in the Earth’s orbital parameters result in a


510


Latitude equivalent

60°

65°

Millenia Würm I

70°

Riss II

75°

Riss I

300 290 280 270 260 250 240 230 220 210 200 190 180 170 160 150 140 130 120 110 100

60°

Millenia Mindel II

65°

Mindel I

70°

Günz II

90 80 Würm II

70 60 50 40 30

Würm III

20 10 0

600 590 580 570 560 550 540 530 520 510 500 490 480 470 460 450 440 430 420 410 400 390 380 370 360 350 340 330 320 310 300

Günz I

Latitude equivalent

Figure 6 Milankovitch’s reconstructed maximum glacial ice extent for the past 600 000 years. From Milankovitch M (1998) Canon of Insolation and the Ice-Age Problem (orig. publ. 1941). Belgrade: Zavod za Udzbenike I Nastavna Sredstva.



511

(a)

W m−2

550 500 450 400 (b)

W m−2

550 500 450 400 −500

−450

−400

−350

−300 −250 −200 Time before present (millenia)

−150

−100

−50

0

Figure 7 Summer solstice insolation at (a) 651 N and (b) 251 N for the past 500 000 years.

300 280 CO2 (ppmv)

260 240 220 200 180 −360

160

−400 −420

D ice (%% )

−380

−440 800

−460

CH4 (ppbv)

700 600 500 400 300 600 000

500 000

400 000

300 000

200 000

100 000

0

Time (years before 1950)

Figure 8 Variations in deuterium (dD; black), a proxy for local temperature, and atmospheric concentrations of the greenhouse gases carbon dioxide (CO2; red), and methane (CH4; green), from measurements of air trapped within Antarctic ice cores. Data from Spahni R, Chappellaz J, Stocker TF, et al. (2005) Atmospheric methane and nitrous oxide of the late Pleistocene from Antarctica ice cores. Science 310: 1317–1321 and Siegenthaler U, Stocker TF, Monnin E, et al. (2005) Stable carbon cycle-climate relationship during the late Pleistocene. Science 310: 1313–1317.


512


redistribution of insolation with latitude and time of the year, with a negligible impact on global annual mean insolation. Therefore, annual mean models are not adequate when investigating the impact of orbital insolation on climate, as they cannot capture the parts of the insolation variations which are seasonal and translate them into long-term climate change. The second type of models includes a representation of the seasonal cycle, and has been used to investigate the orbital theory of Milankovitch. In this case, the seasonal variations in orbital insolation are resolved. However, as for the first type of models, past changes in ice cover are assumed to follow the simulated variations in the extent of perennial snow. This approach assumes that ice cover and the powerful ice albedo feedback are governed only by temperature, as the extent of snow in these models is fixed to the latitude with a temperature of 0 1C. In reality, the growth and decay of land-based ice sheets are governed by the balance of accumulation and ablation. Therefore, when investigating changes in ice cover, it is necessary to include an appropriate representation of the dynamics and mass balance of ice sheets in the model. The third type of models improves upon this by focusing on modeling past changes in mass balance and size of typical Northern Hemisphere ice sheets, such as the Laurentide. This type of studies was initiated by Weertman in the 1960s who used simple ice sheet models, forced by a prescribed distribution of accumulation minus ablation, to predict ice thickness versus latitude. These models do not calculate the atmospheric energy balance in order to estimate snowfall and surface melt; instead, changes to the prescribed distribution of net accumulation follow variations in mean summer insolation. The fourth type of models include zonal mean seasonal climate models coupled to the simple Weertman-type ice sheet model, as well as earth models of intermediate complexity (EMICs) coupled to a dynamic ice sheet. These models give a more realistic representation of the climate as compared with the simpler models. Partly due to the lack of good data on variations in global ice volume older than about half a million years, most model studies have focused on understanding the more recent records dominated by the B100 000 years glacial cycles. All of these models respond with periods close to the precession and obliquity periods of the insolation forcing. However, the amplitude of the response is in most cases significantly smaller than what is observed in the proxy records. At the same time, the dominant B100 000 year cycles of the ice volume record, characterized by rapid deglaciations, are only found when including a time lag in

the response of the model. Such an internal time lag can be produced by taking into account bedrock depression under the load of the ice, or by adding a parametrization of ice calving into proglacial lakes, or marine incursions at the margin of the ice sheet. Alternatively, the B100 000-year cycles have been explained as free, self-sustained oscillations, which might be phase-locked to oscillations in orbital insolation. One of the very few model studies that have investigated variations in ice volume before the late Pleistocene transition (B800 000 years ago) used a two-dimensional climate model developed at Louvain-la-Neuve in Belgium. It falls within the definition of an EMIC and includes a simple atmosphere coupled to a mixed layer ocean, sea ice and ice sheets. By forcing this model with insolation and steadily decreasing atmospheric CO2 concentrations, the model reproduces some of the characteristics of the ice volume record. The 41 000-year periodicity is present in the simulated ice volume for most of the past 3 My and the strength of the 100 000-year signal increases after about 1 My. However, a longer 400 000-year year period is also present and often dominates the simulated Northern Hemisphere ice volume record. This nicely illustrates the remaining questions in the field. It is expected that models responding to the 100 000-year period will also respond to the longer 400 000-year period of eccentricity. However, this later period is not present in the ice volume record. At the same time, the late Pleistocene transition from a dominance of 41 to B100 000-year period oscillations in ice volume is not well understood. Explanations for the transition which have been tested in models are: a steady decrease in CO2 forcing and its associated slow global cooling; or a shift from a soft to a hard sediment bed underlying the North American ice sheet through glacial erosion and exposure of unweathered bedrock. Neither of these changes are in themselves abrupt, but could cause a transition in the response of the ice sheets to insolation as the ice sheets grew to a sufficiently large size. In addition to the challenge of modeling the midPleistocene transition, no model has successfully reproduced the relatively clean 41 000-year cycles preceding the transition. Following the transition, the models only exhibit a good match with the observed glacial cycles when forced with reconstructed CO2 from Antarctic ice cores together with orbital insolation.

Summary The Plio-Pleistocene glacial cycles represent some of the largest and most significant changes in past



climate, with a clear imprint in terrestrial and marine proxy records. Many of the physical mechanisms driving these large cycles in ice volume are not well understood. However, the pursuit to explain these climate changes has greatly advanced our understanding of the climate system and its future response to man-made forcing. New and better resolved proxy records will improve our spatial and temporal picture of the glacial cycles. Together with the advent of comprehensive climate models able to simulate longer periods of the glacial record, scientists will be able to better resolve the interaction of the atmosphere, ocean, biosphere, and ice sheets and the mechanisms linking them to the astronomical forcing.

Nomenclature e ra rp d18O e o

eccentricity aphelion perihelion oxygen isotope ratio (ratio of 16 O relative to a standard) obliquity longitude of perihelion

18

O and

See also Deep-Sea Drilling Methodology. Deep-Sea Drilling Results. Monsoons, History of. Oxygen Isotopes in the Ocean. Satellite Remote Sensing of Sea Surface Temperatures. Stable Carbon Isotope Variations in the Ocean.

513

Further Reading Bard E (2004) Greenhouse effect and ice ages: Historical perspective. Comptes Rendus Geoscience 336: 603--638. Berger A, Li XS, and Loutre MF (1999) Modelling Northern Hemisphere ice volume over the last 3 Ma. Quaternary Science Reviews 18: 1--11. Budyko MI (1969) The effect of solar radiation variations on the climate of the Earth. Tellus 5: 611--619. Crowley TJ and North GR (1991) Paleoclimatology. New York: Oxford University Press. Hays JD, Imbrie J, and Shackleton NJ (1976) Variations in the Earth’s orbit: Pacemakers of the ice ages. Science 194: 1121--1132. Imbrie J and Imbrie KP (1979) Ice Ages, Solving the Mystery. Cambridge, MA: Harvard University Press. Ko¨ppen W and Wegener A (1924) Die Klimate Der Geologischen Vorzeit. Berlin: Gebru¨der Borntraeger. Milankovitch M (1998) Canon of Insolation and the IceAge Problem (orig. publ. 1941). Belgrade: Zavod za Udzbenike I Nastavna Sredstva. Paillard D (2001) Glacial cycles: Toward a new paradigm. Reviews of Geophysics 39: 325--346. Saltzman B (2002) Dynamical Paleoclimatology. San Diego, CA: Academic Press. Siegenthaler U, Stocker TF, Monnin E, et al. (2005) Stable carbon cycle–climate relationship during the late Pleistocene. Science 310: 1313--1317. Spahni R, Chappellaz J, Stocker TF, et al. (2005) Atmospheric methane and nitrous oxide of the late Pleistocene from Antarctica ice cores. Science 310: 1317--1321. Weertman J (1976) Milankovitch solar radiation variations and ice age ice sheet sizes. Nature 261: 17--20.


POLAR ECOSYSTEMS

Introduction The Arctic Ocean and the Southern Ocean together comprise a little under one-fifth of the world’s oceans (the precise fraction depending on how these oceans are defined). The two polar oceans are similar in being cold, seasonal, productive, and heavily influenced by ice, and both have long been fished by man. They differ markedly, however, in geography, age, and many aspects of their biology. In both polar oceans sea water temperatures are typically low, and in many areas are close to freezing ( 1.861C) for long periods. In areas of seasonal ice cover the surface waters undergo a summer warming, but even at the lowest latitudes this rarely amounts to more than a few degrees. Typically summer sea water temperatures in the seasonal sea ice zone of Antarctica are between þ 0.5 and þ 1.51C. The inclination of the earth’s rotational axis means that the seasonality of received radiation increases towards the poles. This seasonality, exacerbated throughout much of the polar oceans by the accompanying seasonal ice cover, means that primary production (the conversion of inorganic carbon and other essential nutrients into organic matter by plants) is also intensely seasonal. The combination of a low and fairly seasonal temperature with a highly seasonal primary production (Figure 1) distinguishes polar marine ecosystems from all others on earth. Temperate systems are typically also highly seasonal, but here both temperature and primary production tend to co-vary. Tropical marine ecosystems experience high temperatures that are fairly aseasonal, but there are often strong seasonal variations in other important environmental variables (for example, rainfall and fresh water run-off in monsoon areas). These features of the polar regions set two particular challenges for organisms living there. The first is that the low temperatures will tend to slow the rates of many biological processes. This does not apply to mammals and birds, whose internal body temperatures are maintained by metabolic processes; for these endothermic (‘warm-blooded’) organisms the challenge is to avoid losing heat, which they do

514

General Features of Polar Marine Food Webs Food webs can be constructed to emphasize different aspects of their structure or dynamics (for example, to highlight trophic interactions, energy flux, or biomass). When comprehensive, such food webs can be extremely complex. There is, however, usually

18 16 _


Chlorophyll (mg m 3)

Copyright & 2001 Elsevier Ltd.

by a variety of mechanisms but principally by enhanced insulation. The subtle molecular mechanisms used by ectothermic (‘cold-blooded’) fish, invertebrates, plants, and bacteria are the subject of extensive physiological investigation. Despite these adaptations, however, it is a widespread observation that many important biological processes proceed slowly in polar systems. The second challenge for polar organisms is that the marked seasonality of primary production means that many of these processes are constrained to the summer months. These two features of generally slow physiological rates and intense seasonality together have a profound influence on the structure and dynamics of polar marine ecosystems.

14 12 10 8 6 4 2

Temperature (˚C)

A. Clarke, British Antarctic Survey, Cambridge, UK

0 5 0 _5 J

J

A

S

O

N D Month

J

F

M A

M

Figure 1 Seasonal variation in chlorophyll biomass (40.2 mm) and temperature at a typical polar locality (Signy Island, maritime Antarctic, 601S). Both plots show weekly means over the period 1988–1994. The horizontal lines show zero chlorophyll biomass and 1.861C (the equilibrium freezing point of sea water). The marked seasonality of chlorophyll biomass coupled with the small annual variation in water temperature is typical of polar oceans.


POLAR ECOSYSTEMS

Sea Ice

Higher predators Whales, seals seabirds Nekton Fish, squid, gelatinous Microbial loop Zooplankton Krill, copepods

515

Bacteria DOC Protozoa Primary production Mixed layer Midwater

Benthos

Figure 2 The basic structure of all marine food webs. Carbon fixed in primary production flows in three directions: to zooplankton herbivores (and thence to zooplanktonic and nektonic carnivores, and vertebrate higher predators), to the microbial loop, and via sedimentation to the benthos. In oceanic areas, the vertical flux of phytodetritus (from primary production) and particulate matter from zooplankton (primarily fecal pellets) passes out of the mixed layer and through the midwater to the benthos. In coastal areas, the benthos is within the mixed layer. Although specific food webs for individual oceanic locations will vary in complexity and the relative importance of different pathways, all have the same underlying structure. DOC, dissolved organic carbon.

a basic underlying structure, and a simplified food web for all marine systems can be produced (Figure 2). The key feature of this basic structure is that it is nonlinear; energy fixed by marine plants (principally phytoplankton but also macroalgae in shallow water areas) passes in three important directions. These are (1) to planktonic herbivores and through a sequence of higher trophic levels eventually to top predators (including man); (2) to the microbial loop; and (3) through vertical flux (sedimentation) to the benthos. In deeper water the flux passes through the vast range of the midwater before reaching the abyssal plain. This basic structure applies to all marine ecosystems, but there are a number of extra features that are peculiar to polar regions. Particularly important are the influence of sea ice, the dominance at times of short linear food chains, and the importance of top predators.

Sea ice has profound and pervasive effect on polar marine ecosystems. It forms a barrier to the exchange of energy, momentum, and gases, thereby acting as a major regulator of primary production in the water column beneath. In particular where snow lies on the ice, surface incident radiation can be reduced to a sufficiently low level to prevent photosynthesis. The isolation of the water column from wind also reduces turbulence and many biological particles (both living and inanimate) sediment out of the photic zone. Sea ice also has a positive effect on primary production in polar oceans. In areas where sea ice is melting (the marginal ice zone) the release of fresh water can lead to an area of increased water column stability that encourages rapid primary production. It is not at present clear to what extent the enhanced primary productivity is caused by the increased stability, by seeding by phytoplankton cells released from the ice, or by other oceanographic features; indeed, it is likely that each of these factors may be important in different places at different times. A second important influence by sea ice is its role as a habitat, supporting a diverse microbial community in the interstices between ice crystals. This distinctive assemblage (sometimes called the sympagic or epontic community) can grow throughout annual sea ice but is almost always most highly developed at the interface between the ice and the underlying sea water. The growth of phytoplankton and other microbial primary producers can be so intense as to color the undersides of the ice deep green or brown, colors that are easily seen as floes are overturned by ships passing through ice in late winter or spring. The contribution of ice-associated primary production to production in the Arctic or Southern Oceans as a whole is not known with any certainty. Current estimates for the Southern Ocean are of the order of 25–30%. In studies of sea ice communities much attention has been directed to diatoms, but the microbial community growing within ice can be rich and diverse. It includes bacteria and flagellates as primary producers and a variety of flagellates, ciliates, and other protozoans as consumers. Associated with these are a range of meiofauna, and the whole assemblage is grazed by consumers. Particularly noteworthy among these are amphipods (in both Arctic and Antarctic systems) and euphausiids (principally in the Antarctic, where sea ice plays an important role in the biology of the Antarctic krill, Euphausia superba). These herbivores are themselves subject to predation by specialist consumers associated with the undersides of ice, most notably gelatinous zooplankton


516

POLAR ECOSYSTEMS

such as ctenophores and in the Antarctic the cryopelagic fish Pagothenia borchgrevinki. Primary Production and the Microbial Loop

The microbial loop (Figure 2) is known to be a feature of all ocean systems, though its importance varies from place to place. It is undoubtedly present in polar regions, and can be important at times. Summer chlorophyll biomass is, however, dominated by diatoms rather than the small cells that contribute directly to the microbial loop. Nevertheless, all the components of the microbial loop are well documented from polar oceans, and there is considerable release of dissolved organic matter from rapidly growing phytoplankton cells in summer. It is likely that the microbial loop is present and active in all polar seas, but that production is dominated by larger cells for much of the summer open water season. One group of primary producers that are important in many oceanic areas, but that are absent from truly polar seas, are coccolithophorids. This may be because the low temperatures are unfavorable for calcification. Short Linear Food Chains

The polar regions have long been famous for the presence of short, linear food chains. The classic example of this is the trophic relationship that characterizes the lower latitude regions of the Southern Oceans: diatoms–krill–whales. This is a three trophic level food chain through which energy

progresses in just two steps from microscopic cells to the largest organisms the world has ever seen. This particular food chain has long formed a paradigm for polar oceans in general (Figure 3). While there can be no doubt that in some places at some times energy flux through polar marine ecosystems can be dominated by such short, linear food chains, the system as a whole is more complex and nonlinear (Figure 2). Vertical Flux

Vertical flux of biological material is usually measured by deploying sediment traps below the euphotic zone. Where this has been done on a year-round basis, vertical flux has been shown to be markedly seasonal. Almost all of the annual flux occurs in summer, and winter sedimentation rates can be almost nonexistent. Maximal rates appear to be associated with the ice edge, and swarms of zooplankton can produce a significant flux of fecal pellets. Midwater

The midwater zone is one of the least explored in polar regions. It is known that many species characteristic of lower latitudes extend into polar regions, and this is probably because the oceanographic features that define the polar regions are typically surface features and their influence may not extend to the mesopelagic realm. As with the midwater elsewhere, the food web in polar mesopelagic waters is characterized by Arctic

Antarctic Higher predators

Higher predators

Fish, squid

Capelin

Krill

ML

Phytoplankton

Benthos

Copepods

ML

Phytoplankton

Benthos

Figure 3 Representative polar food webs for the Antarctic (South Georgia) and the Arctic (Barents Sea), showing how energy flow can be dominated in some places at some times by a short, effectively linear, food chain. The major routes of energy flow are highlighted by the shaded boxes and the thicker arrows. The South Georgia example has become something of a paradigm for polar oceans in general. ML, microbial loop.


POLAR ECOSYSTEMS

crustaceans, fish, and gelatinous zooplankton. The latter are little known from polar regions because they are badly damaged by traditional net sampling techniques, and to date there have been no submersible expeditions to polar waters to complement those from temperate midwaters where gelatinous zooplankton have been shown to be very important. Myctophid fish are a major component of the midwater fauna and they form an important link to the surface food web through their vertical migration. They rise into surface waters by night, and here they are both an important consumer of crustacean zooplankton and an important prey item for sea birds and other top predators. Benthos

The benthos is an important component of the marine food web in all oceans. In shallow waters benthic organisms can feed directly within the zone of surface production, whereas in deeper waters they rely primarily on vertical flux from the euphotic zone. In polar regions suspension feeders are particularly important, with sponges, ascidians (sea squirts), bryozoans, brachiopods, holothurians (sea cucumbers) and fan-worms all well represented. Top Predators

Polar marine ecosystems have long been recognized as being rich in top predators (seabirds, seals, and whales). In the Arctic the important top predators are seabirds (especially alcids and gulls) and the smaller toothed wales, although in the past baleen whales were also important. In the Southern Ocean the important top predators are penguins, albatrosses, and petrels among the sea birds, and both toothed and baleen whales. These top predators achieve greater biomass in polar regions than anywhere else in the world’s oceans, and their role within the marine food web is very significant. Many feed directly on crustacean zooplankton, thereby creating the short linear food chains that can characterize the system at some times. Others feed on fish and squid, and in some polar regions the status of these top predators is being used as a simple measure expressing the overall integrated well-being of the polar marine ecosystem. Seasonality

Two features contribute to the intense seasonality of the polar marine ecosystem. The first is that primary production, and hence food availability for herbivores, is constrained almost completely to the summer months. In consequence, the growth and

517

reproduction of organisms at all higher levels in the food web (both plankton and benthic) is itself often constrained to summer. Many species, however, have life cycles longer than one year and must overcome the long winter period of food shortage. This they do either through the utilization of reserves, typically fat, laid down the previous summer (many zooplankton), or by utilizing general body tissue and ‘de-growing’ over winter (some zooplankton and many benthos). This seasonality leads to very large differences in energy flux through the polar food web between summer and winter. There is also an important seasonal variation in food web structure in that many top predators migrate to low latitudes in winter, thereby decreasing predation pressure on lower levels of the food web. Man and the Polar Food Web

Although the polar regions lie far from many home ports and can be extremely inhospitable, they have been subject to intense disturbance by man. The Arctic has long been a traditional fishing ground for both demersal and midwater fish, and at both poles a variety of top predators have been subject to intense exploitation. In neither the Arctic nor the Southern Ocean can the structure or dynamics of the marine ecosystem be regarded as pristine (in the sense of reflecting the position before the intervention of man). In the Arctic three species of large whale have been either effectively eradicated or reduced to very low populations; several species of fish are now at levels that render them economically unviable; and a major benthic crustacean predator (Alaskan king crab) has been reduced to very low population levels. In the Antarctic the large baleen whales have been reduced to very low populations (though the smaller Minke whale may have increased as a result) and the Southern fur seal is recovering from the brink of extinction. Many demersal fish stocks have been reduced to very low levels, and intermediate levels of the food web (Antarctic krill) are now being fished. We can only speculate as to the structure of these polar food webs prior to the intervention of man. The complexity of the system and the nonlinearity of many of the dynamical processes involved make both prediction and hindcasting almost impossible. It is distinctly possible, however, that following the disturbance of these systems, they have settled (or will eventually do so) to a new and different stable structure. We may never again have a Southern Ocean dominated by the great whales, no matter how long we wait. Such complexities also make it


518

POLAR ECOSYSTEMS

very difficult to predict the effect of further disturbances, such as the development of a new fishery, or climate change. Although there are many general features common to both polar marine food webs, there are also features peculiar to each.

The Arctic Food Web The Arctic is a deep-water basin surrounded by wide shallow continental shelves and almost completely enclosed by large land masses. Extensive river systems drain into the Arctic basin, delivering large volumes of fresh water and sediment. Exchange of water with lower latitudes is highly constrained, and most of the surface is covered with multiyear ice. The Arctic basin food web is probably relatively young, as it is likely that very little planktonic or benthic fauna was present at the height of the last glaciation. Relatively little is known of the food web of the high Arctic basin as it is so difficult to work there. The benthos is low in diversity, though it is not clear to what extent this is a function of the relative youth of the system, the heavy input of fresh water and sediment in some areas, or the intense disturbance from bottom-feeding marine mammals. At lower latitudes there are important stocks of shoaling plankton-feeding fish, which have long been exploited by man.

The Southern Ocean Food Web The Antarctic is in many ways a mirror image of the Arctic. It is a large continental land mass entirely surrounded by a deep ocean. The marine food web is likely to have been in existence for many millions of years, and while the zooplankton community

appears to be relatively low in species richness, the benthos exhibits a diversity fully comparable with all but the richest habitats elsewhere. The summer phytoplankton bloom is formed predominantly of the larger diatoms, and the haptophyte Phaeocystis appears not to be as important here as in the Arctic. The zooplankton is dominated by copepods and euphausiids, with Euphausia superba at lower latitudes and E. crystallarophias closer to the ice. Midwater planktivorous fish are almost absent, and the fish fauna is dominated by the radiation of two predominantly benthic/demersal groups: notothenioids on the continental shelf and lipariids in the deeper water of the continental slope. As with the Arctic, relatively little is known of the deep sea.

See also Arctic Ocean Circulation. Baleen Whales. Current Systems in the Southern Ocean. Fisheries and Climate. Marine Mammal Overview. Marine Mammals: Sperm Whales and Beaked Whales. Microbial Loops. Seabird Foraging Ecology. Southern Ocean Fisheries.

Further Reading Knox GA (1994) The Biology of the Southern Ocean. Cambridge: Cambridge University Press. Fogg GE (1998) The Biology of Polar Habitats. Oxford: Oxford University Press. El-Sayed SZ (ed.) (1994) Southern Ocean Ecology. Cambridge: Cambridge University Press. Smith WO (ed.) (1990) Polar Oceanography, vols 1 and 2. San Diego: Academic Press.


POLLUTION, SOLIDS C. M. G. Vivian and L. A. Murray, The Centre for Environment, Fisheries and Aquaculture Sciences, Lowestoft, UK & 2009 Elsevier Ltd. All rights reserved.

Introduction A very wide variety of solid wastes have been dumped at sea or discharged into the oceans via pipelines or rivers, either deliberately or as a consequence of other activities. The main solid materials involved are dredged material, particulate wastes from sand/gravel extraction and land reclamation, industrial wastes, including mining wastes and munitions, and plastics and litter. Regulation

The London Convention 1972 has regulated dumping at sea of these materials on a global basis since it came into force in 1974 and currently has 83 contracting parties (mid-2008). In 1996, following a detailed review that began in 1993, the Contracting Parties to the London Convention 1972 agreed a Protocol to the Convention to update and modernize it. This came into force for those states that had ratified the Protocol on 24 March 2006 after the 27th state ratified it and currently (mid-2008) has 34 contracting parties. Regional conventions also regulate the dumping of these materials in some parts of the world, for example, the OSPAR Convention 1992 for the North-East Atlantic, the Helsinki Convention for the Baltic (updated in 2004), the Barcelona Convention 1995 for the Mediterranean, the Cartagena Convention 1983 for the Caribbean, and the Noumea Convention 1986 for the South Pacific. Where materials are discharged into the oceans via pipelines or rivers, this usually falls under national regulation but regional controls or standards may strongly influence this regulation, for example, within the countries of the European Union. When human activities introduce solid materials as a by-product of those activities, this is commonly subject only to national regulation, if any at all. Impacts

All solid wastes have significant physical impacts at the point of disposal when disposed of in bulk. These impacts can include covering of the seabed, shoaling,

and short-term local increases in the levels of watersuspended solids. Physical impacts may also result from the subsequent transport, particularly of the finer fractions, by wave and tidal action and residual current movements. All these impacts can lead to interference with navigation, recreation, fishing, and other uses of the sea. In relatively enclosed waters, the materials with a high chemical or biological oxygen demand (e.g., organic carbon-rich) can adversely affect the oxygen regime of the receiving environment while materials with high levels of nutrients can significantly affect the nutrient flux and thus potentially cause eutrophication. Biological consequences of these physical impacts can include smothering of benthic organisms, habitat modification, and interference with the migration of fish or shellfish due to turbidity or sediment deposition. The toxicological effects of the constituents of these materials may be significant. In addition, constituents may undergo physical, chemical, and biochemical changes when entering the marine environment and these changes have to be considered in the light of the eventual fate and potential effects of the material.

Dredged Material Dredging

Dredging is undertaken for a variety of reasons including navigation, environmental remediation, flood control, and the emplacement of structures (e.g., foundations, pipelines, and tunnels). These activities can generate large volumes of waste requiring disposal. Dredging is also undertaken to win materials, for example, for reclamation or beach nourishment. Types of material dredged can include sand, silt, clay, gravel, coral, rock, boulders, and peat. Mining for ores is considered further within the section on industrial wastes. Most dredging activities, particularly hydraulic dredging, generate an overspill of fine solid material as a consequence of the dredging activity. The particle size characteristics of this material and the hydrodynamics of the site will determine how far it may drift before settling and this will influence the extent and severity of any physical impact outside the immediate dredging site. The chemical characteristics of this material have to be taken into consideration in any risk assessment of the consequences to the environment.


519

520

POLLUTION, SOLIDS

Dredged Material

Waterborne transport is vital to domestic and international commerce. It offers an economical, energy efficient and environmentally friendly transportation for all types of cargo. Dredging is essential to maintain navigable depths in ports and harbors and their approach channels in estuaries and coastal waters (maintenance) as well as for the development of port facilities (capital). Maintenance dredged materials tend generally to consist of sands and silts, whereas capital dredged material may include any of the materials mentioned under dredging. Dredged material can be used for land reclamation or other beneficial uses but most of that generated in marine areas is disposed of in estuaries and coastal waters by dumping from ships or barges. The disposal of dredged material is the largest mass input of wastes deposited directly into the oceans. However, it does not follow that dredged material is the most environmentally significant waste deposited in the ocean, since much dredged material disposal is relocation of sediments already in the marine environment, rather than fresh input. London Convention data indicate that Contracting Parties dumped around 400 million tonnes dredged material in 2003. However, not all coastal states have signed and ratified the Convention and less than half of the Contracting Parties regularly report on their dumping activities. Data from the International Association of Ports and Harbours indicated that in 1981 around 580 million tonnes of dredged material were being dumped each year and that survey only had data from half the countries contacted. Thus, it would seem likely that actual quantities of dredged material dumped in the world’s ocean could be up to 1000 million tonnes per annum. Regulation

The Specific Guidelines for Assessment of Dredged Material (SGADM) of the global London Convention 1972 is widely accepted as the standard guidance for the management of dredged material disposal at sea. These guidelines were last updated in 2000 and are available on the London Convention website. Regional conventions will often have their own guidance documents that will usually be more specific to take account of local circumstances, for example, The Guidelines for the Management of Dredged Material of the OSPAR Convention last revised in 2004. Characterization

Dredged material needs to be characterized before disposal as part of the risk assessment procedure if

environmental harm is to be minimized. This will usually require consideration of its physical and chemical characteristics and may require assessment of its biological impacts, either inferred from the chemical data, or directly through the conduct of tests for toxicity, bioaccumulation, etc. The physical properties and chemical composition of dredged materials are highly variable depending on the nature of the material concerned and its origin, for example, grain size, mineralogy, bulk properties, organic matter content, and exposure to contamination. Impacts

In considering the impacts of dredged material disposal, a distinction may be made between natural geological materials not generally exposed to anthropogenic influences, for example, material excavated from beneath the seafloor in deepening a navigation channel and those sediments that have been contaminated by human activities. Although physical, and associated biological, impacts may arise from both types of materials, chemical impacts and their associated biological consequences are usually of particular concern from the latter. In common with the other materials covered in this article, the physical impact of dredged material disposal is most often the most obvious and significant impact on the aquatic environment. The seabed impacts can be aggravated if the sediment characteristics of the material are very different from that of the sediment at the disposal site or if the material is contaminated with debris such as pieces of cable, scrap metal, and wood. Most of the material dredged from coastal waters and estuaries is, by its nature, either uncontaminated or only lightly contaminated by human activity (i.e., at, or close to, natural background levels). Sediments in enclosed docks and waterways are often subject to more intense contamination, both because of local practices, and the reduced rate of flushing of such areas. This contamination may be by metals, oil, synthetic organics (e.g., polychlorinated biphenyls, polyaromatic hydrocarbons, and pesticides) or organometallic compounds such as tributyl tin. The latter has been a major concern in recent years due to its previous widespread use in antifouling paints, its consequential occurrence in estuarine sediments in particular, and its wide range of harmful effects to many marine organisms. In recognition of this concern, the international community agreed the International Convention on the Control of Harmful Anti-fouling Systems on Ships, which was adopted on 5 October 2001. This will prohibit the use of harmful organotins in antifouling paints used on


POLLUTION, SOLIDS

521

ships and will establish a mechanism to prevent the potential future use of other harmful substances in antifouling systems. The convention will enter into force on 17 September 2008. However, it is generally the case that only a small proportion of dredged material is contaminated to an extent that either major environmental constraints need to be applied when depositing these sediments or the material has to be removed to land for treatment or to specialized confinement facilities.

Island reclamation project initiated in 1999 was projected to require 220 Mm3 of material over 3 years. However, in 2003 Singapore estimated that it needed 1.8 billion cubic meters of sand over the following 8 years for reclamation works including Tuas View, Jurong Island, and Changi East. The planned Maasvlakte 2 extension of Rotterdam Port in the Netherlands will reclaim some 2000 ha and require 400 Mm3 of sand with construction due to take place between 2008 and 2014.

Beneficial Uses

Regulation

Dredged material is increasingly being regarded as a resource rather than a waste. Worldwide, around 90% of dredged material is either uncontaminated or only lightly contaminated by human activity so that it can be acceptable for a wide range of uses. The London Convention SGADM recognizes this and requires that possible beneficial uses of dredged material be considered as the first step in examining dredged material management options. Beneficial uses of dredged materials can include beach nourishment, coastal protection, habitat development or enhancement and land reclamation. Operational feasibility is a crucial aspect of many beneficial uses, that is, the availability of suitable material in the required amounts at the right time sufficiently close to the use site. PIANC produced practical guidance on beneficial uses of dredged material in 1992 that is currently (2008) in the process of being revised.

National authorities generally carry out the regulation of sand and gravel extraction and they may have guidance on the environmental assessment of the practice. There is currently no accepted international guidance other than for the Baltic Sea area under the Helsinki Convention and the North-East Atlantic under the OSPAR Convention that has adopted the ICES guidance.

Sand/Gravel Extraction Introduction

Sand and gravel is dredged from the seabed in various parts of the world for, among other things, land reclamation, concreting aggregate, building sand, beach nourishment, and coastal protection. The largest producer in the world is Japan at around 80–100 Mm3 per year, Hong Kong at 25–30 Mm3 per year, the Netherlands and the UK regularly producing some 20–30 Mm3 per year, and Denmark, the Republic of Korea, and China lesser amounts. Reclamation with Marine Dredged Sand and Gravel

In recent years, some very large land reclamations have taken place for port and airport developments particularly in Asia. For example, in Hong Kong some 170 Mm3 was placed for the new artificial island airport development with another 80 Mm3 used for port developments over the period 1990–98. Demand projections indicate a need for a further 300 Mm3 by 2010. Also, in Singapore the Jurong

Impacts

The environmental impacts of sand and gravel dredging depend on the type and particle size of the material being dredged, the dredging technique used, the hydrodynamic situation of the area and the sensitivity of biota to disturbance, turbidity, or sediment deposition. Screening of cargoes as they are loaded is commonly employed when dredging for sand or gravel to ensure specific sand:gravel ratios are retained in the dredging vessel. Exceptionally, this can involve the rejection of up to 5 times as much sediment over the side of the vessel as is kept as cargo. When large volumes of sand or gravel are used in land reclamation, the runoff from placing the material may contain high levels of suspended sediment. The potential effects of this runoff are very similar to the impacts from dredging itself. The ICES Cooperative Research Report No. 247 published in 2001 deals with the effects of extraction of marine sediments on the marine ecosystem. It summarizes the impacts of dredging for sand and gravel. An update of this report is due for completion in 2007. Physical impacts The most obvious and immediate impacts of sand and gravel extraction are physical ones arising from the following.

•

Substrate removal and alteration of bottom topography. Trailer suction dredgers leave a furrow in the sediment of up to 2 m wide by about 30 cm deep but stationary (or anchor) suction dredgers may leave deep pits of up to 5 m deep or more. Infill of the pits or furrows created depends


522

•

•

POLLUTION, SOLIDS

on the natural stability of the sediment and the rate of sediment movement due to tidal currents or wave action. This can take many years in some instances. A consequence of significant depressions in the seabed is the potential for a localized drop in current strength resulting in the deposition of finer sediments and possibly a localized depletion in dissolved oxygen. Creation of turbidity plumes in the water column. This results mainly from the overflow of surplus water/sediment from the spillways of dredgers, the rejection of unwanted sediment fractions by screening, and the mechanical disturbance of the seabed by the draghead. It is generally accepted that the latter is of relatively small significance compared to the other two sources. Recent studies in Hong Kong and the UK indicate that the bulk of the discharged material is likely to settle to the seabed within 500 m of the dredger but the very fine material (o0.063 mm) may remain in suspension over greater distances due to the low settling velocity of the fine particles. Redeposition of fines from the turbidity plumes and subsequent sediment transport. Sediment that settles out from plumes will cover the seabed within and close to the extraction site. It may also be subject to subsequent transport away from the site of deposition due to wave and tidal current action since it is liable to have less cohesion and may be finer (due to screening) than undredged sediments.

Chemical impacts The chemical effects of sand and gravel dredging are likely to be minor due to the very low organic and clay mineral content of commercial sand and gravel deposit from geological deposits not generally exposed to anthropogenic influences and to the generally limited spatial and temporal extent of the dredging operations. Biological impacts The biological impacts of sand and gravel dredging derive from the physical impacts described above and the most obvious impacts are on the benthic biota. The consequences are as follows.

•

Substrate removal and alteration of bottom topography. This is the most obvious and immediate impact on the ecosystem. Few organisms are likely to survive intact as a result of passage through a dredger but damaged specimens may provide a food source to scavenging invertebrates and fish. Studies in Europe on dredged areas show very large depletions in species abundance, number of taxa, and biomass of benthos immediately

•

•

following dredging. The recolonization process following cessation of dredging depends primarily on the physical stability of the seabed sediments and that is closely related to the hydrodynamic situation. Significant recovery toward the original faunal state of the benthos depends on having seabed sediment of similar characteristics to that occurring before dredging. The biota of mobile sediments tends to be more resilient to disturbance and able to recolonize more quickly than the more long-lived biota of stable environments. Generally, sands and sandy gravels have been found to have recovery times of 2–4 years. However, recovery may take longer where rare slow-growing animals were present prior to dredging. The fauna of coarser deposit are likely to recover more slowly, taking from 5 to 10 years. Recolonizing fauna may move in from adjacent undredged areas or result from larval settlement from the water column. Creation of turbidity plumes in the water column. As these tend to be limited in space and time, they are not likely to be of great significance for water column fauna unless dredging takes place frequently or continuously. Redeposition of fines from the turbidity plumes and subsequent sediment transport. Deposition of fines may smother the benthos in and around the area actually dredged but may not cause mortality where the benthos is adapted to dealing with such a situation naturally, for example, in areas with mobile sediments. The deposited fines may be transported away from the site of deposition and affect more distant areas, particularly if the deposited sediment is finer than that naturally on the seabed as a result of screening.

Industrial Solids Introduction

Industries based around coasts and estuaries have historically used those waters for waste disposal whether by direct discharge or by dumping at sea from vessels. Many of these industries do not generate significant quantities of solid wastes but some, particularly mining and those processing bulk raw materials, may do so. The sea may be impacted by mining taking place on land, where waste materials are disposed into estuaries and the sea by river inputs, direct tipping, dumping at sea or through discharge pipes. Examples include china clay waste discharged into coastal waters of Cornwall (southwest England) via rivers,


POLLUTION, SOLIDS

colliery waste dumped off the coast of northeast England, and tailings from metalliferous mines in Norway and Canada tipped into fjords and coastal waters. Alternatively, mining may itself be a waterborne activity, such as the mining of diamonds in Namibian waters or tin in Malaysian waters, with waste material discharged directly into the sea. Fly ash has been dumped at sea off the coast of northeast England since the early 1950s, although this ceased in 1992, and was also carried out off the coast of Israel in the Mediterranean Sea between 1988 and 1998. Fly ash is derived from ‘pulverized fuel ash’ from the boilers of coal power stations but is mainly ‘fly ash’ extracted from electrostatic filters and other air pollution control equipment of coaland oil-fired power stations. This fine material is mainly composed of silicon dioxide with oxides of aluminum and iron but may carry elemental contaminants condensed onto its surfaces following the combustion process. Surplus munitions, including chemical munitions, were disposed of at sea in various parts of the world in large quantities after both World Wars I and II. In addition, a number of countries have made regular but smaller disposals at sea since 1945. In earlier times much of this material was deposited on continental shelves, but since around 1970 most disposals have taken place in deep ocean waters. OSPAR has collated information from its contracting parties on munitions disposal into a database that is available on its website. Regulation

The London Convention 1972 has regulated the dumping of industrial wastes at sea since it came into force in 1974. Annex I of the Convention was amended with effect from 1 January 1996 to ban the dumping at sea of industrial wastes by its Contracting Parties. However, not all coastal states have signed and ratified the Convention and less than half of the Contracting Parties regularly report on their dumping activities. The 1996 Protocol to the London Convention restricts the categories of wastes permitted to be considered for disposal at sea to the 7 listed in Annex 2 of the Protocol, which is very similar to the amended Annex I of the Convention referred to above. The protocol was amended in late 2006 to allow the sequestration of carbon dioxide in sub-seabed geological structures as one of a range of climate change mitigation options. In the period 2000–04 Japan dumped some 1.0 million tonnes per annum of bauxite residues into the sea. This activity was controversial as some contracting parties to the Convention regarded the material as industrial

523

waste, and thus banned from disposal, while Japan maintained that it was an Inert Material of Natural Origin and thus allowed to be disposed of at sea. However, Japan announced in 2005 that it would phase out this disposal by 2015.

Characterization

Like dredged material, these materials need to be characterized before disposal can be permitted to ensure that significant environmental harm is not caused as a result of its disposal.

Impacts

The impacts of the disposal of solid industrial wastes into the sea are as described in the introduction, although chemical and biological impacts may be more significant than for some other solid wastes due to the constituents of the materials. However, physical impacts are usually dominant due to the bulk properties of these materials. Fly ash has a special property (pozzolanic activity) that causes the development of strong cohesiveness between the particles on contact with water. This leads to aggregation of the particles and the formation of concretions that may cover all or part of the seabed.

Beneficial Uses

Beneficial uses can often be found for these materials but depend strongly on the characteristics of the material and the opportunities for use in the area of production. Inorganic materials may be used for example for fill, land reclamation, road foundations, or building blocks, and organic materials may be composted, used as animal feed or as a fertilizer.

Plastics and Litter Introduction

There has been growing concern over the last decade or two about the increasing amounts of persistent plastics and other debris found at sea and on the coastline. Since the 1940s there has been an enormous increase in the use of plastics and other synthetic materials that have replaced many natural and more degradable materials. These materials are generally very durable and cheap but his means that they tend to be both very persistent and readily discarded. They are often buoyant so that they can accumulate in shoreline sinks.


524

POLLUTION, SOLIDS

Regulation

A number of international conventions including the London Convention 1972, MARPOL 73/78, and regional seas conventions ban the disposal into the sea of persistent plastics. Annex V of the International Convention for the Prevention of Pollution from Ships 1973, as modified by the 1978 Protocol (known as MARPOL 73/78), states that ‘‘the disposal into the sea of all plastics, including but not limited to synthetic ropes, synthetic fishing nets and plastic garbage bags, is prohibited.’’ An exception is made for the accidental loss of synthetic fishing nets provided all reasonable precautions have been taken to prevent such loss. This Annex came into force in 1988. Guidelines for implementing Annex V of MARPOL 73/78 suggest best practice for waste handling on vessels. Many parties to MARPOL 73/ 78 have built additional waste reception facilities at ports and/or expanded their capacity in order to encourage waste materials to be landed rather than disposed of into the sea. Annex I of the London Convention 1972 prohibits the dumping of ‘‘persistent plastics and other persistent synthetic materials, for example netting and ropes, which may float or remain in suspension in the sea in such a manner as to interfere materially with fishing, navigation or other legitimate uses of the sea.’’ Since these materials are not listed in Annex 2 of the 1996 Protocol to the London Convention, this instrument also bans them. All land-based sources of these materials fall under national regulation. Impacts

Persistent plastics and other synthetic materials present a threat to marine wildlife, as well as being very unsightly and potentially affecting economic interests in recreation areas. They are a visible reminder that the ocean is being used as a dump for plastic and other wastes. These materials also present a threat to navigation through entangling propellers and blocking of cooling water systems of vessels. In 1975 the US Academy of Sciences estimated that 6.4 million tonnes of litter was being discarded each year by the shipping and fishing industries. According to other estimates in a 2005 UNEP report some 8 million items of marine litter are dumped into seas and oceans everyday, approximately 5 million of which (solid waste) are thrown overboard or lost from ships. It has also been estimated that over 13 000 pieces of plastic litter are floating on every square kilometer of ocean today. The UNEP Regional Seas Programme is currently (2006) developing a series of global and regional activities aimed at controlling, reducing, and abating the problem. The debris that is

most likely to be disposed of or lost can be divided into three groups.

Fishing gear and equipment Nets that are discarded or lost can continue to trap marine life whether floating on the surface, on the bottom, or at some intermediate level. Marine mammals, fish, seabirds, and turtles are among the animals caught. In 1975 the UN Food and Agriculture Organization estimated that 150 000 tonnes of fishing gear was lost annually worldwide.

Strapping bands and synthetic ropes These are used to secure cargoes, strap boxes, crates, or packing cases and hold materials on pallets. When pulled off rather than cut, the discarded bands may encircle marine mammals or large fish and become progressively tighter as the animal grows. They also entangle limbs, jaws, heads, etc., and affect the animal’s ability to move or eat. Plastic straps for four or six packs of cans or bottles can affect smaller animals in a similar way.

Miscellaneous other debris This covers a wide variety of materials including plastic bags or sheeting, packing material, plastic waste materials, and containers for beverages and other liquids. There are numerous studies of turtles, whales, and other marine mammals that were apparently killed by ingesting plastic bags or sheeting. Turtles appear particularly vulnerable to this type of pollution, perhaps by mistaking it for their normal food, for example, jellyfish. A potentially more serious problem may be the increasing quantities of small plastic particles widely found in the ocean, probably from plastics production and insulation, and packing materials. The principal impacts of this material are via ingestion affecting animals’ feeding digestion processes. These plastic particles have been found in 25% of the world’s sea bird species in one study and up to 90% of the chicks of a single species in another study.

See also Antifouling Materials. Benthic Organisms Overview. International Organizations. Law of the Sea. Marine Mammal Overview. Marine Policy Overview. Pollution: Approaches to Pollution Control. Pollution: Effects on Marine Communities. Seabirds as Indicators of Ocean Pollution.


POLLUTION, SOLIDS

Further Reading Arnaudo R (1990) The problem of persistent plastics and marine debris in the oceans. In: Technical Annexes to the GESAMP Report on the State of the Marine Environment, UNEP Regional Seas Reports and Studies No. 114/1, Annex I, 20pp. London: UNEP Duedall IW, Ketchum BH, Park PK, and Kester DR (1983) Global inputs, characteristics and fates of oceandumped industrial and sewage wastes: An overview. In: Duedall IW, Ketchum BH, Park PK, and Kester DR (eds.) Wastes in the Ocean, Industrial and Sewage Wastes in the Ocean, vol. 1, pp. 3--45. New York, NY: Wiley. IADC/CEDA (1996–99) Environmental Aspects of Dredging Guides 1–5. The Hague: International Association of Dredging Companies. IADC/IAPH (1997) Dredging for Development. The Hague: International Association of Dredging Companies. ICES (2001) Effects of extraction of marine sediments on the marine ecosystem. International Council for the Exploration of the Sea, Cooperative Research Report 247. Kester DR, Ketchum BH, Duedall IW, and Park PK (1983) The problem of dredged material disposal. In: Kester DR, Ketchum BH, Duedall IW, and Park PK (eds.) Wastes in the Ocean, Dredged-Material Disposal in the Ocean, vol. 2, pp. 3--27. New York: Wiley.

525

Newell RC, Seiderer LJ, and Hitchcock DR (1998) The impact of dredging on biological resources of the seabed. Oceanography and Marine Biology Annual Review 36: 127--178. Thompson RC, Olsen Y, Mitchell RP, et al. (2004) Lost at sea: Where is all the plastic? Science 304: 838. United Nations Environment Programme (2005) Marine Litter: An Analytical overview. Nairobi: UNEP.

Relevant Websites http://www.helcom.fi – Helsinki Commission: Baltic Marine Environment Protection Commission. http://www.ices.dk – International Council for the Exploration of the Sea. http://www.londonconvention.org – London Convention 1972. http://www.ospar.org – OSPAR Commission. http://www.pianc-aipcn.org – PIANC. http://www.unep.org – United Nations Environment Programme (UNEP). http://www.woda.org – World Organisation of Dredging Associations.


POLLUTION: APPROACHES TO POLLUTION CONTROL J. S. Grayw, University of Oslo, Oslo, Norway J. M. Bewers, Bedford Institute of Oceanography, Dartmouth, NS, Canada & 2009 Elsevier Ltd. All rights reserved.

Approaches applied to the control of pollution have altered substantially over the last four decades. Historically, emphasis was given to management initiatives to ensure that damage to the marine environment was avoided by limiting the introduction of substances to the sea. This was typified by the attention given to contaminants such as mercury and oil in early agreements for the prevention of marine pollution. Marine environmental protection was achieved through prior scientific evaluations of the transport and effects of substances proposed for disposal at sea and defining allowable amounts that were not thought to result in significant or unacceptable effects. This reflects largely a management and control philosophy. In the closing stages of the twentieth century, however, the philosophy underlying pollution control has undergone substantial revision. Recent policy initiatives rely less on scientific assessments and place greater emphasis on policy and regulatory controls to restrict human activities potentially affecting the marine environment. During this period of change, practical pollution control and avoidance procedures have been adapted to improve their alignment with these new policy perspectives. Simultaneously, it has been widely recognized that pollutants represent only part of the problem. Other human activities such as overexploitation of fisheries, coastal development, land clearance, and the physical destruction of marine habitat are equally important, and often more serious threats to the marine environment. In recent years, the concept of marine pollution has been broadened to consider the adverse effects on the marine environment of all human activities rather than merely those associated with the release of substances. This is a most positive development, partly influenced by improved scientific understanding that has led to an improved balance of attention among the sources of environmental damage and threats.

w

Deceased

526

Background In this article, the term ‘pollution’ implies adverse effects on the environment resulting from human activities. This is consistent with, but broader than, the definition of pollution formulated by the United Nations Joint Group of Experts on Marine Environmental Protection (GESAMP) in 1969 that is restricted to adverse effects associated with the introduction of substances to the marine environment from human activities. The term ‘contamination’ infers augmentation of natural levels of substances in the environment but without any presumption of associated adverse effects. Indeed, early approaches to marine pollution prevention reflected the distinction between these terms while more recent approaches are based on more or less identical interpretations of these expressions with both implying adverse effects.

Early Agreements on Marine Pollution Prevention The earliest international marine pollution prevention agreements of the modern era were the Oslo and London conventions of 1972. These conventions were developed at the same time as the heightened awareness of marine pollution issues led to the first major international conference on the topic, the United Nations Conference on the Human Environment, that took place in Stockholm in the same year. Both the original formulation of the two conventions and the results of this conference reflect a commitment to management actions toward the prevention of marine pollution caused primarily by the release of contaminants from human activities. The Oslo and London conventions adopted ‘black’ and ‘gray’ lists of substances and a set of measures to prevent pollution resulting from the dumping of wastes and other matter into the sea. Black list substances are essentially prohibited substances that may not be dumped in the ocean except in trace amounts. Gray list substances are those requiring specific special care measures to be considered in judging their suitability for disposal at sea. In addition, these conventions require that all candidate materials for dumping at sea require a prior assessment to ensure that they do not cause significant adverse effects on the marine environment or pose unacceptable risks to human health.


POLLUTION: APPROACHES TO POLLUTION CONTROL

In light of the rate of introduction of new chemicals into the market economy, flexibility is required in the assignment of substances to ‘black’ and ‘gray’ lists. Thus there is need for an international mechanism for reviewing and updating the list of substances based on an evaluation of their properties. GESAMP provides such a service for the assessment of hazards posed by chemicals transported by ships. There are no similar mechanisms for substances either dumped at sea or entering the marine environment from land-based activities. Meanwhile, each year around a thousand new chemicals are being produced in volume. Their toxicity to a sufficiently wide spectrum of marine species cannot be ensured before use and thus there are constant surprises about the effects of chemicals that were originally thought to entail low risk. The black and gray list approach is too simplistic and does not have the necessary supporting mechanisms to make it reasonably effective. The ocean has the ability to assimilate some finite amount of most substances without adverse effect consistent with the concept of contamination as distinct from pollution. Thus, while not representing a wholly scientific approach, the adoption of these conventions was a major step forward in the introduction of management measures to minimize the risks of marine pollution. However, as will be demonstrated, more recently perceived deficiencies in the provisions of these agreements resulted in their later revision.

Early Approaches to Marine Environmental Protection The oldest strategy for protecting the marine environment from the adverse effects of the disposal of waste in the sea is that based on the application of ‘water quality standards’. This concept was borrowed from practices in freshwater environments, such as rivers, to which it had been applied successfully. The use of water quality standards is based on an assumption that the levels at which contaminants become damaging are well established. Even for chemicals having known effects, for which the severity of effect is proportional to exposure with an assumed threshold for the induction of adverse effects, this assumption has been shown repeatedly to be erroneous as more is learnt about their properties and interactions (see later discussion of the effects of tributyltin (TBT)). For some contaminants, no ‘safe’ level can be established from entirely scientific considerations because they are postulated to pose risks of adverse effect at any concentration. Such substances have what are termed ‘stochastic effects’

527

where the probability of an adverse effect is a function of exposure without any assumption of threshold. The regulation and management of radionuclides and the effects of nuclear practices are based on a postulation of stochastic effects at low doses. A further problem is that the concentrations of contaminants in the marine environment are frequently lower than in fresh water, making measurement more difficult. In coastal areas, the concentrations of heavy metals, for example, are low and thus are difficult to monitor. Chemical contaminants of most concern are generally particlereactive with a strong tendency to attach to particles that end up in sediments, especially in depositional areas. Accordingly, marine sediments are usually a more appropriate focus for assessing the quality of the environment and for monitoring than seawater. However, sediments accumulate relatively slowly and, even where deposition rates are high, biological activity tends to mix sediment layers in the vertical, thus smearing out the record of particle accumulation and of the contaminants co-deposited with particles.

Scientific Perspectives There has long been recognition by scientists that protection of the marine environment per se is inappropriate. The overall approach to environmental protection should be holistic and take account of all human activities and their effects on all compartments of the environment – land, sea, and air. It has similarly been noted that contemporary government and intergovernmental arrangements and structures are inappropriately designed for such a task because they are segregated by development sector, and, often, approaches to environmental protection are considered compartment by compartment rather than comprehensively. The development of protection measures for specific environmental compartments is entirely appropriate but it should follow, rather than precede, the formulation of a holistic framework for environmental management and protection. One of the longest-standing management systems is the radiological protection system, largely developed by the International Commission on Radiological Protection (ICRP), primarily for the protection of human health from the effects of ionizing radiation. This is a scientifically based system that has pioneered concepts such as justification and optimization that require demonstration of the net benefits to society of new practices and minimization of the additional exposures to radiation resulting


528


from all practices. Yet, this same system is predicated, on weakly supported arguments, that if human health is adequately protected, protection of the environment is ensured. While this assumption has been shown to be invalid in some rather special instances, only recently has there been significant professional scientific pressure to extend the protective focus of the system of radiological protection specifically to the environment.

Recent Changes in Policy Perspectives There have also been some significant shifts in policy and management perspectives regarding marine pollution during the last three decades. In the 1960s and 1970s, primary concerns about marine pollution were expressed as concerns about chemical substances disseminated by human activities. With growing evidence of the physical effects of human activities on the land environment and increased public desires to protect the environment, especially areas of natural beauty and wildlife abundance, policy perspectives regarding the range of human activities that could cause adverse effects on the marine environment broadened considerably. Indeed, the term ‘marine pollution’ became perceived as a much broader topic than that defined by the Joint Group of Experts on the Scientific Aspects of Marine Environmental Protection (GESAMP) in 1969. This led to greater consideration of the physical effects on the marine environment of coastal development and watershed activities and resource exploitation, for example. However, this broadening in perspective came about relatively slowly and preoccupation with chemical contaminants was, and remains, evident in international agreements of recent years such as the Global Programme of Action on the Protection of the Marine Environment from Land-Based Activities concluded in 1995. Indeed, the process leading up to this agreement resulted in the adoption of the term ‘activities’ as the final word as a replacement for ‘sources’ at a relatively late stage. Nevertheless, increasingly concerns became extended toward manifestations of human activities on the marine environment other than those solely associated with chemical contaminants. Despite the foregoing, there has been little policy change regarding the adoption of holistic environmental management. Indeed, most of the international and national instruments for the protection of the marine environment are just that – they do not specifically consider the effects of human activities on other environments. It is this fact that creates the

impression that the marine environment is being accorded preferential protection, at least in the context of international agreements. This may well be due to the largely international nature of ocean space, whereas other environments, particularly land and fresh water, lie within national jurisdiction making governments less inclined to reaching international agreements potentially infringing on their sovereignty. Nevertheless, the fact that the marine environment has been the prime subject for the initial advancement and adoption of new policy initiatives, especially the precautionary approach, supports the impression that the sea is being accorded a greater degree of protection than other compartments of the environment. There was also a growing perception among the public, which soon became part of the policy perspective, that in some way previous regulatory approaches to environmental protection had been a failure. Whether this was associated with some disillusionment about the benefits and efficacy of science is a matter of conjecture, but it was abundantly clear that scientists working on behalf of governments or industrial proponents were regarded with skepticism if not outright suspicion when giving professional judgments on environmental matters. The growth in membership and advocacy of the various green lobbies throughout the world is a reflection of this distrust. This, in turn, led to demands for more stringent measures, including extreme policy decisions, to reduce the effects of anthropogenic activities on the environment. More recently advocated approaches to the control of marine pollution reflect these altered perspectives.

Recent Approaches to Marine Environmental Protection A more recent approach to pollution prevention was the so-called ‘best available technology’ strategy that requires the best technology to be applied in human practices in order to minimize associated effects on the environment. At first sight, this approach appears to be logical and appropriate, but it is flawed fundamentally. It provides neither a guarantee of environmental protection nor of the effective use of resources. This approach may result in ineffective environmental protection because the substances being regulated in this way still have adverse effects on the environment, even at the lower release rates achieved. This results in a waste or misdirection of resources. On the other hand, the approach may result in far greater expenditure of effort than that required to prevent environmental damage, thereby also being wasteful of resources that could be used



to better purpose. The point is that, in itself, such an approach achieves little without the knowledge base that allows the tailoring of technology to the needs of society while providing appropriate protection of the environment. A good example is insisting on nitrogen removal for all sewage discharged into the coastal areas of Europe. Primary treatment and phosphorus removal are relatively cheap. Nitrogen treatment is expensive and the amounts of nitrogen discharged in human sewage are small compared to the huge amounts in agricultural wastes that are washed down rivers. In areas with poor water exchange that suffer from eutrophication (enhanced biological production associated with adverse effects such as increased light attenuation, toxic effects on organisms, and increased oxygen demand), such as the inner Oslofjord, there is no doubt that nitrogen treatment of sewage is warranted. However, it is a waste of money to apply the best available sewage technology in other areas where there is no significant risk of eutrophication because there exists sufficient dispersion to ensure that the nutrients are assimilated with little change in the rates and distribution of primary production. There may not be appropriate technology to solve some waste problems, so that even the best contemporarily available technology remains wholly inadequate. For example, some organic chemicals are known to result in widespread effects even at very low concentrations so that contemporary technology, no matter how effective and expensive, is unable to prevent these effects. Various attempts have been made to improve the best technology approach by referring to terms such as ‘best practical technology’, but all these derivatives suffer from similar difficulties in achieving environmental protection at optimal cost. One of the more recent approaches to pollution prevention that has been widely advocated and adopted in agreements during the last decade is the so-called ‘precautionary approach’. Initially, it had been promoted in the form of ‘the precautionary principle’. This appears to have been an outgrowth of a policy development in Germany called the Vorsorgeprinzip (or literally the ‘principle of foresight’). In its original form, the explanation adopted by the German Ministry of the Environment avoided many of the pitfalls that later became intrinsic components of its application in other arenas. The German version, for example, states that not all effects should be treated as representing significant damage and that all risks cannot be avoided. It also placed considerable emphasis on science as a basis for defining risks whose acceptability could be judged in management and policy contexts. There have been many subsequent definitions of the precautionary approach. Such revised versions

529

were later adopted as legal instruments in: the North Sea Declaration of 1987; the declaration adopted at the Rio Conference on Environment and Development (UNCED) in 1992; and several other international agreements, most notably the UN Law of the Sea Convention and the Global Plan of Action for the Prevention of Marine Pollution from LandBased Activities in 1995. The general form of these new formulations has been given as: When an activity raises threats of serious or irreversible harm to human health or the environment, precautionary measures that prevent the possibility of harm shall be taken even if the causal link between the activity and the possible harm has not been proven or the causal link is weak and the harm is unlikely to occur. (Holm and Harris, 1999)

Essentially, as later developed from its German origin, the precautionary approach implies that any lack of knowledge about the environmental hazards associated with a practice justify the adoption of special precautionary measures. Subsequently, such precaution has been deemed to warrant the adoption of extreme preventive measures including the banning of certain practices or chemicals. Several of the agreements that have adopted the precautionary approach emphasize its application to contaminants that are described as ‘‘persistent, toxic and liable to bioaccumulate.’’ However, threshold values for these three properties at which greater precaution is warranted are not specified, either individually or in combination. This is a fundamental flaw in the application of precaution expressed in this way as all substances have the properties of persistence, toxicity, and liability to bioaccumulate to some degree. Clearly, more precision is needed if the precautionary approach is to be of any practical value. The precautionary approach has been debated and criticized from both practical and philosophical perspectives and is not the panacea for the prevention of environmental problems its exponents claim it is. Its biggest danger is that it makes science essentially redundant; mere suspicion of an effect or lack of complete scientific knowledge is a goodenough argument to initiate bans on the use and dissemination of a substance. It has resulted in the replacement of the black and gray list approaches with so-called ‘reverse lists’ of materials that can be considered for dumping at sea under the Oslo and London conventions. These severely curtail management flexibility in the options available to management with little probability of increased environmental protection as a whole. Following pressure from ‘green’ lobbies, for example, the Swedish government is considering banning the discharge of chemicals unless they are proven to be safe.


530


At the 1996 Esbjerg Ministerial Conference on the North Sea, the participants went one step further, requiring that chemicals be proven not to cause effects before they are allowed to be discharged to the sea. The problem here is that safety is the opposite of risk and nothing is devoid of risk. Accordingly, complete safety can neither be proven nor guaranteed no matter how good the associated science. Are we to stop all human activities because of the inherent risk (lack of absolute safety) they entail? Surely not. Another element of recent approaches to pollution prevention generally is ‘environmental impact assessment’ (EIA). It is a procedure for prior evaluation of the environmental consequences of some proposed human activity such as the construction and operation of a new industrial facility. It is a requirement of organizations of all kinds under European and most national legislation where impacts on the environment may occur. As such, EIA is a useful component of the arsenal of environmental protection measures. Yet, far too frequently, EIAs are simply paper exercises incorporating inadequate accountability if their predictions are wrong and the environment is destroyed. This is deplorable and yet need not be the rule. This concept can be applied fairly widely not only to specific industrial installations but also to potential investments in entire new industries and other human activities such as coastal development and tourism, for example. In this sense, it is somewhat analogous to the justification of practices in radiological protection. New ideas regarding EIAs provide environmental authorities improved means of assessing and controlling potentially damaging activities. The process needs detailed and careful science to:

• • •

make quantitative and realistic predictions of effects; suggest criteria for testing such predictions; and design proper and effective monitoring programs with regulatory feedback.

In this context, it should be noted that the process of preparing EIAs needs to involve all parties, including the green movement. It should not be solely the prerogative of the company concerned and its experts. It must also act as a basis for designing scientifically based assessment and monitoring programs.

Predicting the Effects of Chemicals on the Marine Environment The classical way of predicting effects of chemicals on the marine environment is by first conducting toxicity tests. These usually involve testing a variety

of organisms from bacteria through algae to small animals and then fish. Likely effects are predicted, assuming by extension from freshwater models that a concentration of 10% or 1% of the LC50 (50% lethality concentration) is safe. This does not always work as can be demonstrated by data pertaining to the effects of TBT used as an antifoulant for vessel hulls and marine structures. Typical toxicity data show that concentrations of TBT of the order of 1 mg l 1 induce toxic effects. A level of 1/100th of this value should not lead to negative effects. However, misshapen oysters, round as golf balls, were found in the Blackwater Estuary in the United Kingdom and elsewhere. Through excellent detective work it was later shown that TBT was responsible for such effects even at concentrations below 2 ng l 1. These effects had not been predicted because toxicity tests are generally short-term-lasting, at best, for 48 h. Growth effects occur over much longer periods of time. Other scientists discovered that TBT caused the female gastropod snail Nucella lapillus to develop a penis. The most extreme effect was the appearance of a penis having the same size as the male, a condition known as imposex, with the affected females being infertile. Imposex in the field has since been used as an excellent marker for longterm exposures to TBT (due to the use of TBTcontaining antifouling paints). France was the first country to introduce a ban on the use of such paints on vessels less than 25 m in length because of the value of its shellfish industry. Many other countries subsequently followed suit. These lessons teach us that laboratory toxicity tests are unable to predict the long-term effects of some chemicals. TBT, with its effects on reproductive organs, now falls into a general category known collectively as hormone-disrupting chemicals. There exist a large range of chemicals having little commonality in chemical structure, other than that they are all organic, that produce similar effects. Organic chemicals are now justifiably the clear focus of toxicity research rather than the ‘heavy metals’, which were regarded as the key contaminants in the 1970s and 1980s. Recently, a range of new techniques for assessing effects on individuals have been produced. They are referred to as ‘biomarkers’ that reflect stress in marine organisms. One example of such techniques involves the measurement of an enzyme Ethoxyresorufin-o-deethylase (EROD) in flatfish that is induced by exposures to polycyclic aromatic hydrocarbons (PAHs). Another is the measurement of acetylcholinesterase production that is inhibited by chemical stressors, primarily organophosphates in fertilizers used in agriculture. The successful



development of biomarker techniques suggests that strategies for monitoring the condition of the marine environment should emphasize the application of biological rather than chemical measurements. A suite of biomarkers covering the range of responses from genetic to whole organism, combined with the use of multivariate statistical techniques now in widespread use for the analysis of marine community data, now offer both the sensitivity and efficiency required for impact detection in the environment. These can be followed up by chemical measurements when indicative biological response signals are detected. In this way, science is providing methods needed for an effective environmental management and protection framework.

Uncertainty, Risk Assessment, and Power Analyses Generalized frameworks for the management of activities potentially affecting the marine environment have been devised. These involve initial desk studies of the sources and amounts of chemicals released and physical disturbance planned. These can then be combined with specifications of the physical and chemical properties of substances (i.e., hazards, including the properties of toxicity, persistence, and liability to bioaccumulate) and biological effects information. Such information can then be considered in the context of understanding of the physical, chemical, and biological characteristics of the potentially affected area to yield potential changes in these conditions caused by the human activity proposed including the exposures of marine organisms and humans to chemicals. This will provide a basis for assessing the consequences of the proposed activity including the risks to human health, the effects on marine organisms posed by chemical exposures, and any effects on other marine resources and amenities caused by changed sedimentation rates, altered physical dynamics, etc. This constitutes a risk-assessment process. The risk assessment will indicate what effects are likely and the uncertainties that require to be considered in judging the associated risk to the marine environment, its resources and amenities, and to human health. The risk assessment should incorporate appropriate degrees of pessimism that is the equivalent of conservatism to allow for uncertainties in scientific terms and should correspond to precaution in policy terms. Ultimately, the acceptability of effects and risks is not a scientific matter – it lies within the policy and management spheres (although scientists may be consulted). If

531

certain risks are deemed unacceptable, the regulatory authority will legitimately require additional mitigation measures and the risk assessment incorporating the new measures iterated. If this reveals no significant problem from management and protection perspectives, the conditions and predictions should be used as a basis for compliance and effects monitoring of the activity. In this context, the purpose of environmental monitoring is to ensure that the predicted changes are within expectations and not exceeded. If they are, feedback from monitoring to management should ensure that regulatory constraints are revised to reduce the impact further. Equally important is that the results of monitoring can be used to reveal deficiencies or invalid assumptions in previous risk assessments, thereby offering the benefits of future improvements in predictive ability. Previous environmental-monitoring activities have been widely criticized as being ill-conceived, illconducted, and the results inadequately evaluated. In too few cases have monitoring programs been designed around testable hypotheses that enable rigorous scientific evaluation. Furthermore, there has been an unwillingness to evaluate results periodically to ensure that a program is meeting its objectives and yielding useful information. Far too often, the results of monitoring are archived without benefit of human analysis, thereby constituting a waste of resources. These tend to be more general criticisms but there are also improvements that could be made at a more detailed level. Power analysis to determine the ability of a measurement sequence to detect a change of a given amount is not used sufficiently and greater attention to type II, rather than type I, statistical errors is warranted. A type I error occurs when it is accepted that a harmful effect occurs when it does not. We guard against making this error by accepting at least 95% probability (po0.05) and thereby allowing a 5% error due to pure chance. A type II error, on the other hand, is where it is accepted that a harmful effect has not occurred when in fact it has. Several scientists have argued that this is a far more serious error in the context of marine environmental protection than a type I error and yet is rarely considered. Commonly, the criterion adopted for the probability of type II errors is not 95% but 80%.

Improving Marine Environmental Protection Some of the improvements needed in scientific endeavours supporting marine environmental protection have been outlined. Use of pessimistic


532


approaches to take account of uncertainties, improvements in the design and objectives of monitoring programs, and greater consideration of type II statistical errors are among the more important of these. Scientists have been poor at explaining the benefits and limitations of science not only to environmental managers but more crucially to the public. Indeed, previous failures of environmental protection have often been attributed to scientific deficiencies rather than to those of management faced with compromises between political pressures and environmental protection. This is somewhat analogous to similar conflicts and failures in fisheries management. The green movement has been far better at getting its message across not only to managers and legislators, but more importantly to the general public. Unfortunately, this has led to perception being used as a measure of the severity of environmental damage or threat and to the adoption of unwarranted and extreme policy measures for the resolution of perceived problems. The public, and ultimately governmental, reaction to the proposal to dispose of the used field storage tank, Brent Spar, was out of all proportion to the scale of possible damage to the marine environment. A couple of years have passed and a further $30 million spent, but the platform still lies in a Norwegian fiord demanding surveillance. Herein lies the danger: that extreme measures may not solve existing problems or prevent future problems while placing unnecessarily severe constraints or disincentives on economic and technological development as argued by Holm and Harris. Attention is being diverted from the adoption of rational, considered, and scientifically based approaches to the protection of the environment that will permit the greatest opportunity for social and economic development. Society should be demanding the increased use of prior environmental impact assessments to provide quantitative predictions of what effects are to be expected and of their spatial distribution. These should be backed up by properly designed monitoring programs that have adequate power to detect the changes expected and provide routine feedback to regulatory controls. The necessary science already

exists – what is needed is the will to apply it effectively to ensure that the best possible environmental protection is achieved in the face of the demands for human development.

See also Atmospheric Input of Pollutants. Fiordic Ecosystems. Global Marine Pollution. Metal Pollution. Oil Pollution. Pollution: Effects on Marine Communities. Pollution, Solids. Seabirds as Indicators of Ocean Pollution. Thermal Discharges and Pollution.

Further Reading Bewers JM (1995) The declining influence of science on marine environmental policy. Chemistry and Ecology 10: 9--23. Freestone D and Hay E (1996) The Precautionary Principle and International Law: The Challenge of Implementation, 274pp. The Hague: Kluwer Law International. FRG (1986) Umweltpolitik: Guidelines on Anticipatory Environmental Protection, 43pp. Berlin: Federal Ministry for the Environment, Nature Conservation and Nuclear Safety. GESAMP (1991) Scientific strategies for marine environmental protection. GESAMP Reports and Studies No. 45. London: International Maritime Organization. Gray JS (1998) Risk assessment and management in the exploitation of the sea. In: Calow P (ed.) Handbook of Environmental Risk Assessment and Management, pp. 452--473. Oxford, UK: Blackwell. Holm S and Harris J (1999) Pecautionary principle stifles discovery. Nature 400: 398. Milne A (1993) The perils of green pessimism. New Scientist 1877: 34--37. NRC (1990) Managing Troubled Waters: The Role of Environmental Monitoring. Washington, DC: United States National Research Council, National Academies Press. USEPA (1992) Framework for Ecological Risk Assessment. Washington, DC: Office of Research and Development, United States Environmental Protection Agency. Wynne BG and Mayer S (1993) Science and the environment. New Scientist 1876: 33--35.


POLLUTION: EFFECTS ON MARINE COMMUNITIES R. M. Warwick, Plymouth Marine Laboratory, Plymouth, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2222–2229, & 2001, Elsevier Ltd.

practicalities of sampling different groups of marine organisms, are beyond the scope of this article (but see Further Reading). Rather, it concerns the analysis and interpretation of community data with a view to detecting and monitoring the impacts of pollution and other Man-induced disturbances, and inferring causality.

Introduction Natural communities of all types of marine organisms all over the world are being affected either directly or indirectly by pollution: directly by discharges of industrial and domestic wastes, offshore oil and gas drilling activities for example, and indirectly as a result of atmospheric pollution and global climate change. Some of these community changes are visually obvious in the field, such as the effects on reef corals of increased particle sedimentation or abnormally long periods of high water temperatures (causing bleaching). Others require the analysis of samples brought back to the laboratory, such as the assemblages of organisms living in seabed sediments or the plankton. The determination of changes in the structure of these communities, in particular the sediment-dwelling macrobenthos, is widely used in the detection and monitoring of humans’ impact on the sea for a number of reasons. Attributes of the community level of biological organization reflect integrated environmental conditions over a period of time, whereas methods employing lower levels of organization (biochemical, cellular, physiology of whole organisms) require an experimental approach which reflects the condition of the organisms just at the time of sampling. It is natural communities that are of direct concern, and predicting the consequences for these communities from lower level signals is presently not feasible with any degree of confidence. Monitoring at higher levels of organization, e.g., measuring ecosystem attributes, is often simply not tractable. Community change, especially when measured by multivariate methods (see below) has also been shown to be a very sensitive indicator of environmental conditions. It is not necessary for animals to die for a community response to be elicited, but subtle effects such as shifts in the relative competitive abilities of species or changes in fecundity are detectable. Aspects of survey and sampling design, that will enable valid inferences to be drawn, and the

Analysis of Community Data Typically, community data comprise a (usually) large samples by species matrix (Figure 1), in which the cells are filled with species abundances, biomasses, of some other measure of the relative importance of each species, such as percentage cover of colonial organisms. These matrices are likely to have a high proportion of zero entries. The samples may be taken at a number of sites at one time (spatial data) or at the same site at a number of times (temporal data), or a combination of both. When considering environmental problems operating on larger spatial and temporal scales, such as the effects of eutrophication or global warming over large sea areas, the data may be less quantitative, and comprise simple presence/absence information (species lists). There are essentially three classes of methods for analysing such data. 1. Univariate methods. These reduce all the information on the species composition of a sample to a single coefficient, most commonly a diversity index. The occurrence of ‘pollution indicator’ species also falls into this category. 2. Distributional techniques. Here the relative importance of each species for a single sample is represented by a curve or histogram rather than a single number. Note that in both these categories comparisons between samples are not based on particular species identities, and that two samples could have exactly the same diversity or distributional structure without having a single species in common. This is in contrast to the third method. 3. Multivariate methods. Here comparisons of samples are based on the extent to which they share species at comparable levels of abundance, biomass etc. These similarity coefficients then facilitate either a classification or clustering of samples


533

534

POLLUTION: EFFECTS ON MARINE COMMUNITIES

Sampling stations Species

Abra prismatica Acanthocardia echinata Alcyonium digitatum Ampelisca brevicornis Ampelisca macrocephala Ampharete baltica Ampharete finmarchica Ampharete lindstroemi Ampharete sp. Amphicteis gunneri Amphictene auricoma Amphilocus spencebatei Amphitrite cirrata Amphitritides gracilis Amphiura filiformis ...Etc.

30 36 37 31 03 35 27 25 26 29 32 38 28 33 34 ...Etc. 0 0 0 0 0 0 0 0 3 1 1 18 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

0 0 23 0 10 2 13 11 13 14 31 17 34 0 0 0 0 0 0 0 0 0 1 0 0 0 0 2 0 0 0 0 1 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 6 5 4 4 5 6 2 0 3 5 5 4 0 18 21 18 15 27 18 7 7 34 4 19 17 0 1 4 7 0 3 0 1 1 8 0 3 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 1 0 0 0 0 0 0 0 0 0 4 9 4 11 1 5 7 12 20 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 4 7 3 10 1 7 5 9 2 11

Figure 1 Part of a species by stations data matrix for species abundances of the macrobenthos from stations in the vicinity of the Ekofisk oil rig in the North Sea. There are 39 stations and 174 species in the complete matrix.

into groups that are mutually similar, or an or- used indices include: dination plot in which samples are ‘mapped’ in P such a way that their distance apart reflects their Shanon diversity H 0 ¼ i pi loge pi relative dissimilarity in species composition. ðwhere pi ¼ Xi =N Þ Margalef’s species richness d ¼ ðS 1Þ=loge N Pielou’s evenness J0 ¼ H 0 =loge S Univariate Methods Brillouin’s index H ¼ ð1=N Þloge ðN!=Pi Xi !Þ A variety of different indices can be used as measures P Simpson’s index Do ¼ 1 i fXi ðXi 1Þ=N ðN 1Þg of environmental pollution or disturbance, such as the total number of individuals (N), the total number of species (S), the total biomass (B) and also ratios such as B/N (the average ‘size’ of organisms) and N/S (the average number of individuals per species). These tend to be less informative, however, than measures that describe the way in which the numbers of individuals are divided up among the different species (species diversity indices), or measures of biodiversity based on the degree of taxonomic or phylogenetic relatedness of individuals or species in a sample. Species diversity indices Indices of species (or higher taxon) diversity are amenable to standard univariate statistical analysis to determine, for example, the significance of differences between replicate sets of samples. A bewildering range of indices have been devised which measure attributes related to the number of species for a fixed number of individuals (species richness), the extent to which abundances are dominated by a small number of species (dominance, evenness and equitability indices), or a combination of these. Some commonly

Note that values for all of these, except Simpson’s index, tend to be heavily dependent on sample size (N). This means that it is not valid to compare measured diversity values for samples whose size is not standardized. Also, logs to different bases (e, 2, 10) are used in the literature to calculate these indices, and often the log base used is not stated, which also hampers comparisons. Increasing levels of environmental stress are generally considered to decrease species diversity, richness and evenness. The ‘intermediate disturbance hypothesis’, supported by much empirical evidence, suggests however that diversity is maximal at intermediate levels of disturbance, falling off at very low frequencies and intensities of disturbance due to competitive exclusion between species, and decreasing again at high levels as species become eliminated. Thus, the response to increasing levels of pollution or disturbance is not monotonic, and value judgments concerning the effects of pollutants may be difficult to make (Figure 2). With no way of predicting at what level species diversity should be set under pristine environmental conditions, changes



Shannon Oil spill

Diversity

Evenness

Richness

1977

1978

1979

1980

1981

Figure 2 Changes in Shannon diversity and its components of species richness and evenness for the macrofauna at a station in the Bay of Morlaix, France, taken at approximately 3-monthly intervals spanning the period at which the site was affected by oil pollution from the wreck of the Amoco Cadiz on the Brittany coast. Note the increase in all three of these indices after the pollution incident.

due to pollution can only be assessed by comparisons with reference stations (which may often be difficult to find) or with historical data. Taxonomic relatedness measures A measure of biological diversity ought ideally to say something about how different the inhabitants are from each other. Simply to say whether or not they belong to the same species is clearly insufficient, and recently a variety of different measures have been devised to measure the degree to which species are taxonomically related to each other. Polluted communities usually comprise a limited taxonomic spread of species, whereas under more pristine conditions the species present belong to a wide range of higher taxa. Measures of taxonomic distinctness describe features of this taxonomic spread and they are now beginning to find application in environmental impact assessment. They measure either the average distance apart of all pairs of individuals or species in a sample, traced through a hierarchical taxonomic tree, or the variability in structure across the tree. They overcome many of the problems of species diversity measures noted above, for example, they are independent of sample size, they can utilize simple presence/ absence data (species lists), and in the latter case randomization/permutation procedures can test the null-hypothesis that the species present are a random selection from the full spread of taxa in the regional species pool. ‘Average taxonomic diversity’ (D) is simply the average path length through the hierarchical taxonomic tree between every pair of individuals in a

535

sample. ‘Average taxonomic distinctness’, D * , is D divided by its value when the hierarchical classification collapses to the special case of all species belonging to a single genus, and is more nearly a function of pure taxonomic relatedness of individuals (it is equivalent to dividing D by the Simpson diversity D1). A special case is the use only of presence/absence information for each species, when D and D * reduce to the same statistic, namely Dþ. Another aspect of the taxonomic structure is the ‘evenness’ of the distribution of taxa across the hierarchical taxonomic tree. In other words, are some taxa overrepresented and others underrepresented by comparison with what we know of the species pool for the geographical region? Such a difference in structure should be well reflected in variability of the full set of pairwise distinctness weights making up the average, i.e. the variation in taxonomic distinctness, denoted by Lþ. Pollution indicator species Certain taxa of benthic marine invertebrates are known to increase dramatically in abundance when levels of particulate organic enrichment become abnormally high and have become known as ‘pollution indicator’ species. They tend to be the smaller members of that size category called the macrobenthos, and the larger members of the meiobenthos. Several of these comprise groups of very closely related sibling species, such as the polychaete worms Capitella and Ophryotrocha, the benthic copepod genus Tisbe and the free-living nematode genus Pontonema. No firm protocols have been established with regard to the level at which a community must be dominated by a particular indicator for it to be regarded as polluted, and interpretation of this information remains rather subjective. Distributional Techniques

Diversity curves may take a number of forms. 1. Rarefaction curves are plots of the number of individuals on the x-axis against the number of species on the y-axis. Sample sizes (N) may differ, but the relevant sections of the curves can still be visually compared. 2. Plots of the number of species in 2 geometric abundance classes (i.e., number of species represented by 1 individual, 2–3 individuals, 4–7 individuals, 8–15 individuals etc.). 3. Ranked species abundance (dominance) curves in which species (or higher taxa) are ranked in decreasing order of importance in terms of abundance, biomass etc., and their importance,


536


expressed as a percentage of the total for all species, is plotted against the relevant species rank. 4. k-dominance curves are cumulative ranked abundances, biomasses etc. plotted against species rank, or more usually log species rank. The advantage of dominance curves is that the distribution of species abundances among individuals and the distribution of species biomasses among individuals can be compared on the same terms. Since the two have different units of measurement, this is not possible with diversity indices. This is the basis of the Abundance/Biomass Comparison (ABC) method of assessing disturbance, which has been mainly applied to benthic macrofauna communities. k-dominance curves for abundance and biomass in the same sample are plotted together on the same graph (Figure 3). In undisturbed communities the distribution of numbers of individuals among species is more even than the distribution of numbers of biomass among species, so that the k-dominance curve for biomass lies above that for abundance throughout its length. In grossly polluted communities the reverse is the case, and in moderately polluted ones the two curves are quite coincident and may cross over one or more times. These three conditions are recognizable in single samples and can provide a snapshot of the pollution status of a community without reference to spatial or temporal controls, although the latter are of course desirable. Plotting large numbers of ABC curves can be cumbersome, and the information they contain can be summarized by the W statistic, which is the sum of the B–A values for all species across the ranks, standardized to a common scale so that comparisons can be made between samples with differing numbers of species. When samples are replicated, the W statistic provides a means for testing for significance of observed changes in ABC patterns, using standard univariate procedures.

Multivariate Methods

Multivariate classification, or cluster analysis, aims to find groupings of samples such that those within a group are more similar to each other in biotic composition than samples in different groups. The results of such an analysis are usually represented by a dendrogram (Figure 4A). An ordination, is a map of the samples (Figure 4B), usually in two or three dimensions, in which their distance apart reflects their biological similarity; the samples are not forced into groups, but rather the relationship of each sample with every other is depicted (as far as this is possible in two or three dimensions). There are literally hundreds of clustering methods in existence. Also, several ordination methods have been proposed, each using different forms of the original data and varying in their technique of preserving the true intersample similarities in low-dimensional plots: these include principal components analysis (PCA), principal co-ordinates analysis (PCoA), correspondence analysis and detrended correspondence analysis (DECORANA), and multidimensional scaling (MDS), in particular nonmetric MDS. It is not possible here to give a balanced account of this huge range of techniques, but instead this section focuses on a unified set of protocols based on nonparametric methods which is now gaining worldwide acceptance in the field of environmental impact assessment, and is implemented by the software package PRIMER, developed at the Plymouth Marine Laboratory, UK. Compared with univariate or distributional techniques of data analysis, multivariate methods are much more sensitive in detecting differences between communities, and thus evaluating temporal or spatial change. Pivotal to PRIMER’s approach is the biologically relevant definition of the similarity between two samples and its utilization in simple rank form, i.e., sample A is more similar to sample B than it is to sample C. Statistical assumptions about the data are

0

1

10

s Moderately polluted

Unpolluted 5

om as

e

nc

da

un Ab

s as om i B

1

nce da

Bi

d

Ab un

50

ce an

Ab un

ass om

Bi

Cumulative %

100

5 10 Species rank (log scale)

1

Grossly polluted 5

10

Figure 3 Hypothetical k-dominance curves for species biomass and abundance (ABC plots) showing unpolluted, moderately polluted and grossly polluted conditions.



537

20 30 km E

30

Bray_Curtis similarity

40 50 60 70

2 km

(A)

80 90 100 G GG GE E E E D D D D C C B B CB B C A A A A

(A)

D

(B) C

A A A

D A DD B

C

B C C B B (C) Figure 5 Ekofisk oil rig, North Sea. (A) Map of positions of the 39 sampling stations, coded according to their distance from the active drilling center (W, 43.5 km; &, 1.0–3.5 km; , 250– 1000 m; J, o250 m). (B) Nonmetric MDS ordination based on & transformed species abundance data and the Bray–Curtis similarity measure, using the same coding. (C) PCA of three environmental variables (% mud, log total hydrocarbons and log barium concentration in the sediment) at the same stations. Note the match between (B) and (C).

E E

E E

G

G G G

(B) Figure 4 Macrobenthos from four replicate samples at each of six sitations (A–E, G) in Frierfiord, Norway. (A) Dendrogram for hierarchical clustering based on && transformed species abundance data and the Bray–Curtis similarity measure. (B) Nonmetric MDS ordination based on the same similarity matrix. Note the similarity in the grouping of samples. Stations B, C and D, in the deeper basins of the fiord, suffered from seasonal anoxia.

thus minimized and the resulting nonparametric techniques are of very general applicability. From the starting point of a ranked triangular similarity matrix (usually using Bray–Curtis similarity) between all pairs of samples, the procedures encompass: 1. the display of community patterns by hierarchical agglomerative clustering and nonmetric MDS;

2. identification of the species principally responsible for determining sample groupings (the SIMPER program); 3. statistical tests for differences in community composition in space and time in one-way and two-way layouts: multivariate analogs of analysis of variance (ANOSIM); 4. linking patterns of community difference to patterns of physical and chemical environmental variables at the same locations or times (BIOENV). Although causality of community change can really only be established with certainty by means of controlled experiments, inferences can be drawn


538


A C

D

Oil Spill

E

C

B

F

E G

U

T S O

P

R Q

N L

M

J

I

A

D

B

Oil Spill

T U

H

S

R N

P

Q M

O

K

(A)

L

K J

H

F G I

(B)

Figure 6 Nonmetric MDS for the macrofauna at a station in the Bay of Morlaix, France, taken at approximately 3-monthly intervals spanning the period at which the site was affected by oil pollution from the wreck of the Amoco Cadiz on the Brittany coast (same data as Figure 2). (A) Species level data. (B) The same analysis, but with the species data aggregated up into five major groupings: Annelids, Molluscs, Crustaceans. Echinoderms and ‘others’. Note, in both cases, the community change after the pollution incident, and the gradual evolution of the community back to a new equilibrium state. The pollution response seems to be even more marked at the higher taxon level than the species level.

from the relationships between multivariate patterns in the biological and environmental data, which can be established formally using the BIOENV procedure. In Figure 5, for example, the MDS plot for the macrobenthos in the vicinity of the Ekofisk oil rig in the North Sea closely matches the MDS for those environmental variables that can unequivocally be related to the drilling activity. However, in themselves these multivariate analyses are not intended as measures of biological stress. They do not enable us to put a value judgment (bad, good, neutral) on the community change, in the way that univariate or distributional methods can. The three outer distancezones in the Ekofisk study are indistinguishable in terms of species diversity and k-dominance curves, so do the clear differences in species composition between them revealed by multivariate analysis (Figure 5) really matter? Ways are now being devised for incorporating the full multivariate information into measures of biological stress. A feature of these analyses is that patterns of community change at the species level, in response to pollution or disturbance, are generally reflected at higher taxonomic levels, even up to the level of phylum (Figure 6), due to the fact that related taxa are responding in similar ways. A meta-analysis of a range of well-documented pollution/disturbance incidents on the macrobenthos has shown a commonality of response in terms of phyletic composition in relation to the level of environmental stress. Combining new survey data with these training data, and rerunning the analysis, thus enables the pollution status of the new data to be evaluated on a broadly comparative scale. Multivariate patterns of seriation

of community change in response to natural environmental gradients have been shown to break down with increasing environment stress, and the variability between replicate samples in multivariate terms has also been shown to increase. Both these attributes have been developed into stress measures, the Index of Multivariate Seriation (IMS) and the Index of Multivariate Dispersion (IMD), respectively.

Conclusions The methods described above can be used to detect the biological effects of pollution at a range of spatial and temporal scales ranging from local short-term pollution incidents to basin-wide long-term effects of eutrophication and climate change. For local events, the less mobile components of the biota are perhaps more appropriate for study, such as the macrobenthos and meiobenthos of soft-sediments and the sessile epifauna of hard substrata. Here, with carefully controlled sampling designs, community change can be detected in terms of diversity, distributional curves and multivariate aspects of community composition at the species level. For broader scale comparisons, pelagic organisms such as plankton or fish may be more appropriate for study: Because of the methods of collection these community samples tend to integrate ecological conditions over large areas (see, for example, the Continuous Plankton Recorder Survey of the North-East Atlantic). If benthic components of the biota are to be used, biodiversity measures based on taxonomic relatedness may be more useful than species diversity indicies since they



can utilize species lists rather than strictly quantitative data and are sample-size independent. Also, multivariate analysis at taxonomic levels higher than that of species might be more appropriate, in view of the narrow habitat preferences and restricted geographical distributions of many species.

See also Beaches, Physical Processes Affecting. Continuous Plankton Recorders. Chlorinated Hydrocarbons. Coral Reefs. Ecosystem Effects of Fishing. Eutrophication. Fiordic Ecosystems. Grabs for Shelf Benthic Sampling. Lagoons. Macrobenthos. Mangroves. Meiobenthos. Metal Pollution. Oil Pollution. Pollution, Solids. Rocky Shores. Salt Marshes and Mud Flats. Zooplankton Sampling with Nets and Trawls.

539

Clarke KR and Warwick RM (1994) Change in Marine Communities: an Approach to Statistical Analysis and Interpretation. Plymouth: Plymouth Marine Laboratory. Holme NA and McIntyre AD (eds.) (1984) Methods for the Study of Marine Benthos 2nd edn. Oxford: Blackwell. Magurran AE (1991) Ecological Diversity and its Measurement. London: Chapman and Hall. Sokal RR and Rohlf FJ (1981) Biometry. San Francisco: Freeman. Underwood AJ (1997) Experiments in Ecology: Their Logical Design and Interpretation Using Analysis of Variance. Cambridge: Cambridge University Press. Warwick RM (1993) Environmental impact studies on marine communities: pragmatical considerations. Australian Journal of Ecology 18: 63--80. Warwick RM and Clarke KR (2001) Practical measures of marine biodiversity based on relatedness of species. Oceanographic Marine Biology Annual Review.

Further Reading Clarke KR (1993) Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18: 117--143.


POLYNYAS S. Martin, University of Washington, Seattle, WA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2241–2247, & 2001, Elsevier Ltd.

Introduction Polynyas are large, persistent regions of open water and thin ice that occur within much thicker pack ice, at locations where climatologically, thick pack ice would be expected. Polynyas have a rectangular or oval aspect ratio with length scales of order 100 km; they persist with intermittent openings and closings at the same location for up to several months, and recur over many years. In contrast to polynyas, leads – another open water feature – are long, linear transient features associated with the pack ice deformation, are not restricted to a particular location, and generally have a much smaller area than polynyas. Polynyas occur in both winter and summer. Given that their physical behavior in winter is more complicated than in summer, we begin with the winter case, then follow with a shorter description of their transition to summer. Polynyas can be classified into coastal and openocean polynyas. Costal polynas form where the winter winds advect the adjacent pack ice away from the coast, so that sea water at temperatures close to the freezing point is directly exposed to a large negative heat flux, with the resultant rapid formation of new ice. This new ice is advected away from the coast as fast as it forms. For these polynyas, a typical alongshore length is 100–500 km; a typical offshore length 10–100 km. In contrast, the less common open-ocean polynyas have characteristic diameters of 100 km and are driven by the upwelling of warm ocean water, which maintains a large opening in the pack ice. Because the atmospheric heat loss from the open-ocean polynyas goes into cooling of the water column, they are sometimes called ‘sensible heat’ polynyas; because the heat loss from coastal polynyas goes into ice growth, they are called ‘latent heat’ polynyas. Finally, some polynyas, notably the North Water polynya in Baffin Bay, are maintained by both upwelling and ice advection. Figure 1 shows the locations of some of these polynyas for both hemispheres. In the Northern Hemisphere, most of the polynyas are coastal, with

540

the Kashevarov Bank polynya in the Okhotsk Sea being the only purely open-ocean polynya. The largest coastal polynyas occur in the marginal seas, where the adjacent ice edge provides room for ice divergence; this means that large polynyas occur in the Bering, Okhotsk, and Barents Seas. In the Barents Sea, prominent polynyas occur around Novaya Zemlya, Franz Josef Land, and at occasional coastal sites in the Barents and adjacent Kara Seas. In the Laptev Sea, polynyas occur in early winter along Severnaya Zemlya. There is also a long flaw lead, sometimes referred to as a polynya, which occurs between the shorefast and the pack ice north of the Laptev River delta. In the vicinity of the Bering Strait, polynyas occur in the Chukchi, Beaufort, and Bering Seas. In the Chukchi Sea, an important polynya occurs along the Alaskan coast; in the Beaufort Sea, polynyas form in early winter along the Alaskan coast and in Amundsen Gulf. In the Bering Sea, large polynyas occur along the Alaskan coast and south of the islands, where the most investigated of these occurs south of St. Lawrence Island. Prominent polynyas also occur along the Siberian coast south of the Bering Strait and in Anadyr Gulf. The Canadian islands are also sites of several polynyas, the largest and most studied being the North Water polynya in Baffin Bay. In north-east Greenland, the North-east Water (NEW) polynya is large in summer, and has also been the subject of a recent field study. Finally, the Okhotsk Sea with the adjacent Tatarskiy Strait is the site of several coastal polynyas and the openocean Kashevarov polynya. The Kashevarov polynya occurs over the 200 m deep Kashevarov Bank, where the turbulence associated with a strong tidal resonance generates a heat flux to the surface that creates a region of reduced ice cover with a characteristic diameter of about 100 km. The Southern Hemisphere is characterized by many coastal polynyas and by two or three openocean polynyas. The open-ocean polynyas include the Weddell Sea polynya, which only occurred in the 1970s, and the smaller Maud Rise and Cosmonaut Sea polynyas. The coastal polynyas occur at different locations along the Ronne, Amery, and Ross Ice Shelves, where the Ross polynya has recently been studied, and at many locations along the coast. The coastal polynyas form in the lee of headlands, islands, ice shelves, and grounded icebergs and downwind of ice tongues such as the Mertz, Dibble, and Dalton tongues, which extend as much as


Gre enl and

POLYNYAS

North Atlantic

Baffin North Bay Water

NEW

Lancaster Sound

Svalvard

Barents Sea

Franz Josef Land

Kara Sea

Novaya Zemlya

Severnaya Zemlya

Laptev Sea

Anadyr Gulf Shelikhov Bay

Laptev River

Tatarskiy Strait Japan Sea

Weddell Sea Larson Ice Shelf

K Okhotsk Sea

North Pacific

C Syowa

Halley Bay Filchner Ronne Ice Shelf

Bellingshausen Sea ANTARCTICA Ruppert Ross Coast Ice Shelf Ross Sea

(B)

SLIP Bering Sea

M W

Ice Shelf

Amundsen Sea

Alaska Bering Strait

Chukchi Sea

Flaw Lead

(A)

Amundsen Gulf

Beaufort Sea

Amery Ice Shelf

Enderby Land

West Ice Shelf

Shackleton Ice Shelf Cape Poinsette Dalton Ice Tongue d an Porpoise Bay Wilkes L Dibble Ice Tongue Mertz Ice Tongue Terra Nova Bay

Antarctic Ocean

Figure 1 Geographic distribution of polynyas. (A) The Arctic, where SLIP is the St. Lawrence Island Polynya, NEW is the North-east Water, and K is the Kashevarov Bank polynya. (B) The Antarctic, where W is the Weddell polynya, M is the Maud Rise polynya and C is the Cosmonaut Sea polynya. On both figures the dashed line indicates the position of the maximum ice edge. Polar stereographic map projection courtesy of the National Snow and Ice Data Center (NSIDC).

100 km off the coast. These polynyas are created by a combination of the dominant easterly winds and the very cold powerful katabatic winds that sweep down the glacial drainage basins and across the pack ice for

541

a few kilometers before slowing. The response to these winds is a series of large polynyas; for example, the overwinter average area of the Mertz Ice Tongue polynya is about 20 000 km2. Given the prevalence of coastal polynyas compared with mid-ocean polynyas, most of the water mass modification takes place in the former. Although these polynyas occupy only a small fraction of the areal winter pack ice extent, because the polar pack ice is a good insulator, very large atmospheric heat losses occur from both polynya types. Given these heat losses, the coastal polynyas are regions where large amounts of ice are generated, the ocean is cooled and salt is added to the underlying waters, while the open-ocean polynyas cool the upwelled water, and in the Southern Ocean are suspected to contribute to modification of the Antarctic Bottom Water. As winter progresses into spring, the polynya regions remain important. Because the predominant winds sweep the polynyas free of ice, their ice cover at the end of winter consists of either open water or thin ice. This means that as spring approaches, when the air temperature rises above freezing and the incident solar radiation increases, the polynya ice is either swept away without replacement, or melts away faster than the surrounding pack ice. In the Okhotsk and Bering Seas, for example, the onset of open water in spring occurs both from melting at the offshore ice edge, and from the disappearance of ice at the coastal polynya sites. Because the other Arctic and Antarctic coastal polynyas behave similarly, the coastal polynyas become seasonally open approximately one month earlier than the interior pack. The open ocean polynyas, such as the Kashevarov, Maud, and Cosmonaut, also serve as regions for initiation of the spring melt. As is shown below, the early spring melt of the coastal polynyas has biological consequences.

Physical Processes within the Two Polynya Types Coastal Polynyas

Because of their relative accessibility and more frequent occurrence, we know much more about coastal than about open-ocean polynyas. Figure 2, a schematic drawing of a coastal polynya, shows that as the winds advect the pack ice away from the coast, open water is exposed to the cold winds. This generates a wind-wave field on the open water surface, where the wave amplitudes and wavelengths increase away from the coast. As the polynya width increases, and if the wind speed is greater than about 5–10 m s 1, the interaction of the waves with the wind stress creates


542

POLYNYAS

Side view Frazil ice Ice export pile-up

Wind

Pack ice Frazil ice Ocean

Bottom

Top view

A

Open water Pack ice

A′

Figure 2 A schematic drawing of a coastal polynya and top and vertical views.

Figure 3 A RADARSAT image of the St. Lawrence Island polynya on 9 January 1999. The length of the island is approximately 150 km. On the image, (a) is the island, (b) is the fast ice south of the island, (c) is open water, (d) is the Langmuir plumes, (e) is the pack ice, (f) is the piled-up grease ice, and (g) is the thick first-year ice north of the island. The wind is blowing from top right to bottom left. Image courtesy of the Alaska SAR Facility, with processing courtesy Harry Stern and copyright the Canadian Space Agency, 1999.

Langmuir circulation within the water column. This circulation consists of rotating vortices with the rotor axes approximately parallel to the surface winds, where adjacent rotors turn in opposite directions and the rotor diameter is approximately equal to either the bottom depth in well-mixed waters or to the halocline depth. Because of the wave and Langmuir mixing, the initial ice formation occurs as follows. If the sea water temperature is above freezing, the combination of the surface heat loss with the mixing cools the entire column to the freezing point, and sometimes even causes a slight supercooling. This means that once freezing begins, ice formation occurs throughout the water column as small millimeter-scale crystals, called frazil crystals, which float slowly to the surface. As the crystals form, they reject salt to the underlying water column, leading to an oceanic brine flux. Once these crystals reach the surface, the circulation herds them into slurries taking the form of long bands or plumes of floating ice crystals located at the Langmuir convergence zones. The slurries have thicknesses of order 10 cm, are highly viscous, and damp out the incident short waves. This damping gives the slurries a greasy appearance, so that following old whaling terminology, they are sometimes called grease ice. As the plumes grow downwind, they become wider and increase in thickness. As their

thicknesses increase, their surface begins to freeze. The longer ocean swell propagating through the ice, breaks the surface ice into floes with diameters of 0.3 to 0.5 m, called pancake ice. Because of wave-induced collisions, the swell also causes the growth of raised rims around the pancakes, which increase both the wind-drag on the ice and the radar reflectivity. As these ice growth processes proceed, the ice is advected downwind by the wind stress, where it piles up against the edge of the solid pack ice. As time goes on, the width of this region of piled-up ice grows slowly upwind. As an example, Figure 3 shows a 100 m resolution RADARSAT image of the St. Lawrence Island polynya. The figure shows the region of open water and Langmuir plumes surrounded by pack ice south of the island, where the plumes are approximately parallel to the wind, and the polynya area is about 5000 km2. Within the plumes, the figure also shows the downwind increase in brightness associated with the growth of pancake ice. If the wind speeds are slow enough that the Langmuir circulation either does not occur or is not strong enough to circulate the ice crystals the polynya behavior is less well understood. Observations suggest that the downwind transport of ice continues to occur, but with either frazil or thin ice forming immediately adjacent to the coast. Given the

A

Front view, vertical section A′


POLYNYAS

offshore transport of ice in both the Langmuir and non-Langmuir cases, the question arises of what determines the winter polynya size. The crosswind polynya scale is set by the coastline configuration over which the ice divergence occurs. The downwind scale is set by a balance between the production of new ice within the polynya, its export downwind to the pack ice edge, and the subsequent upwind growth of the piled up new ice. All else remaining constant, a greater polynya area leads to more ice production and a faster upwind growth of the new ice. An equilibrium size occurs when the retreat of the pack ice edge equals the advance of the piled-up new ice. This balance between production and advection sets the polynya offshore length scale, typically 10–100 km. When the winds stop, the export stops, and the frazil ice freezes into a solid ice cover. For the same wind speed, very cold temperatures yield smaller polynyas because the ice production is so much greater; relatively warm temperatures correspond to large polynyas with less ice production. Open-ocean Polynyas

The open-ocean polynyas generally occur away from the coast, and are driven by the upwelling of warm sea water (Figure 4). In the Northern Hemisphere, the Kashevarov polynya is driven by a tidal resonance, and the North Water and NEW are maintained by a combination of wind-induced ice advection and oceanic upwelling. In the Southern Hemisphere, there have been three open-ocean polynyas. The most famous, intriguing, and mysterious of these occurred in the Weddell Sea during 1973–76, and was only observed by remote sensing. The Weddell polynya had an open water and thin ice area of about 2–3 105 km2, which is about the area of Oregon, USA. Because this polynya first occurred over Maud Rise, its subsequent evolution may have been caused by a large eddy that separated from the rise and migrated westward across the Weddell Sea. For the other open-ocean polynyas, both the Cosmonaut and Atmosphere Open water Pack ice Ocean

Upwelled heat

543

Maud polynyas have maximum areas of about 105 km2. These occur in oceanic regions with large reservoirs of relatively warm water just beneath a weak pycnocline, where upwelling brings the warm water to the surface. These polynyas are self-maintaining, in that as heat is lost to the atmosphere, the surface water becomes denser and sinks, generating a turbulent convection which brings warm deep water to the surface. The convection ceases when the atmosphere warms in spring, or if sufficient fresh water, either produced locally by melting, or advected into the region, places a low-salinity cap on the convection.

Remote Sensing Observations Even though polynyas were probably first observed by Native American whalers and hunters, our detailed knowledge of their location and variability comes from satellite observations. In the 1970s and early 1980s, the forerunners of the AVHRR (Advanced Very High Resolution Radiometer) instrument provided 1 km resolution visible and thermal imagery of polynyas, but only under cloud-free conditions. Passive microwave instruments such as the SMMR (Scanning Multichannel Microwave Radiometer), which operated between 1978 and 1987, and the SSM/I (Special Sensor Microwave/ Imager, operating between 1987 and the present, made it possible to obtain cloud-independent lowresolution imagery of the entire polar pack at intervals of every other day for the SMMR and daily intervals for the SSM/I. These observations led to the discovery of all of the mid-ocean polynyas, and many of the coastal polynyas, and provided time series of their area versus time. The major problem with the SMMR and SMM/I is their low spatial resolution; the 37 GHz SSM/I channel has a 25 km resolution, and the more water vapor-sensitive channel at 85 GHz has a 12.5 km resolution. The planned AMSR (Advanced Microwave Scanning Radiometer), which is scheduled for launch in 2000, has twice the resolution of the SSM/I and should greatly improve our polynya studies. The high-resolution active microwave satellite instruments that are just beginning to be used in polynya research include the 100–500 km swath width synthetic aperture radars (SAR) with their approximately 3-day repeat cycle, and resolutions of 12.5–100 m (Figure 3).

Isopycnal

Physical Importance Turbulent convection

Figure 4 A schematic drawing of an open-ocean polynya, showing mean circulation and mixing.

The physical importance of the coastal polynyas is that, because the new ice is being constantly swept away, the polynyas serve as large heat sources to the


544

POLYNYAS

atmosphere, and as powerful ice and brine factories. For example, while a typical Bering Sea ice thickness is about 0.5 m, the seasonal ice growth in the St. Lawrence polynya is about 5 m, or an order of magnitude greater. Combined with the persistent north-easterly winds, this means that the Bering ice cover can be approximated as a conveyor belt, where the ice forms in the coastal polynyas, then is advected to the ice edge where it melts. This simple conveyorbelt model may also apply to the ice covers of the Weddell and Okhotsk Seas. The brine generated by the polynyas also contributes to the oceanic water masses. The large shallow continental shelves where most of the Arctic coastal polynyas occur provide a dynamic constraint on the brine-generated dense water that permits its density and volume to increase. Because of this constraint, when the dense water eventually drains off the shelves into the deeper basins, it is dense enough to contribute to the intermediate and deep water masses. In this way, the polynyas in the Chukchi, Bering, Beaufort, and Barents Sea polynyas contribute to the cold halocline layer of the Arctic Ocean. (The Bering Sea dense water is transported through the Bering Strait into the Arctic by the pressure difference between the Bering Sea and the Arctic Ocean.) In the Okhotsk Sea, the polynya water contributes to the Okhotsk Intermediate Water and to the North Pacific Intermediate Water. The dynamics and fate of this dense water are topics of current scientific investigation. In the Antarctic, estimates of the polynya ice growth rates are 0.1–0.2 m d 1, or about 10 m per season. Although the Antarctic shelves are not as broad as the Arctic shelves, the Antarctic coastal polynyas also generate dense shelf water. The drainage of this water contributes to the Antarctic bottom water in the western Weddell Sea, the Ross Sea, and along the Wilkes Land Coast. In the Weddell Sea, the dense water flows beneath the Filchner-Ronne Ice Shelf, causing melting of the glacial ice, and creating a lower-salinity form of the bottom water. The Antarctic open-ocean polynyas cool the warm upwelled deep ocean water, which leads to modification of the intermediate-depth water into cold bottom water. The Arctic polynyas also contribute to sediment transport. Because freezing occurs at all depths during the initiation of ice formation in coastal polynyas, nucleating ice crystals adhere to rocks and sediments on the bottom, forming what is called ‘anchor ice.’ Although this has never been directly observed in polynyas because of the hazardous conditions, observations by Alaskan divers following severe autumnal storms have found sediment-laden frazil ice on the surface, and anchor ice on the

bottom. Downwind of these polynya regions, there are also many observations of pack ice containing layers with large sediment concentrations. These observations suggest that as the amount of anchor ice increases, the buoyancy of the sediment/ice mixture lifts the material to the surface, where it accumulates downwind. At river deltas such as the Laptev River, this means that sediments carried by the river into the delta can be incorporated into the frazil ice, for later export across the Arctic. Laboratory studies also suggest that the Langmuir circulation may also directly mix bottom sediments into the water column, where these sediments are then incorporated into the frazil ice and carried to the surface. By a combination of these routes, the coastal polynyas probably serve as a source of the observed sediments in the polar ice. This mechanism also provides a route for sediments laden with contaminants or radionuclides carried into the polynya regions by the rivers, to be incorporated into the pack ice, and then be transported across the Arctic.

Biological Importance In the winter Canadian Arctic, because the marine mammals living under the ice need breathing holes, these mammals tend to concentrate in the polynya regions. For example, the North Water contains large concentrations of white whales, narwhals, walruses, and seals, with polar bears foraging along the coast. Also, the major winter bird colonies in the Canadian islands are located adjacent to polynyas, especially the North Water, and archeological evidence shows that the coastal region adjacent to the NEW was the site of human settlements. All of this suggests that polynyas are vital for the overwinter survival of arctic species. In spring and summer, the polynya regions continue to have large concentrations of marine mammals. This is because, as the polynya regions become ice-free earlier than the rest of the pack, they annually absorb more solar radiation than an adjacent region covered with thick pack ice. As a result, the polynya regions in summer have a much greater primary productivity than regions with heavy winter pack ice, so that these regions are important feeding areas for whales. Also, in the US and Canadian Arctic, most of the whales migrate into the region through passages generated by the spring melt of polynyas. For example, the opening of the polynya region along the Alaskan Chukchi coast provides a whale migration route to the Beaufort Sea, and in the Canadian Arctic the early opening of the polynya regions provides migration routes for narwhals in Lancaster Strait.


POLYNYAS

Conclusions Polynyas are persistent openings in the ice cover that in winter ventilate the warm ocean directly to the cold atmosphere. The major physical importance of the coastal polynyas is due to their large production of ice and brine, where the resultant dense water contributes to various Arctic, Antarctic, and North Pacific water masses. In the Southern Hemisphere, the winter convection within the mid-ocean polynyas also leads to cooling of the upwelled ocean water and its modification into Antarctic bottom water. Also, although this has not been elaborated on because of space considerations, the polynyas may enhance the exchange of gases such as carbon dioxide between the atmosphere and ocean. The biological importance of the polynyas is that at least in the US and Canadian Arctic, they serve as an important winter and summer habitat for marine birds and mammals. The polynyas also play a role in spring and summer as regions of early open water formation, and as sites of significant increases in primary productivity associated with the early absorption of incident solar radiation in the water column.

Acknowledgment This material is based upon work supported by the National Science Foundation under Grant No. 9811097.

545

Passive-Microwave Measurements of Sea Ice. Satellite Remote Sensing SAR. Sea Ice. Sea Ice: Overview. Sub Ice-Shelf Circulation and Processes. Surface Gravity and Capillary Waves. Weddell Sea Circulation.

Further Reading Comiso JC and Gordon AL (1998) Interannual variability in summer sea ice minimum, coastal polynyas and bottom water formation in the Weddell Sea. In: Jeffries MO (ed.) Antarctic Sea Ice, Physical Processes, Interactions and Variability. Antarctic Research Series 74, pp. 293--315. Washington, DC: American Geophysical Union. Gordon AL and Comiso JC (1988) Polynyas in the Southern Ocean. Scientific American 256(6): 90--97. Martin S, Steffen K, Comiso JC, et al. (1992) Microwave remote sensing of polynyas. In: Carsey F (ed.) Microwave Remote Sensing of Sea Ice. Geophysical Monograph 68, pp. 303--312. Washington, DC: American Geophysical Union. Overland JE, Curtin TB, and Smith WO (eds.) (1995) Special Section: Leads and Polynyas. Journal of Geophysical Research 100: 4267--4843. Smith SD, Muench RD, and Pease CH (1990) Polynyas and leads: an overview of physical processes and environment. Journal of Geophysical Research 95(C6): 9461--9479. Stirling I (1997) The importance of polynyas, ice edges, and leads to marine mammals and birds. Journal of Marine Systems 10: 9--21.

See also Arctic Ocean Circulation. Current Systems in the Southern Ocean. Langmuir Circulation and Instability. Marine Mammal Migrations and Movement Patterns. Marine Mammal Overview. Marine Mammals, History of Exploitation. Okhotsk Sea Circulation. Open Ocean Convection. Satellite


POPULATION DYNAMICS MODELS Francois Carlotti, C.N.R.S./Universite´ Bordeaux 1, Arachon, France Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2247–2256, & 2001, Elsevier Ltd.

Introduction The general purpose of population models of plankton species is to describe and eventually to predict the changes in abundance, distribution, and production of targeted populations under forcing of the abiotic environment, food conditions, and predation. Computer-based approaches in plankton ecology were introduced during the 1970s with the application of population models to investigate large-scale population phenomena by the use of mathematical models. Today, virtually every major scientific research project of population ecology has a modeling component. Population models are built for three main objectives: (1) to estimate the survival of individuals and the persistence of populations in their physical and biological environments, and to look at the factors and processes that regulate their variability; (2) to estimate the flow of energy and matter through a given population; and (3) to study different aspects of behavioral ecology. The study of internal properties of a population, like the various effects of individual variability, and the study of interactions between populations and successions of population are also topics related to population models. The field of biological modeling has diversified and, at present, complex mathematical approaches such as neural networks, genetic algorithms, and dynamical optimization are coming into use, along with the application of supercomputers. However, the use of models in marine research should always be accompanied by extensive field data and laboratory experiments, for initialization, verification or falsification, or continuous updating.

Approach for Modelling Plankton Populations Population Structure and Units

A population is defined as a group of living organisms all of one species restricted to a given area and with limited exchanges of individuals from other populations. The first step in building a population

546

model is to identify state variables (components of the population) and to describe the interactions between these state variables and external variables of the system and among the components themselves. The components of a population can be (1) the entire population (one component); (2) groups of individuals identified by a certain states: developmental stages, weight or size classes, age classes (fixed numbers of components); or (3) all individuals (varying numbers of components). The usual unit in population dynamics models is the number of individuals per volume of water, but the population biomass can also be used (in g biomass or carbon (C) or nitrogen (N)). When all individuals or groups of individuals are represented, the individual weight can be considered as a state variable. The forcing variables influencing the population dynamics are biological factors, mainly nutrients and predators, and physical factors, mainly temperature, light, advection, and diffusion. Individual and Demographic Processes

Population models usually work with four major processes: individual growth, development, reproduction, and mortality. Growth is computed by the rate of individual weight change. Development is represented by the change of states (phases in phytoplankton and protozoan cell division, developmental stages in zooplankton) through which each individual progresses to reach maturity. Reproduction is represented by the production of new individuals. Mortality induces loss of individuals, and can be divided in two components: natural physiological mortality (due, for instance, to starvation) and mortality due to predation. Combination of the four processes permits one to stimulate (1) increase in terms of number of individuals in the population, (2) the body growth of these individuals, and (3) by combination of the two previously simulated values, the increase in total biomass of the population (which is usually termed ‘population growth’). Plankton Characteristics

Essential information to be built into models of plankton organisms are (1) the individual life duration (a few hours for bacteria; one to a few days for phytoplankton and unicellular animals; several weeks to years for zooplankton and ichthyoplankton organisms); (2) the range of change in size or weight between the beginning and the end of a life cycle; and


POPULATION DYNAMICS MODELS

Plankton Population Models

The most modeled component in marine planktonic ecosystems is phytoplankton production. Most of the phytoplankton models simulate the growth of phytoplankton as a whole, using only the process of photosynthesis. Few models deal with phytoplankton population growth dynamics at the species level. Existing models of other unicellular plankton organisms (bacterioplankton, species of microzooplankton) usually treat them as a single unit, except for a few models simulating phytoplankton and microbial cell cycles. In contrast, mesozooplanktonic organisms, including the planktonic larval stages of benthic species (meroplankton), and fish that have complex life cycles are extensively modeled at the population level.

Dynamics of Single Species Population Models Described by the Total Density

When a population is observed at timescales much larger than the individual life span, and on a large number of generations, models with one variable (the total number of individuals or the total biomass in that population) are the simplest. These models postulate that the rate of change of the population number, N, is proportional to N (eqn [1], where r is the difference between birth and death rates). dN ¼ rN dt

½1

The logistic equation ([2]) represents limitation due to the resources or space (see Figure 1). dN N ¼ rN 1 dt K

½2

where K is the carrying capacity. Population growth of bacteria, phytoplankton, or microzooplankton can be simulated adequately by the logistic equation. With addition of a time delay term into the logistic equation, oscillations of the population can be represented.

K Observations

Number of individuals

(3) the number of individuals produced by a mother individual (from two individuals up to thousands of individuals). When developmental stages in the life cycle are identified, the stage durations are needed. The observation time step has to be defined to adequately follow the timescale of the chosen variables, and thus should be smaller than the duration of the shortest phases.

547

Fitting of the logistic equation

N0 Time step between observations

Time Figure 1 The growth of a plankton population with density regulation and its fitting by the logistic equation.

Population Models of Organisms with Description of the Life Cycle

If observations of a population are made with a time step shorter than the life cycle duration (Figure 2), the population development pattern is a succession of periods with decreasing abundance of individuals due to mortality, and increasing abundance due to recruitment of new individuals in periods of reproduction (cell division or egg production). Recruitment is defined as the input flux of individuals in a given state (stage, size class, etc.). To represent such patterns, it is necessary to identify different phases in the life cycle, based on age, size, developmental stages, and so on. Two types of models can be developed:

• •

Structured population models, which consider the flux of individuals through different classes Individual-based models, which simulate birth, growth and development through stages and death of each individual

The major distinction between physiologically structured population models and individual-based models in a stricter sense is that individual-based models track the fate of all individuals separately over time, while physiologically structured population models follow the density of individuals of a specific type (age or size classes, stages). These models are particularly used for representing the complex life cycles of zooplankton and ichthyoplankton, but have also been useful for studying the population growth of bacteria, phytoplankton, and microzooplankton, particularly division synchrony in controlled conditions (chemostat).


548


Suppose there are m age classes numbered 1, 2, y, m, each covering an interval t. If Nj;t denotes the number of individuals in age class j at time t and Gj denotes the fraction of the population in this age class that survive to enter age class j þ 1, then eqn [3] applies.

Production of new individuals

Loss by mortality Number of individuals

Njþ1; tþ1 ¼ Gj Nj;t

½3

Individuals of the first age class are produced by mature individuals from older age classes and eqn [4] applies, where Fj is the number of age class 1 individuals produced per age class i individual during the time step t. Life cycle

N1;tþ1 ¼

m X

Fj Nj;t

½4

j¼1

Time Figure 2 Total population abundance in controlled conditions. The population is initiated with newborn individuals, and decreases, first owing to mortality, until the maturation period, when a new generation is produced. At the end of the recruitment period of new individuals of second generation, the total abundance decreases, again owing to mortality. Owing to individual variability in development, loss of synchronism in population induces a broader recruitment period for the third generation. In nonlimiting food conditions, recruitment of new individuals is continued after few generations. The three generations correspond to the exponential phase presented in Figure 1 (i.e., without density regulation).

Structured Population Models

A population can be structured with respect to age (age-structured population models), stage (stagestructured population models), or size or weight (size or weight-structured population models). Two types of equations systems are usually used: matrix models, which are discrete-time difference equation models, and continuous-time structured population models. Matrix models constitute a class of population models that incorporate some degree of individual variability. Matrix models are powerful tools for analyzing, for example, the impact of life history characteristics on population dynamics, the influence of current population state on its growth potential, and the sensitivity of the population dynamics to quantitative changes in vital rates. Matrix models are convenient for cases where there are discrete pulses of reproduction, but not for populations with continuous reproduction. They are not suitable for studying the dynamics of populations that live in fluctuating environments. The Leslie matrix is the simplest type of agestructured dynamic considering discrete classes.

The system of eqns [3] and [4] can be written in matrix form (eqn [5]). 2

N1

6N 6 2 6 6 N3 6 6 . 6 . 4 .

3

2

0 6G 7 6 1 7 6 7 6 7 7ðt þ 1Þ ¼ 6 0 6. 7 6. 7 4. 5

Nm

0

F2 0 G2

32

F3

?

Fm

0 0

? ?

N1 76 N 76 2 76 76 N3 76 76 . 76 . 54 . Nm

& &

?

0 0 .. .

0

Gm1

0

3 7 7 7 7 7ðtÞ 7 7 5

½5 The Leslie matrix can easily be modified to deal with size classes, weight classes, and developmental stages as the key individual characteristics of the population. Organisms grow through a given stage or size/ weight class for a given duration. There are several variations of matrix models, differing mainly in the expression of vital rates, which can vary with time depending on external factors (e.g., temperature, food concentration, competitors, predators) or internal (e.g., densitydependent) factors. The earlier type of continuous-time structured model is usually referred to as the McKendrick–von Foerster equation, and uses the age distribution on a continuous-time basis in partial differential equations. This type of model has been developed to the extent that it can be used to describe population dynamics in fluctuating environments. In addition, it also applies to situations in which more than one physiological trait of the individuals (e.g., age, size, weight, and energy reserves) have strong influences on individual reproduction and mortality. The movement of individuals through the different structural classes is followed over time. Age and weight are continuous variables, whereas stage is a discrete variable.



The general equation is eqn [6], where n is abundance of individuals of age a and mass m at time t. @nða; w; tÞ @nða; w; tÞ @gða; w; tÞnða; w; tÞ þ þ @t @a @w ¼ mða; w; tÞnða; w; tÞ

½6

where mða; w; tÞ is the death rate of the population of age a, weight w at time t. The von Foerster equation describes population processes in terms of continuous age and time (agestructured models) according to eqn [7]. @nða; tÞ @nða; tÞ þ ¼ mða; tÞnða; tÞ @t @a

½7

549

Using upwind difference discretization to solve the equations, and recasting the representation in terms of the number of individuals in the ith weight class, Ni ðtÞEni ðtÞ Dwi , the dynamic equation becomes [13], where mi ðtÞ replaces mðwi ; tÞ dNi gi1 gi ¼ Ni1 Nimi Ni dt Dwi1 Dwi

½13

This describes the dynamics of all weight classes except the first (i ¼ 2) and last (i ¼ Q). If RðtÞ represents the total rate of recruitment of newborns to the population, and all newborns are recruited with the same weight w1 , then the dynamic of the weight class covering the range Dw1 is described by eqn [14].

The equation has both an initial age structure j at t ¼ 0 (eqn [8]) and a boundary condition of egg production at a ¼ 0 (eqn [9]).

dN1 g1 ¼R N1 m1 N1 dt Dw1

nða; 0Þ ¼ j0 ðaÞ ðN Fða; SR Þnða; tÞda nð0; tÞ ¼

If we assume that individuals in only the Qth weight class are adult, and that adult individuals expend all assimilated energy on reproduction rather than growth, the population dynamics of the adult population is given by eqn [15].

½8 ½9

0

F is a fecundity function that depends on age (a) and the sex ratio of the population SR . These kinds of equations are mathematically and computationally difficult to analyze, especially if the environment is not constant. The same type of equation as [7] can be used where the age is replaced by the weight (weightstructured models (eqn [10]). @nðw; tÞ @gðw; T; PÞnðw; tÞ þ ¼ mðw; tÞnðw; tÞ ½10 @t @w The weight of the individual w and the growth g are influenced by the temperature T, the food P, and by the weight itself through allometric metabolic relationships. The equation has both an initial age structure j at t ¼ 0 (eqn [11]) and a boundary condition of egg production at w ¼ w0 (eqn [12]). nðw; 0Þ ¼ j0 ðwÞ

N ð0; tÞ ¼

ðN

½11

Fðw; SR Þnðw; tÞdw

½12

0

F is the fecundity function, which depends on weight (w) and the sex ratio of the population SR . The numerical realization of this equation requires a representation of the continuous distribution nðw; tÞ by a set of discrete values ni ðtÞ that are spaced along the weight axis at intervals Dwi ¼ wiþ1 wi .

dNQ gQ1 ¼ NQ1 mQ NQ dt DwQ1

½14

½15

The rate of recruitment of newborns to the population is given by [16] where bðtÞ represents the per capita fecundity of an average adult at time t. RðtÞ ¼ bðtÞNQ ðtÞ

½16

The weight intervals Dwi increase with class number i as an allometric function. The growth rate gðw; tÞ can be calculated by a physiological model. Stage-structured Population Models

Plankton populations often have continuous recruitment and are followed in the field by observing stage abundances over time. A large number of zooplankton population models deal with population structures in term of developmental stage, using ordinary differential equations (ODEs). A single ODE can be used to model each development stage or group of stages: for instance, a copepod population can be subdivided into four groups: eggs, nauplii, copepodites, and adults. The equation system is eqn [17]–[20], where R is recruitment, a is the transfer rate to next stage, and m is the mortality rate. Eggs


dN1 ¼ R a1 N1 m1 N1 dt

½17

550


R

1

Eggs 1

3

2

Nauplii

Copepodites 2

Adults

3

4

Figure 3 Schematic representation of the population dynamics mathematically represented by eqn [17]–[20]. ai ¼ transfer rate of stage i to stage i þ 1; mi ¼ mortality rate in stage i; R ¼ recruitment: number of eggs produced by females per day.

Nauplii

dN2 ¼ a1 N1 a2 N2 m2 N2 dt

½18

Copepodids

dN3 ¼ a2 N2 a3 N3 m3 N3 dt

½19

adults

dN4 ¼ a3 N3 m4 N4 dt

½20

The system of ODEs is solved by Euler or Runge– Kutta numerical integration methods, usually with a short time step (approximately 1 hour). In the model presented in Figure 3, the transfer rate of animals from stage to stage and the mortality at each stage are expressed as simple linear functions, which induce a rapidly stable stage distribution. To represent the delay of growth within a stage, more refined models consider age-classes within each stage, or systems of delay differential equations. They have a high degree of similarity with observed cohort development in mesocosms or closed areas (Figure 4). Individual-based Models of a Population

Individual-based models (IBMs) describe population dynamics by simulating the birth, development, and eventual death of a large number of individuals in the population. IBMs have been developed for phytoplankton, zooplankton, meroplanktonic larvae, and early life history of fish populations. Object-oriented programming (OOP) and cellular automata techniques have been applied to IBMs. As powerful computers become more accessible, numerous IBMs of plankton populations have been developed, mainly to couple them with 1D-mixed layer models (phyto- and zooplankton) and circulation models (zooplankton). IBMs treat populations as collections of individuals, with explicit rules governing individual biology and interactions with the environment. Each

biological component can change as a function of the others. Each individual is represented by a set of variables that store its i-state (age, size, weight, nutrient or reserve pool, etc.). These variables may be grouped together in some data structure that represents a single individual, or they may be collected into arrays (an array of all the ages of the individuals, an array of all the sizes of the individuals, etc.), in which case an individual is an index number in the set of arrays. The i-state of an individual changes as a function of the current i-state, the interactions with other individuals, and the state of the local environment. The local environment can include prey and predator organisms that do not warrant explicit representation as individuals in the model. Population-level phenomena (e.g., temporal or spatial dynamics) or vital rates can then be inferred directly from the contributions of individuals in the ensemble. The model starts with an initial population and the basic environment, then monitors the changes of each individual. At any time t, the i-state of individual j changes as eqn [21]. Xi;j ðtÞ ¼ Xi;j ðt dtÞ þ f X1;j ðt dtÞ; y; Xi;j ðt dtÞ ½21 Xi;j ðtÞ is the value of the i-state of individual j, and f is the process modifying Xi;j , as a function of the values of different i-states of the organism and external parameters such as the temperature. When the fate of all individuals during the time-step dt has been calculated, the changes to the environment under the effects of individuals can be updated. Any stochastic process can be added to eqn [21]. This type of model can add a lot of detail in the representation of physiological functions. Individual growth can be calculated as assimilation less metabolic loss, and the interindividual variation in physiology can be represented by adding stochastic processes or parameters describing the characteristics of each individual (growth and development parameters, mortality coefficient, and parameters connected with reproduction). The end results are unique life histories, which when considered as a whole give rise to growth/size distributions that provide a measure of the state of the population. Calibration of Parameters

Parametrization of a model can range from very simplistic to extremely complex depending upon the amount of information known about the population under consideration. Bioenergetic processes (ingestion, egestion, excretion, respiration, and egg



30 Temperature (°C)

Percent of initial population density

500

551

400

300

25

20

15 0

2

4

6

200

8 10 Days

12

14

16

100

0 4

2

0

6

8

(A)

200

200

14

16

200 N3

100

100

100

0 100

0 100

0 40

N4

Percent of initial population density

12

N2

N1

N5

N6

50

50

20

0

0

0

40

20

20

C1

C2

C3

20

10

10

0 20

0 20

0 40 Adults

C5

C4

20

10

10

0

0

0 0 (B)

10

Days

4

8

12

16

0

4

8

12

16

0

4

8

12

16

Days

Figure 4 Simulation of the cohort development of the copepod Euterpina acutifrons in mesocosms with a structured stage and agewithin-stage model. (A) Total population during development, with variable temperature and constant food supply (points ¼ experimental data; line ¼ simulation). The initial density decreases owing to mortality and then increases to newborn individuals (as the first part in Figure 2). (B) Naupliar stages (N1 to N6), copepodite stages (C1 to C5), and adults during development (points ¼ experimental data; line ¼ simulation). The simulation start with similar N1 of same age. Reproduced from Carlotti F and Sciandra A, 1989. Population dynamics model of Euterpina acutifrons (Copepoda: Harpacticoida) coupling individual growth and larval development. Mar. Ecol. Prog. Ser., 56, 3, 225–242.


552


production) are usually modeled from experimental results, whereas biometrics (size, weight,y) and demographic (development rate, mortality rate, y) parameters are estimated by combining data from life tables collected in the field or from laboratory studies. To solve for the unknown parameters, new techniques have been developed such as inverse methods and data assimilation by fitting simulations to data.

Spatial Distribution of Single Plankton Populations An important development in plankton population modeling is to make full use of the increased power of computers to simulate the dynamics of plankton (communities or populations) in site-specific situations by coupling biological and transport models, giving high degrees of realism for interpreting plankton population growth, transport, spatial distribution, dispersion, and patchiness. Structured population models and individual-based models allow detailed simulations of zooplankton populations in different environmental conditions. Physical–biological models of various levels of sophistication have been developed for different regions of the ocean. Spatial Plankton Dynamics with Advection– Diffusion–Reaction Equations

Equation [22] is the general physical–biological model equation used to describe the interaction between physical mixing and biology. @C þ r ðva CÞ r ðKrCÞ ¼ ‘biological teams’ @t ½22 Cðx; y; z; tÞ is the concentration of the biological variable, which is a functional group (phytoplankton, microzooplankton, or zooplankton), a species or a developmental stage, or a size class (in which case the number of equations would equal the number of stages or size classes) at position x; y; z at time t. The concentration can be expressed as numbers of organisms or biomass of organisms per unit volume. va ðua ; va ; wa Þ represents the advective fluid velocities in x; y; z directions. Kx ; Ky ; Kz are diffusivities in x; y; z directions. D ¼ ð@=@x; @=@y; @=@zÞ is the Laplacian operator. On the left-hand side of eqn [22], the first term is the local change of C, the second term is advection caused by water currents, and the third term is the diffusion or redistribution term. The right-hand side

of eqn [22] has the biological terms that represent the sources and sinks of the biological variable at position x; y; z as a function of time. The biological terms may or may not include a velocity component (swimming of organisms, migrations, sinking, y), and the complexity of the biological representation can vary from the dispersion of one (the concentration of a cohort) to detailed population dynamics. Physical–biological models of various levels of sophistication have been developed recently for different regions of the ocean. Biological models can be configured as compartmental ecosystem models in an upper-ocean mixed layer, where phyto-, microzoo-, and mesozooplankton are represented by one variable. In extended cases, the model takes into account several size classes of phyto-, microzoo-, and zooplankton. Such types of ecosystem model have been coupled to one-dimensional physical, and embedded into twodimensional and three-dimensional circulation. Studies of plankton population distribution in regions where plankton may be aggregated (e.g., upwelling and downwelling regions, Langmuir circulations, eddies) can be undertaken with populations described by equations of the McKendrick– von Foester type coupled with 2D or 3D hydrodynamical models. In 1982, Wroblewski presented a clear example with a stage-structured population model of Calanus marshallae, a copepod species, embedded in a circulation system simulating the upwelling off the Oregon coast. Simulations of the dynamics focused on the interaction between diel vertical migration and offshore surface transport. The zonal distribution of the life stage categories Ci of C. marshallae over the Oregon continental shelf was modeled by the two-dimensional ðx; z; tÞ equation [23], where wbi is the vertical swimming speed of the ith stage, assumed to be a sinusoidal function of time: wbi ¼ wsi sinð2ptÞ, with wsi the maximum vertical migration speed of the ith stage. @Ci ðx; z; tÞ @ ½ua ðx; z; tÞCi ðx; z; tÞ þ @t @x @ ½wa ðx; z; tÞCi ðx; z; tÞ þ @z ½23 @ @Ci ðx; z; tÞ @ @Ci ðx; z; tÞ Kðx; tÞ Kðz; tÞ @x @x @z @z @ ½wbi ðx; z; tÞCi ðx; z; tÞ ¼ population dynamics þ @z The population dynamics model was presented in eqns [17]–[20].



Distance offshore (km)

10

0

0.3

10 20 30 40 50

0.2 0.1 Eggs (A) 40

50

20

30

10 0.5

0

0.6

10

0.4

20

0.3

0.2

30

Nauplii 0.1

40

(B)

50

0.1

30

0.2

20

10 0.3

0 10 20

Copepodites CI_III

30

Depth (m)

40

50

(C)

Depth (m)

20

30

Depth (m)

40

50

40 50

Figure 5 Model of Calanus marshallae in the Oregon upwelling zone. The figure shows the simulated zonal distribution of (A) eggs, (B) nauplii, and (C) early copepodites at noon on 15 August. Concentrations of each stage are expressed as a fraction of the total population (all stages m3). (Reproduced with permission from Wroblewski JS, 1982. Interaction of currents and vertical migration in maintaining Calanus marshallae in the Oregon upwelling zone—a simulation. Deep Sea Research 29: 665–686.

The upwelling zone extended 50 km from the coast down to a depth of 50 m, and was divided into a grid with spacing 2.5 m in depth and 1 km in the horizontal. The author used a finite-difference scheme with a time step of 1 h, which fell within the bounds for computational stability (Figure 5). Coupling IBMs and Spatially Explicit Models

Individual-based models are more and more frequently used to assess the influence of space on the population dynamics, namely on the time course of population abundances and the pattern formation of populations in their habitats. This approach uses simulated currents from sophisticated 3D hydrodynamic models driving Lagrangian models of particle trajectories to examine dispersion processes. The approach is relatively straightforward and is a first step in formulating spatially explicit individual based models. Given a ‘properly resolved’ flow field, particle (larval fish/zooplankton/meroplanktonic larvae) trajectories are computed (generally with standard Runge–Kutta integration methods of the

553

velocity field). Specifically, hydrodynamic models provide the velocity vector v ¼ ðu; v; wÞ as a function of location x ¼ ðx; y; zÞ and time t, and the particle trajectories are obtained from the integration of eqn [24]. dx ¼ vðx; y; z; tÞ dt

½24

The simplest model of dispersion is a random walk model in which individuals move along a line from the same starting position. These trajectories could be modified by turbulent dispersion as described in the section below. Once the larval/particle position is known, additional local physical variables can be estimated along the particle’s path: e.g., temperature, turbulence, light, etc., and input to the IBM. The physical quantities are then included in biological (physiological or behavioural) formulations of IBMs (see below). Simulations considering trajectories of plankton as passive particles are a necessary step before considering any active swimming capability of planktonic organisms. They show the importance of physical features in the aggregation or dispersion of the particles. Plankton transport models that include biological components typically use a prescribed vertical migration strategy for all or part of an animal’s life history or a vertical motion (sinking or swimming) that is determined by animal’s development and growth. The simulated plankton distributions from these models tend to compare better with observed distributions than do models that use passive particles. Sensitivity studies show that behavior is an important factor in determining larval transport and/ or retention. The coupling of IBMs of zooplankton and fish populations and 3D circulation models is a recent field of study, even for fish models. Generally, models that describe the spatial heterogeneity of the habitat have been designed to answer questions about the spatial and temporal distribution of a population rather than questions about the numbers and characteristics of surviving individuals. They allow us to explore the potential effects of habitat alteration on these populations. Using this approach, biological mechanisms that are strongly dependent on habitat and that are not fully understood could be studied by examining different scenarios. Modeling Behavioral Mechanisms, Aggregation and Schooling, and Patches

Different types of models have been built, some of them focusing on the structure and shape of


554


aggregations depending on internal and external physical forces, others dealing with the benefits for individuals of living in groups with regard to feeding (foraging models) and to predation. The Lagrangian approach can take into account the behavior of individual organisms and the effects of the physical environment upon them. Although Eulerian approaches are mathematically tractable, the methods do not explicitly address the density dependence of aggregating individual behavior within a patch. Dynamic optimization allows descriptions of the internal state of individuals, which may lead to both variable and fluctuating motivations among individuals over short time periods.

Interactions between Populations Models with Plankton Populations in Interaction

Simple models of two species interactions take the form of eqns [25] and [26]. dN1 ¼ r1 N1 k1 N1 N2 dt

½25

dN2 ¼ r2 N2 k2 N1 N2 dt

½26

These population models represent some special experimental situations or typical field situations. Interactions between two species have been rarely treated by population models with description of the life cycle, although structured population models as well as IBM models can represent interactions between species such as predation, parasitism, or even cannibalism. As an example, Gaedke and Ebenho¨h presented in 1991 a study on the interaction between two estuarine species of copepods, Acartia tonsa and Eurytemora affinis. They first used a simple model based on eqns [25] and [26] including (a) predation (including self-predation of immature stages) by Acartia on the two, (b) a term of biomass gain of Acartia by this predation, and (c) a density-dependent loss term caused by predation by invertebrates or by starvation of the two species. This simple model did not result in stable coexistence between the two species with a reasonable parameter range under steady-state conditions. These authors then used two-stage-structured population models with stage-specific interactions (with similar equations to [17]–[20]) allowing the predation of large individuals of A. tonsa (copepodites 4 to adults) on nauplii of both species to be represented. The results of this detailed numerical model were compared with results obtained using

the simpler model with two variables. The predation on nauplii by Acartia tonsa appears to be key factor in the interaction of the two copepod populations.

Food Webs with Population Models

Structured models should be chosen to stimulate the dynamics of several interacting species. The stagebased approach will be acceptable with few species, but quickly become intractable with increasing numbers of species. In this case, a community model based on size structure and using prey–predator size ratio is the alternative approach. There is a continuum of models from detailed size spectrum structure up to large size classes representing functional (trophic) groups in food web models. The detailed size spectrum approach is particularly useful when simulating the predation of a fish cohort on its prey, whereas large functional groups are required for large-scale ecosystem models. Numerous examples include models with size structure of herbivorous zooplankton populations and their prey, and their interactions, in a nutrient–phytoplankton– herbivore–carnivore dynamics model. Size-based plankton model with large entities consider the size range 0.2–2000 mm, picophytoplankton, bacterioplankton, nanophytoplankton, heterotrophic flagellates, phytoplankton, microzooplankton, and mesozooplankton.

See also Biogeochemical Data Assimilation. Carbon Cycle. Fish Larvae. Fish Migration, Vertical. Fish Predation and Mortality. Gelatinous Zooplankton. Krill. Lagrangian Biological Models. Large Marine Ecosystems. Marine Mesocosms. Microbial Loops. Network Analysis of Food Webs. Nitrogen Cycle. Ocean Gyre Ecosystems. Patch Dynamics. Phosphorus Cycle. Plankton. Polar Ecosystems. Small-scale Patchiness, Models of. Small-Scale Physical Processes and Plankton Biology. Upwelling Ecosystems.

Further Reading Carlotti F, Giske J, and Werner F (2000) Modelling zooplankton dynamics. In: Harris RP, Wiebe P, Lenz J, Skjoldal HR, and Huntley M (eds.) Zooplankton Methodology Manual, pp. 571--667. New York: Academic Press. Caswell H (1989) Matrix Population Models: Construction, Analysis and Interpretation. Sunderland, MA: Sinauer Associates.



Coombs S, Harris R, Perry I and Alheit J (1998) GLOBEC special issue, vol. 7 (3/4). DeAngelis DL and Gross LJ (1992) Individual-based Models and Approaches in Ecology: Populations, Communities and Ecosystems. New York: Chapman and Hall. Hofmann EE and Lascara C (1998) Overview of interdisciplinary modeling for marine ecosystems. In: Brink KH and Robinson AR (eds.) The Sea, vol. 10, Ch. 19, pp. 507--540. New York: Wiley. Levin SA, Powell TM, and Steele JH (1993) Patch Dynamics. Berlin: Springer-Verlag. Lecture Notes in Biomathematics 96.. Mangel M and Clark CW (1988) Dynamic Modeling in Behavioral Ecology. Princeton, NJ: Princeton University Press.

555

Nisbet RM and Gurney WSC (1982) Modelling Fluctuating Populations. Chichester: Wiley. Renshaw E (1991) Modelling Biological Populations in Space and Time. Cambridge: Cambridge University Press. Cambridge Studies in Mathematical Biology. Tuljapurkar S and Caswell H (1997) Structured-population Models in Marine, Terrestrial, and Freshwater Systems. New York: Chapman and Hall. Population and Community Biology Series 18. Wood SN and Nisbet RM (1991) Estimation of Mortality Rates in Stage-structured Populations. Berlin: SpringerVerlag. Lecture Notes in Biomathematics 90. Wroblewski JS (1982) Interaction of currents and vertical migration in maintaining Calanus marshallae in the Oregon upwelling zone – a simulation. Deep Sea Research 29: 665--686.


POPULATION GENETICS OF MARINE ORGANISMS D. Hedgecock, University of Southern California, Los Angeles, CA, USA & 2009 Elsevier Ltd. All rights reserved.

Introduction This article provides a brief overview of the principles of population genetics and its applications in ocean science. The specialized vocabulary of genetics is defined, and central concepts and approaches are summarized in an abbreviated historical context. Finally, specific topics that have been addressed in the marine biological literature illustrate major areas of application of population genetics in ocean science.

Definitions and Historical Approaches Population genetics is the branch of genetics that explores the consequences of Mendelian inheritance at the level of populations, rather than families. Population genetics seeks to explain the most elementary step of Darwinian evolution, change in population genetic composition over time. Genetic composition is described by the frequencies, in population samples, of ‘alleles’ (alternative states or forms of a particular gene or genetic marker that arise by mutation or recombination), ‘phenotypes’ (the perceivable properties of organisms, such as their color or the banding patterns of their macromolecules following electrophoretic separation), or ‘genotypes’ (the hereditary constitutions of individuals, usually inferred from their phenotypes), at particular genes or ‘loci’ (singular ‘locus’, the place on a chromosome where a genetic element is located). ‘Genes’ are functional units of inheritance, whether they code for the synthesis of proteins or RNA or serve some regulatory or structural purpose, but nonfunctional portions of the genome are often used as genetic markers. The zygotes of most animals and plants are ‘diploid’, inheriting one copy of each gene from their parents. A diploid genotype, comprising two alleles that are identical either phenotypically or by descent, is said to be ‘homozygous’. A genotype comprising two different alleles is ‘heterozygous’. The proportion of heterozygous genotypes in a population, the ‘heterozygosity’, is an important measure of genetic diversity. Genetically determined traits or markers are said to be ‘polymorphic’ when two or more forms or alleles occur in a population at frequencies greater than expected

556

from recurrent mutation. Conversely, a ‘monomorphic’ or ‘fixed’ genetic trait or marker is represented by a single phenotype within a population comprising only a single allele (homozygotes). The primary forces that change genetic composition over time are recurrent mutations, natural or artificial selection, gene flow among subpopulations, and random changes resulting from sampling errors in finite populations. In the absence of these forces, the genetic composition of a ‘randomly mating’ population (i.e., one in which all possible matings or unions of gametes are equally likely) is stable through time and is described by the Hardy–Weinberg (H–W) principle. Under this basic principle, the frequencies of genotypes at a single, polymorphic locus are given by the expansion of the binomial (with two alleles) or multinomial (with three or more alleles) function of allelic frequencies. For example, at a locus with two alleles of equal selective value, A and a, having relative frequencies p and q (p þ q ¼ 1.0) in a large population, the frequencies of the genotypes, AA, Aa, and aa, are given by the binomial expansion, (p þ q)2 ¼ p2 þ 2pq þ q2, respectively. Populations conforming to this principle are said to be in H–W equilibrium. The H–W principle, which is easily extended to multiple alleles, simplifies the description of populations, by reducing the number of parameters to n alleles at a locus, rather than the n(n þ 1) genotypes formed by n alleles in diploid organisms. There are three distinct and complementary approaches in population genetics. Empirical studies seek to document patterns of genetic variation in natural populations and to explain these patterns in terms of the primary forces acting on population genetic variation and the population biology of the species in question. Experimental studies seek to test the effects of inbreeding and crossbreeding, natural and artificial selection, random genetic drift owing to population bottlenecks or finite population size, or to estimate the frequency and fitness effects of spontaneous mutations. Finally, theoretical studies continue to develop and elaborate the mathematical backbone of population genetics set down in the seminal works of Sewall Wright, Sir Ronald Fisher, and J.B.S. Haldane published in the early 1930s. In the marine sciences, the bulk of population genetic studies fall into the empirical category. The earliest population genetic studies, carried out by Charles Bocquet in France and Bruno Battaglia in Italy in the 1950s and 1960s, described color polymorphisms in natural populations of copepods and


POPULATION GENETICS OF MARINE ORGANISMS

isopods. The introduction, in the late 1960s, of methods for electrophoretic separation of allelic protein variants (or ‘allozymes’) allowed investigation of genetic variation in organisms not amenable to laboratory culture or lacking visible polymorphisms. Among the first allozyme studies of marine organisms were those by Fred Utter and colleagues on Pacific salmon, which provided a scientific basis for management of salmon ocean fisheries. In the 1980s, methods for the analysis of polymorphism in DNA sequences were developed, first for mitochondrial DNA but eventually for single copy nuclear genes. Today, the highly abundant and polymorphic, di-, tri-, and tetra-nucleotide tandem repeat elements known as microsatellites are the most widely used DNA markers, though less polymorphic but even more abundant single nucleotide polymorphisms (SNPs), which are amenable to high-throughput genotyping technologies, promise genome-wide coverage in the near future. Experimental genetic studies were initiated by Bocquet, Battaglia, and others working on small and easily cultured crustaceans but have exploded over the past two decades in conjunction with the growth of aquaculture. In this arena, molecular markers are being used not only to document variation in populations but also to construct linkage maps, which when combined with experimental analysis of complex phenotypes, such as growth and survival, permit the mapping of genes controlling variance in these quantitative traits (known as quantitative trait loci or QTL mapping). Application of technologies for gene-expression profiling and even whole genome sequencing in species with well-developed experimental methods are likely to increase greatly our understanding of phenotypic evolution in marine populations in the near future. Over the past three decades, biological oceanographers and fisheries scientists have had a growing interest in the theory and practice of population genetics, particularly as enhanced by molecular methods. Major scientific questions being addressed by population geneticists working with marine animals and plants can be grouped under five headings: (1) identification of morphologically cryptic, sibling species; (2) amount and spatial structure of genetic diversity within species; (3) temporal genetic change, particularly in relation to recruitment; (4) retrospective analyses of historical collections; and (5) phylogenetic and phylogeographic analyses, which aim, respectively, to determine the evolutionary relationships among species or higher taxa and to reveal the geographical distribution of genes and genealogies within species. Some of the prominent

557

applications of population genetics in the marine environment concern the connectivity among populations, design of marine reserves, aquaculture and stock enhancement and their ecological affects, and mixed-stock analyses in fisheries.

Morphologically Cryptic, Sibling Species A major contribution of population genetic studies to marine biology has been the identification of sibling or cryptic biological species within morphologically defined taxa. ‘Sympatric’ (occupying the same geographical area), cryptic species are most often recognized by obvious failures of the H–W principle at one or more markers. If the cryptic species are fixed for alternative alleles at a marker, one observes AA and BB homozygotes only and no or few AB heterozygotes, contrary to expectations for random mating. When cryptic species are ‘allopatric’ (occupying nonoverlapping areas), inference of biological species status is less clear. Fixed differences at multiple genetic markers suggest an absence of genetic exchange, but additional evidence is needed to ascertain whether the populations are reproductively isolated. Morphologically, cryptic speciation might also be suggested by large divergence in DNA sequences, but, again, ancillary evidence is required to ascertain whether this variation is concordant with reproductive isolation. The percent nucleotide difference maintained within conspecific populations for mitochondrial DNA ranges from a few percent in many species to as much as 18% in the copepod Tigriopus californicus. Cryptic species have been discovered in several well-studied taxa, including the blue mussel Mytilus, the worm Capitella, and the copepod Calanus. These discoveries, mostly serendipitous byproducts of research directed at other questions, suggest that a systematic investigation of the frequency of sibling species in various taxa could eventually increase estimates of marine species diversity by an order of magnitude. Prudence dictates that the taxonomy of all target organisms of oceanographical study should be confirmed by both traditional and molecular methods and that voucher specimens be kept. A major advantage of using DNA markers is that voucher specimens can often be preserved simply in ethanol. Molecular diagnosis of cryptic species or of the cryptic stages of species should enable the extension of species diagnosis to early life stages. Correct species identification of larval stages should enable, in turn, the acquisition of data on species-specific patterns in dispersal and recruitment.


558


Amount and Spatial Structure of Genetic Diversity Within Species Early studies of protein polymorphism revealed widely varying levels of genetic diversity among marine taxa. Comparative studies of taxa differing in life history or ecological traits have provided few compelling general explanations for what maintains different levels of genetic diversity in different taxonomic groups. Certainly, the amount of diversity measured for a particular taxon depends largely on the particular type and set of genetic markers studied. However, the amount of genetic diversity in a species depends more fundamentally on the mating system and the balance among the forces of mutation, migration, selection, and random genetic drift owing to finite population size. We focus on the last factor first. The H–W principle applies to infinitely large or very large populations. However, this important assumption of the H–W principle is often violated in real populations, which are finite in size. Therefore, to understand the maintenance and structure of genetic diversity in finite natural populations, we require an understanding of the concepts of random genetic drift and ‘effective population size’ (Ne). Random genetic drift is defined as change in allelic frequency resulting from the sampling of finite numbers of gametes from generation to generation. One way to illustrate genetic drift is to consider the transition matrix that describes the probability of having i A alleles in generation t þ 1, given j A alleles in generation t. The elements of this matrix are calculated from the binomial probability:

With this transition matrix, one can calculate how the distribution of allelic frequencies among a collection of populations will evolve. The allele frequency distribution is a vector, giving the proportion, yj,t, of populations with j A alleles at generation t, the sum of which is 1.0. We can calculate the change in the allele frequency distribution from one generation to the next by multiplying the matrix, X, of probability transition values (i.e., the xij values) by the vector, Y, of population states (i.e., the yj,t values), such that Yt þ 1 ¼ XYt. This recursive relationship can be generalized as Yt ¼ XtY0, assuming an initial allele frequency distribution Y0. How allele frequency and heterozygosity evolve by random genetic drift in a collection of small populations, each of size three (2N ¼ 6), is illustrated in Table 1. The initial frequency of the A allele in all populations is 0.5 (i.e., y3,0 ¼ 1.0). Several important features of genetic drift are illustrated by Table 1. (1) The proportion of polymorphic populations (those with one to five A alleles) rapidly declines, until all populations become fixed with either zero or six A alleles. (2) The frequency of A across the collection of all subpopulations remains at the mean of 0.5, so that allelic diversity is conserved across groups. (3) Finally, heterozygosity for the total population declines from the initial H–W equilibrium value, 2pq ¼ 2 (0.5) (0.5) ¼ 0.5, to zero, as subpopulations become fixed. The heterozygosity at generation t can be calculated as

Ht ¼ 2

2N X j¼0

ð2NÞ! j 2Ni j i 1 xij ¼ ð2N iÞ! i! 2N 2N where 2N is the number of alleles and the frequency of A at generation t is j/2N.

j 1 2N

j yj;t 2N

The recursive relationship between heterozygosity at generation t and heterozygosity at generation t þ 1 is readily solved, yielding the important fact that heterozygosity declines from one generation to the next by

Table 1 The evolution by random genetic drift of allele frequencies and heterozygosity (H ) in a collection of small populations, each of size three (2N ¼ 6) with initial frequency of allele A equal to 0.5 (see text for complete discussion) Number of A alleles

Generation 0

1

2

4

8

y

N

0 1 2 3 4 5 6

0.0 0.0 0.0 1.0 0.0 0.0 0.0

0.015 6 0.093 8 0.234 4 0.312 5 0.234 4 0.093 8 0.015 6

0.072 8 0.132 6 0.189 3 0.210 6 0.189 3 0.132 6 0.072 8

0.196 0.109 4 0.128 4 0.132 5 0.128 4 0.109 4 0.196

0.353 0.053 8 0.061 7 0.063 1 0.061 7 0.053 8 0.353

y y y y y

0.5 0.0 0.0 0.0 0.0 0.0 0.5

Mean frequency H

0.5 0.5

0.5 0.416 7

0.5 0.347 2

0.5 0.241 1

0.5 0.116 3

y y

0.5 0.0



the factor 1/(2N). This is what makes population size such an important parameter in population and conservation genetics. The finite populations discussed above are mathematically ideal populations that differ from real populations in a number of important ways. In the mathematically ideal population, there are equal numbers of both sexes, adults mate at random, and variance in number of offspring per adult is binomial or Poisson. In actual populations, the sexes may not be equal in number, mating may not be at random, or the variance in offspring number may be larger than binomial or Poisson. The effective size, Ne, is the size of a mathematically ideal population that has either the same rate of random genetic drift or the same rate of inbreeding as the actual population under study. Variance and inbreeding effective sizes are similar unless the population experiences a rapid change in size. Here, we are concerned primarily with the variance effective size, which determines the rate at which genetic variants are lost (p ¼ 0.0) or fixed (p ¼ 1.0) by random genetic drift. The number of adults in the ideal population (N) is, by definition, equal to the effective size, and the ratio, Ne/N ¼ 1.0 in the ideal case. Owing to the nonideal properties of real populations, the Ne/N ratio for most terrestrial vertebrate populations is thought to lie between 0.25 and 0.75. However, as we shall see in the next section, this ratio may be much smaller in highly fecund marine species. In actual populations, Ne is substituted for N in theoretical calculations; for example, heterozygosity is lost at a rate of 1/(2Ne) per generation. The spatial structure of genetic variation within a species is, perhaps, the most thoroughly studied aspect of marine population genetics. The question of how variation in mode of larval development, therefore larval dispersal potential, affects the genetics and evolution of marine species was raised over a quarter of a century ago. To investigate this question, it is useful to review the theory of population divergence developed largely by Sewall Wright. We focus primarily on the balance between random genetic drift and migration. The role of diversifying selection in causing populations to diverge is not easily generalized; moreover, the specific consequences of any selection regime can depend on the underlying balance of drift and migration. The genetic diversity of a species can be partitioned into components within and among population units, ranging from local, randomly mating populations (or demes) to subpopulations to the total species. Wright partitioned genetic variation within a species, using F-statistics, which measure the average genetic correlation between pairs of gametes derived from

559

different levels in a population hierarchy. At the basal level of this hierarchy, the correlation between the two gametes drawn from the same individual is symbolized as FIS. FIS is zero in a randomly mating subpopulation but is positive when there are excesses of homozygotes relative to H–W expectations. Mating among related individuals can cause an excess of homozygotes, in which case FIS is equivalent to the coefficient of inbreeding, f. Let us return to the example of a single locus with two alleles, A and a, having relative frequencies p and q in a particular local population. With partial inbreeding, the frequencies of genotypes, AA, Aa, and aa, are p2 þ fpq, 2pq(1-f), and q2 þ fpq, respectively. This is a generalization of the H–W principle, in which f (or FIS) governs how alleles associate into genotypes. When f (or FIS) is zero, the proportions of each of the three genotypes are the same as the H–W equilibrium; when f is 1.0, the population contains no heterozygotes. Most sexually reproducing marine populations conform to H–W equilibrium, so that FIS is close to zero. Clonal marine populations, such as some sea anemones and corals, often have nonzero FIS, owing to a mixture of sexual and vegetative propagation of genotypes. If a species is subdivided into partially isolated, finite subpopulations, mating among individuals in the total population cannot take place at random. At the same time, there is genetic drift within each subpopulation. The effect on the proportion of genotypes in the species is analogous to the effect of inbreeding. Local populations tend toward fixation, with a decline in heterozygosity, but genetic diversity is preserved among rather than within subpopulations (Table 1). The genetic correlation between gametes drawn from different individuals within a subpopulation, with respect to allelic frequencies in the ¯ pÞ. ¯ In total population, is given by FST ¼ s2p =pð1 ¯ this equation, p is the average frequency of allele A in the total population and s2p is the variance of p among subpopulations. Thus, FST is the ratio of the variance of allelic frequencies among subpopulations to the maximum variance that would be obtained if each subpopulation were fixed for one of the alternative alleles without change in mean allelic frequency. When local populations diverge from one another, there is an excess of homozygotes and a deficiency of heterozygotes, with respect to random mating expectations in the total population. The frequencies of AA, Aa, and aa in the total population are, thus, p¯ 2 þ s2p , 2p¯ q¯ 2s2p , and q¯ 2 þ s2p , respectively. Note the resemblance of these genotypic frequencies to those in an inbreeding population. The principle is readily understood at the extreme, in which each subpopulation is fixed for one allele or another; in this case, there are no heterozygotes in the total


560


population. This partitioning of genetic diversity can be extended to any number of hierarchical levels. In the absence of gene exchange among local populations of finite size, FST increases through time as local populations become fixed through the process of random genetic drift (Table 1). Local fixation can be prevented, however, by ‘gene flow’, the exchange of gametes or individuals among local populations. Suppose that a large population is subdivided into many ¯ subpopulations, with an average allele frequency, p, and that each subpopulation exchanges a proportion, m, of its population with a random sample of the whole population every generation. This is Wright’s well-known ‘‘island’’ model of population structure. If we consider a single subpopulation, the change in allelic frequency per generation will be ¯ Although this model is not realistic Dp ¼ mðp pÞ. (e.g., exchange among all subpopulations is not likely to be equal), it illustrates the principle that the frequency of an allele in a subpopulation will be pulled toward the mean frequency of immigrants. When m is small, the equilibrium between genetic drift and gene flow in the island model is approximated by FST ¼ 1/ (4Nm þ 1), which leads to a simple estimate of Nm, the absolute number of migrants exchanged among demes. Such estimates from data on real populations must be regarded as provisional, however, as the assumption of equilibrium underlying the estimate are likely rarely, if ever, met. Gene flow is a powerful cohesive force holding populations together. Indeed, in the island model, the exchange of one migrant every other generation (Nm ¼ 0.5) is theoretically sufficient to prevent the chance loss or fixation of an allele with an initial mean frequency of 0.5 in most subpopulations. However, owing to departures from the ideal populations in the island model, the exchange of thousands of individuals per generation between neighboring subpopulations of a widely distributed species could be insufficient to prevent considerable random drift. Moreover, the balance between genetic drift and gene flow is established over many generations. Divergence (or convergence) following an interruption (or resumption) of gene flow can take hundreds of generations, so that genetic similarity is not sufficient evidence for current gene flow. On the other hand, genetic similarity owing to high gene flow does not mean that subpopulations are demographically linked. Ecological and genetic processes operate over very different time scales. Spatial structure of genetic diversity or population subdivision is the topic of most interest to oceanographers, because it is tied to the hope that the geographic sources of recruitment to marine animal populations might be identified by their genetic

makeup. However, for most species of interest, those that comprise the zooplankton broadly speaking, this hope is not well founded in logic or fact. Dispersal among geographic populations, perhaps even low levels of dispersal, can eliminate the very genetic differences among populations that would permit identification of provenance. The marine population genetics literature supports the generalization that species with dispersing, planktotrophic (feeding) larvae have, as expected, much less geographic variation or population subdivision than species with poorly dispersing, lecithotrophic (nonfeeding) larvae. FST is generally much less than 0.05 for marine species with planktonic larvae. The oceanographer’s problem is to detect if a sample of zooplankton is a genetic mixture and if so, to determine the contributions of different geographic populations to the mix. This problem has been solved in mixed-stock fisheries, particularly for anadromous species. Sophisticated statistical analysis of genetic data can identify, with accuracy and precision, the relative contributions of discrete salmon stocks to ocean catches. With the advent of highly polymorphic microsatellite DNA markers, it is even possible to assign individuals to their population of origin. However, these methods work well for species like salmon because the source populations are identifiable in space and are genetically distinct, with FST values generally above 0.05 and as high as 0.5. These same methods are not likely to work for the many marine species that lack obvious spatial genetic structure. Although identification of sources and sinks of zooplanktonic populations is critical to understanding their distribution and abundance, it is unlikely to be achieved with indirect genetic methods alone. The fundamental limitation may be the dispersal biology of such species, which can easily homogenize the frequencies of alleles at all loci, whether allozymes or microsatellite DNA markers. Studies of spatial genetic variation should be made as a part of any baseline population genetic description of a marine species. As a rule, one should not expect such studies to yield conclusive information about the sources of recruits or the water masses bearing them. Nevertheless, any given study might reveal an exception to this rule or a previously unrecognized barrier to dispersal that has resulted in a major genetic subdivision.

Temporal Genetic Change Despite the generalization that marine species with planktonic dispersal tend to be genetically homogeneous over large geographic regions, allelic



frequencies among samples, taken sometimes on scales of meters, often show unpredictable patchiness. Planktonic larvae and patterns of recruitment are thought to play a role in such paradoxical observations. Temporal genetic studies, which are few, further suggest that as much genetic variation can be observed among recruits to a single place as can be observed among adult populations on spatial scales of 100s to 1000s of kilometer. Thus, the population genetics of marine planktonic larvae or new recruits are spatially and temporally much more dynamic than expected, given that their adult populations are large and well connected by larval dispersal. This anomalous feature of marine population genetics, like patterns of larval settlement themselves, may be attributed to the dynamics of the ocean environment. Genetic heterogeneity on fine spatial scales and temporal change may both be explained by the hypothesis that highly fecund marine organisms may have a large variance in individual reproductive success. This variance in reproductive success may result from a sweepstakes-chance matching of reproductive effort with oceanographic conditions conducive to spawning, fertilization, larval survival, and recruitment. According to this hypothesis, only small fractions (from 1/100 to 1/100 000) of spawning adults effectively reproduce and replace standing adult populations each generation. In this case, ratios of Ne/N may be far less than those observed in terrestrial animals (0.25–0.75), and random genetic drift of allelic frequencies should be measurable in some populations. This prediction has been borne out by temporal studies, in which observed genetic drift implies effective population sizes that are many orders of magnitude less than the simple abundance of adults. Studies of larval populations and comparisons of recruits and adults confirm a second prediction, that specific cohorts of larvae or recruits should show genetic evidence of having been produced by only a segment of the potential parental pool and may even show significant levels of full- or half-sib relatedness. This hypothesis establishes a connection between oceanography and population genetics in the study of recruitment and may explain how local adaptations and speciation can occur in seemingly large and well-mixed marine populations. Temporal genetic studies are therefore required to make sense of population genetic structure.

Retrospective Analyses of Historical Collections Enzymatic amplification of specific DNA sequences by the polymerase chain reaction (PCR) can potentially

561

be applied to the study of preserved zooplankton or fish scale collections. Molecular methods could facilitate rapid systematic treatment of these collections and establish genetic histories for particular species of interest. The power of such an approach is illustrated by a genetic study of Georges Bank on haddock populations, which revealed significant heterogeneity in the frequencies of mitochondrial DNA genotypes between the 1975 and 1985 year-classes. Temporal change in this case was attributed to immigration rather than variance in reproductive success. On the other hand, in a study of New Zealand red snapper, for which dried scales were available from 1950 to 1986 and fresh material from 1998, microsatellite DNA analysis showed that loss of genetic diversity and changes in allele frequencies in the Tasman Bay stock of nearly 7 million fish were consistent with an effective population size of 176 individuals. The authors attribute this very low Ne/N ratio to successful breeding by relatively few adults.

Phylogenetics, Phylogeography, and Paleoceanography Molecular genetic analyses are being used to reconstruct phylogenies or evolutionary relationships at all taxonomic levels and for many phyla. The application of phylogenetic methods to molecular as well as organismal traits, such as morphology, life history, and behavior, will undoubtedly shed new light on the evolution and systematics of marine organisms. Comparisons of genetic divergence within and between closely related species separated by barriers of known age (e.g., the Bering Strait or the Isthmus of Panama) are useful for calibrating rates of molecular evolution and reconstructing the history of faunal exchanges. Ultimately such studies may provide a basis for the development of biotechnological tools for rapid and automated classification of oceanographic samples and collections. Within the past decade, the application of phylogenetic approaches to molecular variation within species has yielded new insights into population histories and a synthesis of the traditionally separate disciplines of systematics and population genetics. Studies of mitochondrial DNA in several species living along the Gulf of Mexico and Atlantic coasts of the US have revealed concordant patterns of major genetic subdivisions. In all of these species, Gulf genotypes give way to Atlantic genotypes across a previously recognized biogeographical boundary in southeastern Florida. Gulf and Atlantic genotypes represent clades that separated over a million years ago. Thus, ‘phylogeographic’ patterns reflect the


562


persistence of historical events in the gene pools of organisms. Such information is relevant to management and conservation efforts and begs for interdisciplinary, paleoceanographic explanation. Other biogeographic boundaries should be similarly studied to assess whether they generally coincide with genetic discontinuities. A review of genetic differences between conspecific populations on either side of Point Conception, California, failed to find a consistent pattern of discontinuity like that associated with the Florida peninsula. Nevertheless, past dispersal, rather than present oceanographic connections, appears to be the message written into the genetics of contemporary populations.

See also Population Dynamics Models.Population Genetics of Marine Organisms

Further Reading Avise JC (2004) Molecular Markers, Natural History, and Evolution, 2nd edn., 684pp. Sunderland, MA: Sinauer.

Hauser L, Adcock GJ, Smith PJ, Bernal Ramirez JH, and Carvalho GR (2002) Loss of microsatellite diversity and low effective population size in an overexploited population of New Zealand snapper (Pagrus auratus). Proceedings of the National Academy of Sciences of the United States of America 99: 11742--11747. Hedgecock D, Launey S, Pudovkin AI, Naciri Y, Lapegue S, and Bonhomme F (2007) Small effective number of parents (Nb) inferred for a naturally spawned cohort of juvenile European flat oysters Ostrea edulis. Marine Biology 150: 1173--1182. Hedrick PW (2005) Genetics of Populations, 3rd edn., 737pp. Boston, MA: Jones and Bartlett. Hellberg ME, Burton RS, Neigel JE, and Palumbi SR (2002) Genetic assessment of connectivity among marine populations. Bulletin of Marine Science 70(supplement S): 273--290. Luikart G and England PR (1999) Statistical analysis of microsatellite DNA data. Trends in Ecology and Evolution 14: 253--256. Luikart G, England PR, Tallmon D, Jordan S, and Taberlet P (2003) The power and promise of population genomics: From genotyping to genome typing. Nature Reviews Genetics 4: 981--994. Whitlock MC and McCauley DE (1999) Indirect measures of gene flow and migration: FSTa1/(4Nm þ 1). Heredity 82: 117--125.


PORE WATER CHEMISTRY D. Hammond, University of Southern California, Los Angeles, CA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2263–2271, & 2001, Elsevier Ltd.

Introduction As marine sediments are deposited, they trap sea water in the pore space between grains. This pore water (sometimes called interstitial water) may represent more than 90% of the volume of the bulk sediment in fine-grained deposits near the sediment– water interface. The volume fraction of the bulk sediment that is water is called the porosity. Porosity usually decreases rapidly with increasing depth through the uppermost sediments, with the profile depending on lithology and accumulation rate. Ultimately, porosity may decrease to only 5–20% as the sediment becomes lithified. As sediments are compacted during burial, the pore fluid is squeezed upward, traveling through fractures, burrows, or perhaps the sediment pore space itself. As sediment is buried, its composition may be modified by chemical reactions, a process called chemical diagenesis. Pore water provides a medium that permits a solute to migrate from a site where it is produced to another site where it may be removed. For example, in organic-rich sediments, pyrite (FeS2) is a common end product of diagenesis. While the intermediate steps in pyrite formation are not fully understood, they involve sulfate reduction to sulfide at sites where reactive organic matter is found, and reduction of insoluble ferric oxides to form soluble ferrous iron at sites of iron-bearing minerals. Iron sulfides have very low solubility, and their deposition is usually localized at one of the two sites. If sulfide is released faster than ferrous iron, the sulfide diffuses from its site of production on an organic-rich particle to the mineral containing iron, where it forms insoluble iron sulfides. This can produce pyrite overgrowths on iron-bearing minerals. Alternatively, if the iron is more readily released, it diffuses to form iron sulfides near the sites of organic particles such as shells or localized pockets of organic substrate. Because diagenesis may alter only a small fraction of the solid phases, its impact may be difficult to detect from studies of solid phases alone. Pore water chemistry is much more sensitive to such changes. For example, in a sediment of 80% porosity, dissolution

of 0.1 weight percent CaCO3 from the solid phase (near the detection limit measurable in solid phases) would make a change of 6 mmol kg1 in the concentration of dissolved calcium. This change in pore water would be easily detectable because it results in a concentration 60% greater than that in the starting sea water. Of course, this calculation assumes that the pore water acts as a closed system, which is generally not the case as noted below. However, this example illustrates that pore water chemistry is more sensitive than solid phase chemistry to diagenesis. Studies of pore waters have become a standard tool for understanding the biogeochemical processes that influence sediments, and considerable efforts have been invested during the past several decades to develop techniques to collect samples, evaluate whether vertical profiles exhibit artifacts introduced during collection and handling, and develop approaches to model the observed profiles and obtain quantitative estimates of reaction kinetics and stoichiometry. Usually, modeling approaches assume steady-state behavior, but when time-dependent constraints can be established, nonsteady-state approaches can be applied.

Reasons to Study Pore Water Composition The study of pore waters can reveal many processes that are important in regulating the biogeochemical cycles of the ocean and in evaluating the impact of diagenesis on the sedimentary record recorded in solid phases. Some applications of pore water studies are given below. Calculation of Mineral Stability

Thermodynamic calculations can be carried out to determine which solid phases should be dissolving, precipitating, or in equilibrium with the pore fluid chemistry. While these calculations do not guarantee the presence of minerals that are at or above saturation, or the absence of minerals that are undersaturated, they are a very useful indicator of whether it may be worthwhile to search for minerals that could be present in only trace abundance. Identification of Sites of Reaction and Reaction Stoichiometry

Maxima in pore water profiles indicate localized inputs, and minima define sinks for solutes. However,


563

564

PORE WATER CHEMISTRY

defining the transition between source and sink regions requires location of inflection points in the pore water profile. If the pore water profile is in steady-state, material balance calculations can be carried out for solutes if the transport mechanisms are known. This usually involves fitting a reactiontransport model that defines the depth dependence of the net reaction-kinetics. Transport processes are discussed below, and may include molecular diffusion, advection, and macrofaunal irrigation. The impact of reactions occurring at depths beyond the range of sampling may also be evident in pore water profiles. Interpretation of the Sedimentary Record

Are changes in solid phase profiles with depth due to diagenetic reactions during burial, or due to temporal variations in the composition or rain rate of solid phase inputs? Pore water profiles provide a way to evaluate the contemporary rates of reactions (assuming they are in steady-state) and predict the effect of diagenetic reactions on solid phase profiles. Solid phase changes that exceed the diagenetic effects derived from modeling pore water profiles must reflect nonsteady-state behavior in the input of solid phases to the sediment column. Estimation of Benthic Exchange Rates

Sediments are a sink or source for many solutes in the water column. Thus, they can play an important role in regulating the composition of the overlying waters. This approach provides information that can be compared to direct measurements of benthic fluxes (see Platforms: Benthic Flux Landers) Recovery of Deep Pore Water that may be Fossil Water

The trapping of pore fluids as sediments are buried may potentially preserve fluid from a time when ocean composition differed from the present, such as the last glacial period when salinity should have been greater than at present and the isotopic composition of water should have been heavier. However, pore water is an open system, and diffusion facilitates the re-equilibration between fossil pore water and bottom waters. Consequently, relict signals may be difficult to detect, even in the absence of any influence of diagenetic reactions.

Sampling Techniques Initial studies of pore waters utilized retrieval of cores, sectioning them into intervals, and centrifugation to

separate pore waters from the associated solids. This approach works well for many solutes in sediments with high porosity, as long as appropriate precautions are taken to minimize artifacts (changes in composition attributable to recovering and processing samples). During sample processing, it is often critical to regulate temperature and eliminate contact between reducing sediments and oxygen, depending on the solute of interest. For studies related to nearsurface diagenesis, it is essential to obtain cores that have undisturbed interfaces and with bottom water still in contact with the sediment. To extract water from low porosity sediments, or minimize contact with gas phases, squeezing techniques have been developed. Several kinds of squeezing devices have been utilized, but all rely on compressing sediments while permitting water to escape through a filtration assembly. To avoid artifacts associated with retrieving cores from the deep sea, devices have been developed to carry out filtration in situ. These devices avoid the pressure- and temperature-dependent perturbations associated with core retrieval, but have their own logistical difficulties in deployment to minimize leakage and obtain accurate sampling resolution. One strategy drives a probe called a harpoon into sediments; openings at various distances along the harpoon shaft permit water to be drawn through filters into sample reservoirs. Another device collects cores, seals the bottom, and squeezes water by driving a piston and filter pack down onto the core; sequential aliquots of water are collected and assumed to represent water from progressively deeper intervals. Other strategies have relied on inserting probes that contain water that may communicate with pore water through a dialysis membrane. These devices, named ‘peepers’, require several days to equilibrate with pore waters and must be initially filled with a solution that will not significantly contaminate the surrounding sediment with exotic solutes. Several of the artifacts noted above may be avoided through the use of in situ electrodes that can be inserted directly into sediment and measure activities of various solutes. Systems to measure oxygen and pH are often used. Very recently, new electrodes to measure pCO2, sulfide, iron, and manganese have been developed. By using microelectrodes, gradients over short distances can be resolved. Finally, some tools have been developed to retrieve pressurized cores and extract pore fluids onboard ships at in situ pressures (see Deep-Sea Drilling Methodology). These tools are particularly important where high quantities of methane are found, either dissolved in pore fluid or as a gas hydrate.



Limitations of Pore Water Studies Observation of the Net Process

Some solutes may be involved in more than one reaction. This limits the ability to uniquely define reaction stoichiometry.

565

to localized sites of reaction, but the pore water measurement of sectioned cores defines an average for the zone sampled. Most studies are designed to evaluate vertical gradients, as it is usually difficult to evaluate any horizontal gradients, if they are present. Sampling Artifacts

Required Assumptions

Pore water profiles are usually assumed to be steadystate. If boundary conditions vary, pore waters respond, but there is a temporal lag in response that increases with depth. The profile should exhibit concentrations that are roughly averaged over this response time. A second problem may be the existence of unidentified transport processes, such as macrofaunal irrigation (see below). A third problem is that patchiness of organisms on the seafloor may lead to localized effects, such as caches of freshly deposited organic matter in burrows. It is not always possible to collect and process sufficient cores to evaluate the spatial heterogeneity introduced by these effects, that may occur on horizontal scales of centimeters to meters. In the Equatorial Pacific, for example, benthic fluxes of oxygen calculated from pore water profiles collected with replicate cores at the same site have been shown to vary by 30%, and inferences based on a single core have an inherent uncertainty. Sampling Resolution

Gradients may exist over very short vertical or horizontal distances that cannot be easily resolved during sampling. In organic-rich slope sediments, for example, microelectrode measurements show that the thickness of oxygenated sediments may be only 1–2 mm. Furthermore, if micro-environments are present within burrows or inside shells, this can lead Table 1

These may be created by changes in temperature, pressure, exposure to oxygen (or perhaps any gas phase), activities of stressed organisms, or deterioration of samples between collection and storage (see Table 1). Some of these changes appear to be reversible, while others are not. Some changes involve the direct reaction of the solute in question, while others are indirect due to the co-precipitation or adsorption of one solute with the solid formed by the direct reaction of another.

Diagenetic Reactions and Biogeochemical Zonation The seafloor receives a rain of sediment that is a mixture of biogenic debris and detrital materials. Some of these components are rather reactive and undergo diagenesis at very shallow depth. Of paramount importance are reactions related to the oxidation of reduced carbon in organic material to form carbon dioxide. The details of organic carbon diagenesis are not well understood, and a detailed discussion of relevant reactions is beyond the scope of this article. However, organic carbon diagenesis involves microbial catalysis of reactions that result in decreasing the free energy of the system through the transfer of electrons from the organic material to terminal electron acceptors. Dissolved organic carbon is produced, and some escapes from sediments

Known artifacts in pore water studies

1. Changes in temperature. Consistently observed for boron, potassium, and silicon; sometimes observed for acid, calcium, and magnesium. 2. Changes in pressure. Precipitation of carbonate during retrieval of cores from deep water is consistently observed in carbonate-rich sediments. As a consequence, phosphate and uranium are often lost. 3. Exposure to oxygen. Consistently observed for ferrous iron. As a consequence, phosphate, silicon, and perhaps other metals are affected by precipitating ferric oxyhydroxides. 4. Gas exchange. Exposure to a gas phase permits any gas dissolved in the sample to partition among the available phases present. In sediments under pressure, high concentrations of dissolved gases (e.g. methane,) can accumulate and form bubbles when retrieved to the surface. Also, some containers are permeable to certain types of gases. 5. Stressed organisms. Some animals may excrete large amounts of ammonia when they are stressed. Stress can occur due to temperature and pressure changes as cores are retrieved, or due to physical disturbance. Some bacteria may contain vacuoles rich in nitrate; these vacuoles may break when samples are centrifuged. 6. Deterioration during sample storage. Filtering samples can screen out most bacteria, but may not inhibit reactions that might occur inorganically. For example, oxygen will eventually diffuse into plastic sample bottles and oxidize ferrous iron. Precipitation of ferric iron can be inhibited by acidification of the sample.


566


into the overlying water, but most of these reactions result in production of carbon dioxide, which reacts with water to form carbonic acid. A consistent pattern has been observed in the distribution of the principal terminal electron acceptors with increasing distance from an oxic water column. Oxygen, nitrate, manganese dioxide, ferric oxides, sulfate, and finally carbon dioxide serve as the principal electron acceptors. Some representative reactions are illustrated in Table 2. This sequence of reactions has led to the concept of biogeochemical zonation, with each zone named for the solute that serves as the principal electron acceptor or that is the principal product (Figure 1). The sequence of zones is determined by the chemical free energy yields released by possible reactants. The thickness of each zone is dependent on the rate of reaction consuming the electron acceptor, the rate of a reactant’s transport through sediments, and the concentration of the acceptor in bottom waters or in solid phases. Zones may overlap, and tracer studies have shown that they need not be mutually exclusive. The existence of the deeper zones depends on the availability of sufficient reactive organic matter.

Table 2

Boundaries between zones are often interesting sites, and may provide environments where specialized bacteria thrive. As soluble reduced reaction products form at depth, they may diffuse upward into the overlying zone, where they are oxidized. One example is illustrated in Figure 2. In iron-rich systems, ferrous iron produced at depth diffuses upward, until it encounters nitrate or oxygen diffusing downward. At this horizon, ferrous iron is oxidized to ferric iron that precipitates as an oxyhydroxide, often leaving a visible thin red band in the sediments that marks the ferrous/ferric transition. Continued accumulation of new sediments transports the ferric oxyhydroxide downward relative to the sediment– water interface, beyond the penetration depth of oxidants, where the ferric iron is again reduced to ferrous iron; the ferrous iron diffuses upward again and is re-oxidized. Studies of pore water have confirmed this redox shuttle system, which maintains a horizon of sediments rich in ferric iron at a consistent depth relative to the sediment–water interface. Manganese can undergo a similar cycle to produce a manganese-rich horizon. In sulfur-rich sediments that are overlain by oxygenated bottom waters, the

Biogeochemical zonation and proton balance



Corg

567

Electron acceptors _ 2_ (O2, NO3 , MeO, SO4 )

CO2 Particle rain Water Sediment Corg

4+

O2 reduction

_

C + 4e

_

NO3 reduction 2_

SO4 reduction CH4 production

Burial of metal sulfides

Figure 1 Biogeochemical zonation. The rain of organic carbon to the seafloor and its burial provide a substrate for metabolic activity, as microbial communities transfer electrons from the organic carbon to terminal electron acceptors. This results in the conversion of organic carbon into carbon dioxide, and may be accompanied by the conversion of oxidized forms of nitrogen, sulfur, and metals (MeO) into reduced forms: molecular nitrogen that escapes and metal sulfides that are buried. Sediments can be divided into zones, characterized by the principal acceptor that is present, or by the key product (in the case of methane). In some cases, distinct zones may be observed where manganese and iron are the principal acceptors, but these often overlap with the nitrate and sulfate zones. This schematic does not include the details of transport, but acceptors migrate downward from overlying waters, or are produced in an upper zone and diffuse downward. The drawing is not to scale. The relative thickness of each zone varies, depending on input of reactive organic material and the availability of different acceptors, and the deeper zones do not form where the rain of labile organic materials is too low. The oxygen reduction zone may be only a few millimeters thick in margin sediments, tens of centimeters thick under open-ocean equatorial sediments, and many meters thick in open-ocean sediments that underlie oligotrophic waters. The geometry of each zone may be convoluted due to the presence of macrofaunal burrows or other heterogeneities.

Pore water concentration

Solid phase concentration Recycled ferric oxyhydroxides

Oxidant 2+

Fe

Oxidation

Depth

Diffusion

Burial

2+

Fe

Fe (OH)3

Reduction

Fe (OH)3

2+

Fe

Figure 2 Schematic illustration of iron cycling to maintain a diagenetic front at a constant depth near the sediment–water interface, as explained in the text. Only the recycled component of iron is shown.

upward diffusion of sulfide through pore waters brings it into contact with dissolved oxygen. This provides a unique environment that may be exploited by the sulfur-oxidizing bacteria Beggiatoa, an organism that utilizes the energy released by sulfide oxidation and forms bacterial mats (see Marine Mats) frequently found where low oxygen bottom waters may exclude predators.

Diagenetic Modeling Substantial advances have been made in the development of mathematical models to quantify the

effects of the diagenetic processes that influence pore water profiles. Important processes include those mentioned below, but a full discussion of appropriate formulations to describe these processes is beyond the scope of this article (see the Further Reading section for the mathematical development). Chemical Reactions

Some solids dissolve and others precipitate. Many reactions are driven by redox processes associated with the oxidation of organic matter, largely catalyzed by microbial activities as they extract metabolic energy. Some of these are illustrated in


568


Table 2. These redox reactions can also result in the production or consumption of acid, depending on the suite of reactants available. The addition of acid should be buffered by the dissolution of carbonate solids, if they are present, and removal of acid may promote precipitation of carbonates. For example, as shown in Table 2B, utilization of oxygen as a terminal electron acceptor is quite effective in producing acid, and should favor the dissolution of any carbonate minerals in the oxygenated zone. If manganese or iron serves as the terminal electron acceptor, acid is consumed, and carbonate may precipitate from the pore waters. If sulfate serves as the terminal electron acceptor, and no iron oxides are available to permit iron sulfides to precipitate, pore waters are acidified and may dissolve carbonates. A similar behavior occurs if iron can be leached from silicates and exchanged for dissolved magnesium. However, if iron oxyhydroxides are present, they dissolve and favor carbonate precipitation. This example illustrates the importance of iron availability in helping regulate the pH of marine pore waters. The behavior of nitrogen (oxidized to nitric acid if oxygen is present, and consuming a proton if it is released as ammonia in the absence of oxygen) also contributes to pH buffering. Many other solid phases respond to these pH changes, changes that are largely influenced by the oxidation of organic matter and the behavior of carbonate and iron-bearing minerals. Another interesting effect is the co-precipitation of some trace constituents with phases created by cycles of a more abundant substance. For example, as ferrous iron diffuses upward, other trace metals or phosphate may co-precipitate with the oxyhydroxides when oxygen or nitrate is encountered. These constituents may re-dissolve as the ferric oxyhydroxides are buried more deeply.

Reversible Adsorption and Desorption

Pore waters are in intimate contact with solid surfaces, and may exchange solutes reversibly on short timescales. Consequently, a change in pore water concentration is accompanied by an additional change in the adsorbed inventory. If a solute is strongly adsorbed, its transport will be dominated by movement of the particulate phase, rather than by migration through the dissolved phase. The ionic speciation of the solute is often critical in determining the degree of adsorption. Adsorption and desorption equilibria are temperature dependent, and are responsible for some of the artifacts noted earlier.

Diffusion

The random motion of solutes in pore waters results in a net transport from regions of high concentration to regions of low concentration, as described by Fick’s laws for diffusion. This is the principal mechanism for transport over short distance scales, and the rate of diffusion depends on the solute in question, the porosity of the sediment, the temperature, and to some extent on the ensemble of other ions present and their concentration gradients. This last effect results from the speciation of the ion and cross-coupling among ion gradients that produces electrical potentials. A rough estimate of the relationship between length- and time-scales for which diffusion is effective can be obtained from using Fick’s Second Law to evaluate the time for a diffusive front to migrate from a perturbation introduced in a one-dimensional system. The time is given by the relationship t ¼ x2/(2D), where t is the time for response at a distance x from the perturbation in a sediment with a diffusivity of D. In deep-sea sediments, many solutes have diffusivities in the range of 5 106 cm2 s1, so the diffusive front migrates approximately 1 cm in 1 day, but requires 30 years to migrate 1 m and 0.3 million years to migrate 100 m. Advection

It is most convenient to define pore water spatial coordinates relative to the sediment–water interface. If sediments are accumulating and contain some water, pore waters must be moving relative to this interface. In addition, flow may be driven by strong heating at depth, or other externally imposed forcing. These directed flows are considered advection and may be an important transport mechanism. The relative importance of advection in transporting a solute can be evaluated by considering the dimensionless Peclet number, D/UL, where D represents the diffusivity through sediments, U is the flow velocity, and L is the length scale over which transport must be accomplished. If this number is much larger than one, advection can be ignored, and transport is dominated by diffusion. Irrigation

In coastal and slope sediments, organisms create burrows that act as conduits through which bottom waters may move (see Macrobenthos). Flow through a burrow may be driven by active pumping by the organism, or by hydrodynamic effects created by the flow of bottom waters past the burrow orifice. Thus, the presence of burrows creates a complex geometry for the effective shape of the sediment–water



interface and for the biogeochemical zonation that should parallel the burrow walls. Burrows may provide a pathway for communication between deep sediment and overlying water that does not depend

on vertical diffusion. Because communication depends only on transport through the dissolved phase, it is called irrigation. This transport is distinct from bioturbation, a process that may also be important,

Concentration

Concentration

O2

Depth

TCO2

Diffusion + reaction

Diffusion

Salt dissolution at discrete horizon

Cl

569

No reaction

_

(B)

(A)

Concentration

Concentration

O2

Nitrification Denitrification

Reaction rate

Reaction, diffusion, irrigation

_

NO 3 +

NH4

Only reaction + diffusion

Production rate

TCO2 (C)

(D)

Figure 3 Schematic pore water profiles for various combinations of reaction and transport that conceptually illustrate observations made in field studies. Depth scales vary widely for these examples. (A) Reaction at a discrete horizon, with diffusive transport from the site of reaction to the overlying water. An example of this is the profile of chloride in sediments of the Mediterranean Sea created by dissolution of deeply buried (several hundred meters) salt deposits. Another example of this style of profile includes observations in ODP pore waters that reveal interactions of basaltic basement rocks where they contact pore waters in the overlying sediment, removing magnesium and releasing calcium. (B) Coupled reaction–diffusion profiles. In this example, organic carbon is oxidized to carbon dioxide (TCO2) utilizing oxygen as the electron acceptor. The reaction is assumed to occur only above the dashed line, and the curvature defines whether the solute is consumed (O2) or produced (TCO2). (C) Coupled nitrification and denitrification. In this example, it is assumed that any ammonia released by degradation of organic matter is completely oxidized to nitrate (nitrification) if it enters the oxic zone (including ammonia diffusing from below). The nitrate produced diffuses downward and is converted to N2 in the nitrate reduction zone (denitrification). Inflection points in the curves define the horizon separating these zones. Because of the competing production of nitrate in the oxic zone and removal in the nitrate reduction zone, sediments may be a net source or net sink for nitrate in the overlying water column, depending on the relative availabilities of reactive organic matter and oxygen. In this example, the competing reactions balance so there is no nitrate gradient at the sediment–water interface. Open-ocean sediments are relatively efficient at recycling their fixed nitrogen during diagenesis, while denitrification in margin sediments may lose 30–70% of the fixed nitrogen that rains to the seafloor in those locations. (D) Reaction, diffusion, and irrigation. In this example, TCO2 is produced throughout the sediment column, but the rate decreases with increasing depth (dotted line). Reaction products are transported by diffusion and irrigation in the upper zone, and by diffusion only in the lower zone. If sampling defines the average concentration at each depth, the nonlocal irrigation effect provides an apparent sink for TCO2.


570


but it is slower and involves the physical mixing of both solids and water by organismal activities. Accurately quantifying the effect of irrigation has posed a significant modeling challenge, and for convenience it is usually introduced into equations as a parameter called non-local transport. Irrigation effect are commonly observed in estuarine, shelf, and slope locations overlain by oxygenated bottom water. The importance of irrigation in promoting benthic exchange depends on the density of burrows, the depth to which they penetrate, how frequently they are flushed, and the length scale over which diagenetic reactions for the solute of interest take place. Effect of Irrigation on pore water profiles have not been detected in sediments overlain by bottom waters low in oxygen, due to the exclusion of macrofauna, or in deep-sea settings where macrofauna are less abundant than in margin settings. Irrigation effects are greatest in the upper 10 cm, but pore water profiles in some deep basins of the California Borderland indicate observable effects to nearly 2 m.

dependence, but the shape may not have a unique interpretation, particularly if these assumptions are not valid. It is also important to remember that concentration gradients of solutes adjust until transport is equal to the net reactions occurring. If a solute is involved in competing reactions, such as dissolution of one phase and precipitation of a less soluble phase, the net reaction could be zero and no concentration gradient would be created.

See also Authigenic Deposits. Benthic Boundary Layer Effects. Calcium Carbonates. Deep-Sea Drilling Methodology. Macrobenthos. Marine Mats. Ocean Margin Sediments. Sedimentary Record, Reconstruction of Productivity from the. Turbulence in the Benthic Boundary Layer.

Boundary Conditions

Pore water profiles are dependent on boundary conditions. An upper boundary condition is imposed by the solute concentration in the overlying water, although complications exist if the solute is very reactive. Above the sediment–water interface is a diffusive sublayer that may be several hundred micrometers thick in the deep sea, and thinner in more energetic environments (see Turbulence in the Benthic Boundary Layer). A significant concentration gradient may exist through this zone if the scale length characterizing the solute profile in the sediment is small enough to approach this distance. It can be more difficult to identify a lower boundary condition. Sometimes the solute will approach a constant value if its reactions cease at depth. In other cases, the solute may reach a constant value dictated by solubility constraints or by its disappearance.

Interpretation of Pore Water Profiles Some examples of pore water profiles are illustrated in Figure 3, drawn schematically to illustrate the range of behavior that may be observed when different factors are important. Several assumptions have been made in drawing these profiles. One is that they represent steady-state, relative to the sediment– water interface. A second is that the diffusivity is not depth-dependent. A third is that any reactions go to zero at infinite depth. A fourth is that advection has been ignored. The shape of a profile is a clue to interpret what factors are important and their depth

Further Reading Aller RC (1988) Benthic fauna and biogeochemical processes in marine sediments: The role of burrow structure. In: Blackburn TH and Sorensen J (eds.) Nitrogen Cycling in Coastal Marine Environments, pp. 301--338. New York: John Wiley. Berner RA (1974) Kinetic models for early diagenesis of nitrogen, sulfur, phosphorus, and silicon. In: Goldberg ED (ed.) The Sea, vol. 5, pp. 427--450. New York: John Wiley. Berner RA (1980) Early Diagenesis. Princeton: Princeton Press. Boudreau BP (1997) Diagenetic Models and Their Implementation. Berlin: Springer-Verlag. Boudreau BP and Jorgensen BB (2000) The Benthic Boundary Layer: Transport Processes and Biogeochemistry. New York: Oxford University Press. Boudreau BP (2000) The mathematics of early diagenesis: from worms to waves. Reviews of Geophysics 38: 389--416. Burdige DJ (1993) The biogeochemistry of manganese and iron reduction in marine sediments. Earth Science Reviews 35: 249--284. Fanning KA and Manheim FT (eds.) (1982) The Dynamic Environment of the Ocean Floor. Lexington, MA: DC Heath and Co. Froelich PN, Klinkhammer GP, Bender ML, et al. (1979) Early oxidation of organic matter in pelagic sediments of the eastern equatorial Atlantic: Suboxic diagenesis. Geochimica Cosmochimica Acta 43: 1075--1090. Lerman A (1977) Migrational processes and chemical reactions in interstitial waters. In: Goldberg E, McCave I, O’Brien J, and Steele J (eds.) The Sea, vol. 6, pp. 695--738. New York: John Wiley.



Li Y-H and Gregory S (1974) Diffusion of ions in sea water and deep-sea sediments. Geochimica Cosmochimica Acta 38: 703--714. Luther GW III, Reimers CE, Nuzzio DB, and Lovalvo D (1999) In situ deployment of voltammetric, potentiometric, and amperometric microelectrodes from a ROV to determine dissolved O2, Mn, Fe, S(-2), and pH in porewaters. Environmental Science and Technology 33: 4352--4356. Manheim FT and Sayles FL (1974) Composition and origin of interstitial waters of marine sediments, based on deep

571

sea drill cores. In: Goldberg ED (ed.) The Sea, vol. 5, pp. 527--568. New York: John Wiley. Nealson KH (1997) Sediment bacteria: Who’s there, what are they doing and what’s new? Annual Reviews of Earth and Planetary Science 25: 403--434. Van Der Weijden C (1992) Early diagenesis and marine pore water. In: Wolf KH and Chilingarian GV (eds.) Diagenesis III, Developments in Sedimentology, pp. 13--134. New York: Elsevier.


PRIMARY PRODUCTION DISTRIBUTION S. Sathyendranath and T. Platt, Dalhousie University, NS, Canada Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2272–2277, & 2001, Elsevier Ltd.

Introduction The study of the distribution of primary production in the world oceans preoccupied biological oceanographers for most of the twentieth century. Understanding the distribution remains a problem for which there is no facile answer; it is a many-faceted problem to which the approaches differ depending on the temporal and spatial scales at which understanding is sought. The issues relate to limitations of measurement techniques, interpretation of measurements that are available, and paucity of data. In this context, the first point to bear in mind is that every method for the measurement of primary production has an intrinsic timescale associated with it (Table 1). Space and timescales are inextricably

linked in oceanography, such that small timescales are always associated with small length scales, and large timescales with large length scales. Therefore, in studying the distribution of primary production in the ocean, one has to be careful that measurements are made at time and space scales appropriate for the problem at hand. Another important point is that all the available techniques do not measure the same quantity; the measured quantity may be gross primary production, net primary production, or net community production (if carbon-based methods are used), or total production, new production, or regenerated production (if nitrogen-based methods are used) (Table 1). A careful analysis quickly reveals that no methods are currently available for estimating the same component of primary production at all scales of interest in oceanography. It follows that, in comparing results (or in combining methods), one has to make sure that like things are being compared (or combined). Failure to do this can lead to perplexing results. Precise measurements of primary production are time consuming, and paucity of data has been a

Table 1 Various methods that can be used to estimate primary production in the ocean. The nominal timescales applicable to the results are also given. The components of primary production Pg (gross primary production), Pn (net primary production), and Pc (net community production) are based on rates of carbon uptake; PT (total primary production), Pr (regenerated production) and Pnew (new production) refer to nitrogen uptake rates. Sedimentation rate refers to the flux of organic particles from the photic zone Method

Nominal component of production

Nominal timescale

In vitro 14 C assimilation O2 evolution 15 NO3 assimilation 15 NH4 assimilation 18 O2 evolution

PT ( Pn) PT Pnew Pr Pnew ð Pc Þ

Hours Hours Hours Hours Hours

Bulk property NO3 flux to photic zone O2 utilization rate (OUR) below photic zone Net O2 accumulation in photic zone 238 234 U/ Th 3 3 H/ He

Pnew Pnew Pnew Pnew Pnew

Hours to days Seasonal to annual Seasonal to annual 1–300 days Seasonal and longer

Optical Double-flash fluorescence Passive fluorescence Remote sensing

PT PT PT ; Pnew

o1 s o1 s Days to weighted annual

Upper and lower limits Sedimentation rate below photic zone Optimal energy conversion of photons absorbed Depletion of winter accumulation of NO3

Pnew ( Pc ): (lower limit) PT (upper limit) Pnew (lower limit)

Days to months (duration of trap deployment) Instantaneous to annual Seasonal

572


to to to to to

1 1 1 1 1

day day day day day

(duration (duration (duration (duration (duration

of of of of of

incubation) incubation) incubation) incubation) incubation)

PRIMARY PRODUCTION DISTRIBUTION

constant difficulty. This necessitates the use of extrapolation schemes to produce large-scale distributions from a small number of observations. When undertaking such extrapolations, it would be desirable to test the extrapolation schemes by comparing the estimated distributions at large scales against some other independent estimates. However, the lack of techniques for measuring the same component of primary production at different time scales confounds efforts to make independent validations of extrapolation protocols. A profitable approach to dealing with these matters may be to begin by examining the factors that influence primary production. One may anticipate that the distribution of primary production would be influenced by the variability in the forcing fields.

Factors that Influence Variations in Primary Production Primary production is the rate of carbon fixation by photosynthesis. The primary forcing variable is therefore light. The light field varies with depth in the ocean, with time of day, and with time of year. Correspondingly, primary production in the ocean exhibits a strong depth dependence and a strong time dependence. The temporal variations occur on several scales, ranging from seconds (response to clouds and vertical mixing) to diurnal, seasonal, and annual. Adaptations of phytoplankton populations to various light regimes also influence primary production. The adaptations may involve, for example, change in the concentration of chlorophyll-a per cell, change in the number of chlorophyll-a molecules per photosynthetic unit, or changes in the concentrations of auxiliary pigments, which may play either a photoprotective or a photosynthetic role. Understanding photo-adaptation requires that we know the light history of the cells, in addition to the current light levels. Light acts on the state variable, the biomass of phytoplankton. It has been a common practice, in this field, to treat the concentration of the main phytoplankton pigment, chlorophyll-a, as an index of phytoplankton biomass, because it (or a variant called divinyl chlorophyll-a) is present in all types of phytoplankton, because of the fundamental role it plays in the photosynthetic process, and because it is easy to measure. The tendency would be for primary production to increase with chlorophyll-a concentration, though the rate of increase would vary depending on other factors. A third factor that determines the primary production in the ocean is the availability of essential

573

nutrients such as nitrogen. In a stratified, oceanic water column, the upper illuminated layer is typically low in nutrients, with the deeper layers acting as a reservoir of nutrients. Mixing events bring these nutrients to the surface layer, enhancing primary production. In temperate and high latitudes, deep mixing events in winter, and subsequent stratification as the surface warming trend begins, lead to the wellknown phenomenon of the spring bloom and more generally to a pronounced seasonal cycle in primary production. On the seasonal cycle are superimposed the effects of sporadic mixing events in response to passing storms. The short-term increases in primary production associated with these sporadic events are often missed by sampling schemes designed to record the seasonal cycle. Temperature is another factor that influences primary production. It is believed that temperature controls the enzyme-mediated dark-reaction rates of photosynthesis. Thus, from laboratory experiments, it has been shown that photosynthetic rates in phytoplankton increase with temperature up to an optimal temperature, after which they decrease. However, the details of this response may differ with species. In nature, the tendency of primary production rates to increase with temperature is confounded by another effect: upwelling waters with high nutrients tend to have a low temperature, and the increase in primary production in response to the nutrient supply may in fact supersede the counteracting effect of temperature. The recent years have seen an increasing appreciation of the role of micronutrients such as iron as limiting resources for primary production. The idea that iron limits production in certain marine environments was aired in the early twentieth century. However, it was in the 1980s that this idea gained renewed momentum, with the pioneering work by John Martin. It is now believed that iron limits primary production in large tracts of the ocean (the Southern Ocean, the Equatorial Pacific, the subarctic Pacific), leading to regimes known as the high-nutrient, low-chlorophyll regimes; these are environments where the nitrogen in the upper mixed layer is apparently never used up, and the phytoplankton biomass remains low throughout the year. Other contributing factors for the presence of high-nutrient, low-chlorophyll regimes include top-down control of phytoplankton biomass (and hence productivity) by zooplankton grazing, and the supply of nutrients by physical processes that exceeds the demands of biological production. It has been postulated that the variations in the distribution of primary production in response to changes in the availability of iron over geological timescales, and the consequent changes in


574


the draw-down of carbon dioxide from the atmosphere into the ocean, may be implicated in climate change. Species composition and species succession also influence rates of primary production. For example, the size distribution of the cells and the pigment composition of the cells can both change with species composition, thus influencing nutrient uptake and light absorption for photosynthesis. In view of the large number of factors that influence the distribution of primary production at so many temporal and spatial scales, it is desirable to apply mathematical modeling techniques to organize and formalize the study of the distribution of primary production in the world oceans. Light-dependent models of primary production are particularly useful, not only because the models are based on the first principles of plant physiology, but also because of their direct applicability to remote sensing of primary production. According to such models, primary production P at a given location (x, y, z) and a given time (t) can be formalized as: Pðx; y; z; tÞ ¼ Bðx; y; z; tÞf ðIðx; y; z; tÞÞ

½1

where B is the phytoplankton biomass indexed as chlorophyll-a concentration, and the function f describes the biomass-specific, photosynthetic response of phytoplankton to available light I. The response function f is known to have three phases: a lightlimited phase at low-light levels when production increases linearly as a function of available light; a saturation phase at high-light levels, when production becomes independent of light; and a photoinhibition phase at extremely high light levels, when production is actually reduced by increasing light. At the most, three parameters (or only two, if the photoinhibitory phase is negligible, which is often the case) are required to describe such a response. Often, these are taken to be the initial slope aB (typical units: mg C (mg chla)1 h1 (W m2)1); the saturation parameter PBm (typical units: mg C (mg chla)1 h1); and the photoinhibition parameter bB (typical units: mg C (mg chla)1 h1 (W m2)1). In such models, light and biomass are taken, justifiably, to be the principal agents responsible for variations in primary production. The effects of other factors (temperature, nutrients, micronutrients, species, light history and photoadaptation) may be accounted for indirectly, through their influence on the parameters of the response function (the photosynthesis–irradiance curve). The success of such models depends on how well we are able to describe, mathematically, the fields of biomass and light, as well as the parameters of the photosynthesis–

irradiance curve. Such models could be taken to a higher level of sophistication if the photosynthesis parameters were in turn expressed as functions of the various factors that influence them. General relationships valid in the natural environment, that would account for the influences of all known factors on photosynthesis–irradiance parameters, have eluded scientists so far. Therefore, assignment of parameters has to rely heavily on direct observations. With this background, we can look at what we know of the distribution of primary production in the world ocean.

Vertical Distribution It is only the upper, illuminated part of the water column that contributes to primary production. A useful rule of thumb is that practically all the primary production in the water column occurs within the euphotic zone, commonly defined as the zone that extends from the surface to the photic depth at which light is reduced to 1% of its surface value (though there are some arguments for using 0.1% light level as a more rigorous boundary on the euphotic zone). However, it is important to realize that photosynthesis is a quantum process, which depends on the absolute magnitude of light available, rather than on relative quantities as indicated by the photic depth. Thus, regardless of the definition of photic depth that may be used, it will only tell us that the production below that particular depth horizon will be small compared with that above; it tells us nothing about the absolute magnitude of production in the water column. The 1% light level may occur at depths exceeding 100 m in oligotrophic open-ocean waters, whereas it may be less than 10 m in eutrophic or turbid waters. It is noteworthy that phytoplankton themselves are a major factor responsible for modifying the optical properties of sea water, and hence the rate of penetration of solar radiation into the ocean. Another useful depth horizon that is relevant in the study of primary production is the critical depth. If production and loss terms of phytoplankton (grazing, sinking, decay) from the surface to some finite depth are integrated, then the integrated production and loss terms become equal to each other at some depth of integration, which is known as the critical depth. The concept of critical depth was formalized by Sverdrup in 1956. If the mixed-layer depth is shallower than the critical depth, then production in the layer will exceed losses, which is favorable for the accumulation of biomass in the layer, which would, in turn, further enhance mixed-layer



production. If the mixed layer is deeper than the critical depth, the conditions would be unfavorable for the formation of blooms. Thus, the vertical distribution of production relative to the mixed-layer depth plays an important role in determining the potential for enhanced production. The concept of critical depth is built on the fact that, within the mixed layer, production is typically a decreasing function of depth (because the available light decreases exponentially with depth), whereas the biomass and the loss terms are uniformly distributed within the mixed layer. However, it would be erroneous to suppose that the maximum primary production always occurs at the surface of the ocean. In fact, the maximum primary production may well occur at some subsurface depth, where the nutrient availability and light levels are optimal. In a stratified water column, several factors (species composition, photoadaptation, nutrient supply, and light levels) conspire to produce a deep-chlorophyll maximum. Depending on the physiological status of the phytoplankton in the deep-chlorophyll maximum, there may be a subsurface maximum in primary production, which may occur at the same depth as the chlorophyll maximum, or at a shallower depth. If the light levels at the surface are sufficiently high to induce photoinhibition, then also the maximum in production would occur at some subsurface level.

Horizontal Distribution Primary production varies markedly with region and with season. Upwelling regions (e.g., waters off north-west Africa, the north-west Arabian Sea off Somalia, the equatorial divergence zone, the northeast Pacific off California and Oregon, the south-east Pacific off Peru, and south-west Africa) are typically more productive than the central gyres of the major ocean basins, because of the high levels of nutrients that are brought to the surface by upwelling. In general, coastal regions are more productive than open-ocean waters, also due to the availability of nutrients. Temperate and high latitudes show a pronounced seasonal maximum in spring (a consequence of the high supply of nutrients to the mixed layer during deep mixing events in winter, followed by a shallowing of the mixed layer and an increase in incoming solar radiation as the seasons progress). In summer, the depletion of nutrients in the mixed layer leads to reduced production and biomass, and to the migration of the chlorophyll maximum to below the mixed layer. This is often followed in fall by a secondary peak in production, as a result of increased nutrient supply as the mixed layers begin to deepen,

575

and also perhaps a decrease in the grazing pressure. A notable exception to these seasonal cycles is the high-nutrient, low-chlorophyll regions mentioned earlier. Seasonal cycles are far less pronounced in tropical waters, unless they are influenced by seasonal upwelling, as is the case in the Arabian Sea. A summary of what we know of the distribution of primary production in the world’s oceans, and of the physical and biological factors that influence it is available in the literature. Given the highly dynamic nature of the distribution of primary production in the oceanic environment and the extreme paucity of measurements, it was only in the 1970s that compilations of measurements were first used to evaluate the distribution of marine primary production at the global scale. These studies had to combine observations from many years to maximize the number of measurements, so that it was impossible to examine interannual variability, or trends, in primary production. This inherent limitation to the study of the distribution of primary production at large scales was only addressed, at least partially, with the advent of remote-sensing techniques, which are examined briefly below.

The Use of Remote Sensing The last two decades of the twentieth century saw the development of remote sensing to study the distribution of phytoplankton in the ocean. This technology uses subtle variations in the color of the oceans, as monitored by a sensor in space, to quantify variations in the concentrations of chlorophyll-a in the surface layers of the ocean. Since polar-orbiting satellites can provide global coverage at high spatial resolution (1 km or better), it became possible for the first time to see (cloud cover permitting) in great wealth of detail the variations in phytoplankton distribution at synoptic scales. The next logical step in the exploitation of ocean-color data was taken a few years later, when these fields of biomass were converted into fields of primary production. The procedure that has met with the most success builds on models of photosynthesis as a function of available light. The light available at the sea surface can also be estimated using remote-sensing methods: geostationary satellites monitor cloud type and cloud cover, which are used in combination with atmospheric light-transmission models to estimate light available at the sea surface. Optical properties of the water, derived from ocean-color data are then used to compute light available at depth in the ocean.


576


Thus, information on the variations in the forcing field (light), and the state variable on which light operates (phytoplankton biomass) in the process of photosynthesis, are directly available through remote sensing. If these satellite data are supplemented with best estimates of photosynthesis–irradiance parameters based on in situ observations, then all the major elements are in place for computations of primary production by remote sensing. These computations can be further refined by incorporating information on vertical structure in biomass, also derived from in situ observations. In regions where deep chlorophyll maxima are a consistent feature, ignoring their presence can lead to some systematic errors in the results. The remote-sensing approach to computation of primary production has several advantages. It makes use of the vast database on light and biomass available through remote sensing, and in addition, makes use of all the available information on plant physiology obtained through in situ observations to define the parameters of the photosynthesis–irradiance curve. It has the potential to monitor the distributions of primary production at very large spatial

scales, and over several timescales, ranging from the daily to interannual. It emerges as the method of choice for studying the distribution of primary production at large scales, especially when it is considered that remote sensing provides information on chlorophyll-a concentration, which can vary over four decades of magnitude, and on the highly variable light field at the surface of the ocean. In this method, the more sparse in situ observations are made use of only to define the parameters of the photosynthesis–irradiance curve and the parameters that are used to describe the vertical structure in biomass, which have a much lower range of variability compared with chlorophyll-a concentration. The method is not without its problems, however. The major issues relate to differences in the time and space scales at which satellite and in situ measurements are made. This necessitates the development of methods for seamless integration of the two types of data streams. Ideally, the methods would bring to bear variations in the oceanographic environment and the physiological responses of phytoplankton to these variations. This is an active area of research. The first computation of global oceanic primary

75 88 103 120 141 165 193 226 265 310 363 426 498 583 683 800 Figure 1 Distribution of primary production (g C m2 year1) in the world oceans, as estimated using remotely sensed ocean-colour data (Longhurst et al., 1995). Areas of high production are seen in northern high latitudes; these reflect the impact of spring blooms on the distribution of production integrated over an annual cycle. Areas of upwelling (e.g., equatorial upwelling, the Benguela upwelling, the Somali upwelling, the California upwelling and the Peru upwelling) also emerge as areas of high production. The coastal shelf regions are also locations of high production. However, it is recognized that the remote-sensing method used in this calculation could overestimate the biomass, and hence the production in some coastal areas. Recent improvements in ocean-color technology offer the potential to reduce this type of error in the computation.



production using the remote-sensing approach appeared in the literature in 1995 (Figure 1). Other, similar computations have since appeared in the literature. It is a method that will continue to improve, with improvements in satellite technology as well as in the techniques for extrapolation of local biological measurements to large scales.

See also Microbial Loops. Network Analysis of Food Webs. Ocean Gyre Ecosystems. Pelagic Biogeography. Primary Production Methods. Primary Production Processes.

Further Reading Chisholm SW and Morel FMM (eds.) (1991) What Controls Phytoplankton Production in Nutrient-Rich Areas of the Open Sea? vol. 36 Lawrence KS: American Society of Limnology and Oceanography. Falkowski PG and Woodhead AD (eds.) (1992) Primary Productivity and Biogeochemical Cycles in the Sea. New York: Plenum Press.

577

Geider RJ and Osborne BA (1992) Algal Photosynthesis. New York: Chapman Hall. Li WKW and Maestrini SY (eds.) (1993) Measurement of Primary Production from the Molecular to the Global Scale, ICES Marine Science Symposia, vol. 197. Copenhagen: International Council for the Exploration of the Sea. Longhurst A (1998) Ecological Geography of the Sea. San Diego: Academic Press. Longhurst A, Sathyendranath S, Platt T, and Caverhill C (1995) An estimate of global primary production in the ocean from satellite radiometer data. Journal of Plankton Research 17: 1245--1271. Mann KH and Lazier JRN (1991) Dynamics of Marine Ecosystems. Biological–Physical Interactions in the Oceans. Cambridge, USA: Blackwell Science. Platt T and Sathyendranath S (1993) Estimators of primary production for interpretation of remotely sensed data on ocean color. Journal of Geophysical Research 98: 14561--14576. Platt T, Harrison WG, Lewis MR, et al. (1989) Biological production of the oceans: the case for a consensus. Marine Ecolog Progress Series 52: 77--88.


PRIMARY PRODUCTION METHODS J. J. Cullen, Department of Oceanography, Halifax, NS, Canada Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2277–2284, & 2001, Elsevier Ltd.

production with a broad variety of methods and for interpreting the measurements in a range of scientific contexts. It is nonetheless useful to define three components of primary production that can be estimated from measurements in closed systems:

•

Introduction Primary production is the synthesis of organic material from inorganic compounds, such as CO2 and water. The synthesis of organic carbon from CO2 is commonly called carbon fixation: CO2 is fixed by both photosynthesis and chemosynthesis. By far, photosynthesis by phytoplankton accounts for most marine primary production. Carbon fixation by macroalgae, microphytobenthos, chemosynthetic microbes, and symbiotic associations can be locally important. Only the measurement of marine planktonic primary production will be discussed here. These measurements have been made for many decades using a variety of approaches. It has long been recognized that different methods yield different results, yet it is equally clear that the variability of primary productivity, with depth, time of day, season, and region, has been well described by most measurement programs. However, details of these patterns can depend on methodology, so it is important to appreciate the uncertainties and built-in biases associated with different methods for measuring primary production.

Definitions Primary production is centrally important to ecological processes and biogeochemical cycling in marine systems. It is thus surprising, if not disconcerting, that (as discussed by Williams in 1993), there is no consensus on a definition of planktonic primary productivity, or its major components, net and gross primary production. One major reason for the problem is that descriptions of ecosystems require clear conceptual definitions for processes (e.g., net daily production of organic material by phytoplankton), whereas the interpretation of measurements requires precise operational definitions, for example, net accumulation of radiolabeled CO2 in particulate matter during a 24 h incubation. Conceptual and operational definitions can be reconciled for particular approaches, but no one set of definitions is sufficiently general, yet detailed, to serve as a framework both for measuring planktonic primary

578

• •

Gross primary production (Pg) is the rate of photosynthesis, not reduced for losses to excretion or to respiration in its various forms Net primary production (Pn) is gross primary production less losses to respiration by phytoplankton Net community production (Pnc) is net primary production less losses to respiration by heterotrophic microorganisms and metazoans.

Other components of primary production, such as new production, regenerated production, and export production, must be characterized to describe foodweb dynamics and biogeochemical cycling. As pointed out by Platt and Sathyendranath in 1993, in any such analysis, great care must be taken to reconcile the temporal and spatial scales of both the measurements and the processes they describe. Marine primary production is commonly expressed as grams or moles of carbon fixed per unit volume, or pet unit area, of sea water per unit time. The timescale of interest is generally 1 day or 1 year. Rates are characterized for the euphotic zone, commonly defined as extending to the depth of 1% of the surface level of photosynthetically active radiation (PAR: 400–700 nm). This convenient definition of euphotic depth (sometimes simplified further to three times the depth at which a Secchi disk disappears) is a crude and often inaccurate approximation of where gross primary production over 24 h matches losses to respiration and excretion by phytoplankton. Regardless, rates of photosynthesis are generally insignificant below the depth of 0.1% surface PAR.

Photosynthesis and Growth of Phytoplankton Primary production is generally measured by quantifying light-dependent synthesis of organic carbon from CO2 or evolution of O2 consistent with the simplified description of photosynthesis as the reaction: B8hn

CO2 þ 2H2 O 2- ðCH2 OÞ þ H2 O þ O2

½1

Absorbed photons are signified by hn and the carbohydrates generated by photosynthesis are


PRIMARY PRODUCTION METHODS

represented as CH2O. Carbon dioxide in sea water is found in several chemical forms which exchange quickly enough to be considered in aggregate as total CO2 (TCO2). In principle, photosynthesis can be quantified by measuring any of three light-dependent processes: (1) the increase in organic carbon; (2) the decrease of TCO2; or (3) the increase of O2. However, growth of phytoplankton is not so simple: since phytoplankton are composed of proteins, lipids, nucleic acids, and other compounds besides carbohydrate, both photosynthesis and the assimilation of nutrients are required. Consequently, many chemical transformations are associated with primary production, and eqn [1] does not accurately describe the process of light-dependent growth. It is therefore useful to describe the growth of phytoplankton (i.e., net primary production) with a more general reaction that describes how transformations of carbon and oxygen depend on the source of nutrients (particularly nitrogen) and on the chemical composition of phytoplankton. For growth on nitrate: 1:0NO3 þ 5:7CO2 þ 5:4H2 O -ðC5:7 H9:8 O2:3 NÞ þ 8:25O2 þ 1:0OH

½2

The idealized organic product, C5.7H9.8O2.3N, represents the elemental composition of phytoplankton. Ammonium is more reduced than nitrate, so less water is required to satisfy the demand for reductant: 1:0NHþ 4 þ 5:7CO2 þ 3:4H2 O -ðC5:7 H9:8 O2:3 NÞ þ 6:25O2 þ 1:0Hþ

½3

The photosynthetic quotient (PQ; mol mol1) is the ratio of O2 evolved to inorganic C assimilated. It must be specified to convert increases of oxygen to the synthesis of organic carbon. For growth on nitrate as described by eqn [2], PQ is 1.45 mol mol1; with ammonium as the source of N, PQ is 1.10. The photosynthetic quotient also reflects the end products of photosynthesis, the mixture of which varies according to environmental conditions and the species composition of phytoplankton. For example, if the synthesis of carbohydrate is favored, as can occur in high light or low nutrient conditions, PQ is lower because the reaction described in eqn [1] becomes more important. Uncertainty in PQ is often ignored. This can be justified when the synthesis of organic carbon is measured directly, but large errors can be introduced when attempts are made to infer carbon fixation from the dynamics of oxygen. Excretion of organic material would have a small influence on PQ and is not considered here.

579

Approaches Primary production can be estimated from chlorophyll (from satellite color or in situ fluorescence) if carbon uptake per unit of chlorophyll is known. Therefore, ‘global’ estimates of primary production depend on direct measurements by incubation. The technical objectives are to obtain a representative sample of sea water, contain it so that no significant exchange of materials occurs, and to measure lightdependent changes in carbon or oxygen during incubations that simulate the natural environment. Methods vary widely, and each approach involves compromises between needs for logistical convenience, precision, and the simulation of natural conditions. Each program of measurement involves many decisions, each of which has consequences for the resulting measurements. Several options are listed in Tables 1 and 2 and discussed below.

Light-dependent Change in Dissolved Oxygen

The light-dark oxygen method is a standard approach for measuring photosynthesis in aquatic systems, and it was the principal method for measuring marine primary production until it was supplanted by the 14C method, which is describged below. Accumulation of oxygen in a clear container (light bottle) represents net production by the enclosed community, and the consumption of oxygen in a dark bottle is a measure of respiration. Gross primary production is estimated by subtracting the dark bottle result from that for the light bottle. It is thus assumed that respiration in the light equals that in the dark. As documented by Geider and Osborne in their 1992 monograph, this assumption does not generally hold, so errors in estimation of the respiratory component of Pg must be tolerated unless isotopically labelled oxygen is used (see below). Methods based on the direct measurement of oxygen are less sensitive than techniques using the isotopic tracer 14C. However, careful implementation of procedures using automated titration or pulsed oxygen electrodes can yield useful and reliable data, even from oligotrophic waters of the open ocean. Interpretation of results is complicated by containment effects common to all methods for direct measurement of primary production (see below). Also, a value for photosynthetic quotient must be assumed in order to infer carbon fixation from oxygen production. Abiotic consumption of oxygen through photochemical reactions with dissolved organic matter can also contribute to the measurement, primarily near the surface, where the effective ultraviolet wavelengths penetrate.


580


Table 1

Measurements that can be related to primary production

Measurement

Advantages

Disadvantages

Comments

Change in TCO2

Direct measure of net inorganic C fixation

Not generally practical for open-ocean work

Change in oxygen concentration (high precision titration)

Direct measures of O2 dynamics can yield estimates of net and gross production Very sensitive and relatively easy. Small volumes can be used and many samples can be processed No problems with radioactivity

Relatively insensitive: small change relative to large background Small change relative to large background Interpretation of light-dark incubations is not simple Tracer dynamics complicate interpretations Radioactive – requires special precautions and permission Less sensitive and more work than 14C method Larger volumes required

A common choice when 14 C method is impractical

Requires special equipment

A powerful research tool, not generally used for routine measurements

Incorporation of 14 C-bicarbonate into organic material (radioactive isotope) Incorporation of 13 C-bicarbonate into organic material (stable isotope) Measurement of 18 O2 production from H18 2 O

Table 2

Measures photosynthesis without interference from respiration

Very useful if applied with great care. Requires knowledge of PQ to convert to C-fixation The most commonly used method in oceanography

Approaches for incubating samples for the measurement of primary production

Incubation system

Advantages

Disadvantages

Comments

Incubation in situ

Best simulation of the natural field of light and temperature

Limits mobility of the ship Vertical mixing is not simulated

Simulated in situ

Many stations can be surveyed Easy to conduct time-courses and experimental manipulations

Artifacts possible if deployed or recovered in the light Special measures must be taken to stimulate spectral irradiance and temperatureo Vertical mixing not simulated

Not perfect, but a good standard method if a station can be occupied all day

Photosynthesis versus irradiance (P versus E) incubator (14C)

Data can be used to model photosynthesis in the water column With care, vertical mixing can be addressed

Extra expenses and precautions are required Spectral irradiance is not matched to nature Results depend on timescale of measurement Analysis can be tricky

Light-dependent Change in Dissolved Inorganic Carbon

Changes in TCO2 during incubations of sea water can be measured by several methods. Uncertainties related to biological effects on pH-alkalinity-TCO2 relationships are avoided through the use of coulometric titration or infrared gas analysis after acidification. Measurement of gross primary production and net production of the enclosed community is like that for the light-dark oxygen method, but there is no need to assume a photosynthetic quotient. However, precision of the analyses is not quite as good as

Commonly used when many stations must be sampled. Significant errors possible if incubated samples are exposed to unnatural irradiance and temperature A powerful approach when applied with caution

for bulk oxygen methods. Extra procedures, such as filtration, would be required to assess precipitation of calcium carbonate (e.g., by coccolithophores) and photochemical production of CO2. These processes cause changes in TCO2 that are not due to primary production. The TCO2 method is not used routinely for measurement of primary production in the ocean. The

14

C Method

Marine primary production is most commonly measured by the 14C method, which was introduced by Steemann Nielsen in 1952. Samples are collected and



the dissolved inorganic carbon pool is labeled with a known amount of radioactive 14C-bicarbonate. After incubation in clear containers, carbon fixation is quantified by liquid scintillation counting to detect the appearance of 14C in organic form. Generally, organic carbon is collected as particles on a filter. Both dissolved and particulate organic carbon can be quantified by analyzing whole water after acidification to purge the inorganic carbon. It is prudent to correct measurements for the amount of label incorporated during incubations in the dark. The 14C method can be very sensitive, and good precision can be obtained through replication and adequate time for scintillation counting. The method has drawbacks, however. Use of radioisotopes requires special procedures for handling and disposal that can greatly complicate or preclude some field operations. Also, because 14C is added as dissolved inorganic carbon and gradually enters pools of particulate and dissolved matter, the dynamics of the labeled carbon cannot accurately represent all relevant transformations between organic and inorganic carbon pools. For example, respiration cannot be quantified directly. The interpretation of 14C uptake (discussed below) is thus anything but straightforward. The

13

C Method

The 13C method is similar to the 14C method in that a carbon tracer is used. Bicarbonate enriched with the stable isotope 13C is added to sea water and the incorporation of CO2 into particulate matter is followed by measuring changes in the 13C:12C ratio of particles relative to that in the TCO2 pool. Isotope ratios are measured by mass spectrometry or emission spectrometry. Problems associated with radioisotopes are avoided, but the method can be more cumbersome than the 14C method (e.g., larger volumes are generally needed) and some sensitivity is lost. The

18

581

The 18O method is sufficiently sensitive to yield useful results even in oligotrophic waters. It is not commonly used, but when the measurements have been made and compared to other measures of productivity, important insights have been developed.

Methodological Considerations Many choices are involved in the measurement of primary production. Most influence the results, some more predictably than others. A brief review of methodological choices, with an emphasis on the 14C method, reveals that the measurement of primary production is not an exact science. Sampling

Every effort should be made to avoid contamination of samples obtained for the measurement of primary production. Concerns about toxic trace elements are especially important in oceanic waters. Trace metalclean procedures, including the use of specially cleaned GO-FLO sampling bottles suspended from KevlarTM line, prevent the toxic contamination associated with other samplers, particularly those with neoprene closure mechanisms. Frequently, facilitates for trace metal-clean sampling are unavailable. Through careful choice of materials and procedures, it is possible to minimize toxic contamination, but enrichment with trace nutrients such as iron is probably unavoidable. Such enrichment could stimulate the photosynthesis of phytoplankton, but only after several hours or longer. Exposure of samples to turbulence during sampling can damage the phytoplankton and other microbes, altering measured rates. Also, significant inhibition of photosynthesis can occur when deep samples acclimated to low irradiance are exposed to bright light, even for brief periods, during sampling.

O Method

Gross photosynthesis can be measured as the production of 18O-labeled O2 from water labeled with this heavy isotope of oxygen (see eqn [1]). Detection is carried out by mass spectrometry. Net primary production of the enclosed community is measured as the increase of oxygen in the light bottle, and respiration is estimated by difference. In principle, the difference between gross production measured with 18O and gross production from light-dark oxygen changes is due to light-dependent changes in respiration and photochemical consumption of oxygen. Respiration can also be measured directly by 18 O2. tracking the production of H18 2 O from

Method of Incubation

Samples of seawater can be incubated in situ, under simulated in situ (SIS) conditions, or in incubators illuminated by lamps. Each method has advantages and disadvantages (Table 2). Incubation in situ ensures the best possible simulation of natural conditions at the depths of sampling. Ideally, samples are collected, prepared, and deployed before dawn in a drifting array. Samples are retrieved and processed after dusk or before the next sunrise. If deployment or retrieval occur during daylight, deep samples can be exposed to unnaturally high irradiance during transit, which can lead to


582


artifactually high photosynthesis and perhaps to counteracting inhibitory damage. Incubation of samples in situ limits the number of stations that can be visited during a survey, because the ship must stay near the station in order to retrieve the samples. Specialized systems both capture and inoculate samples in situ, thereby avoiding some logistical problems. Ship operations can be much more flexible if primary productivity is measured using SIS incubations. Water can be collected at any time of day and incubated for 24 h on deck in transparent incubators to measure daily rates. The incubators, or bottles in the incubators, are commonly screened with neutral density filters (mesh or perforated metal screen) to reproduce fixed percentages of PAR at the surface. Light penetration at the station must be estimated to choose the sampling depths corresponding to these light levels. Cooling comes from surface sea water. This system has many advantages, including improved security of samples compared with in situ deployment, convenient access to incubations for time-course measurements, and freedom of ship movement after sampling. Because the spectrally neutral attenuation of sunlight by screens does not mimic the ocean, significant errors can be introduced for samples from the lower photic zone where the percentage of surface PAR imposed by a screen will not match the percentage of photosynthetically utilizable radiation (PUR, spectrally weighted for photosynthetic absorption) at the sampling depth. Incubators can be fitted with colored filters to simulate subsurface irradiance for particular water types. Also, chillers can be used to match subsurface temperatures, avoiding artifactual warming of deep samples. Artificial incubators are used to measure photosynthesis as a function of irradiance (P versus E). Illumination is produced by lamps, and a variety of methods are used to provide a range of light levels to as many as 24 or more subsamples. Temperature is controlled by a water bath. The duration of incubation generally ranges from about 20 min to several hours, and results are fitted statistically to a P versus E curve. If P versus E is determined for samples at two or more depths (to account for physiological differences), results can be used to describe photosynthesis in the water column as a function of irradiance. Such a calculation requires measurement of light penetration in the water and consideration of spectral differences between the incubator and natural waters. Because many samples, usually of small volume, must be processed quickly, only the 14C method is appropriate for most P versus E measurements in the ocean.

Containers

Ideally, containers for the measurement of primary production should be transparent to ultraviolet and visible solar radiation, completely clean, and inert (Table 3). Years ago, soft glass bottles were used. Now it is recognized that they can contaminate samples with trace elements and exclude naturally occurring ultraviolet radiation. Glass scintillation vials are still used for some P versus E measurements of short duration; checks for effects of contaminants are warranted. Compared with soft glass, laboratory-grade borosilicate glass bottles (e.g., PyrexTM) have better optical properties, excluding only UV-B (320–400 nm) radiation. Also, they contaminate less. Laboratory-grade glass bottles are commonly used for oxygen measurements. Polycarbonate bottles are favored in many studies because they are relatively inexpensive, unbreakable, and can be cleaned meticulously. Polycarbonate absorbs UV-B and some UV-A 320–400 nm radiation, so near-surface inhibition of photosynthesis can be underestimated. The error can be significant very close to the surface, but not when the entire water column is considered. TeflonTM bottles, more expensive than polycarbonate, are noncontaminating and they transmit both visible and UV (280–4000 nm) radiation. When the primary emphasis is an assessing effects of UV radiation, incubations are conducted in polyethylene bags or in bottles made of quartz or TeflonTM. The size of the container is an important consideration. Small containers (r50 ml) are needed when many samples must be processed (e.g. for P versus E) or when not much water is available. However, small samples cannot represent the planktonic assemblage accurately when large, rare organisms or colonies are in the water. Smaller containers have greater surface-to-volume ratios, and thus small samples have greater susceptibility to contamination. If it is practical, larger samples should be used for the measurement of primary production. The problems with large samples are mostly logistical. More water, time, and materials are needed, more radioactive waste is generated, and some measurements can be compromised if handling times are too long. Duration of Incubation

Conditions in containers differ from those in open water, and the physiological and chemical differences between samples and nature increase as the incubations proceed. Unnatural changes during incubation include: extra accumulation of phytoplankton due to exclusion of grazers; enhanced inhibition of photosynthesis in samples collected from mixed



Table 3

583

Containers for incubations

Container

Advantages

Disadvantages

Comments

Polycarbonate bottle

Good for minimizing trace element contamination Nearly unbreakable Affordable More transparent to UV Incompressible

Excludes UV radiation Compressible, leading to gas dissolution and filtration problems for deep samples More trace element contamination Breakable Contaminate samples with trace elements and Si Exclude UV radiation More difficult to handle Requires caution with respect to contamination Relatively expensive

Many advantages for routine and specialized measurements at sea

Laboratory grade borosilicate glass (e.g., PyrexTM) Borosilicate glass scintillation vials Polyethylene bag

Quartz, TeflonTM

Small volume (1–25 ml)

Large volume (1–20 l)

Inexpensive Practical choice for P versus E Inexpensive Compact UV-transparent UV-transparent TeflonTM does not contaminate Good for P versus E Samples can be processed by acidification (no filtration) Some containment artifacts are minimized Potential for time-course measurements

layers and incubated at near-surface irradiance; stimulation of growth due to contamination with a limiting trace nutrient such as iron; and poisoning of phytoplankton with a contaminant, such as copper. When photosynthesis is measured with a tracer, the distribution of the tracer among pools changes with time, depending on the rates of photosynthesis, respiration, and grazing. All of these effects, except possibly toxicity, are minimized by restricting the time of incubation, so a succession of short incubations, or P versus E measurements, can in principle yield more accurate data than a day-long incubation. This requires much effort, however, and extrapolation of results to daily productivity is still uncertain. The routine use of dawn-to-dusk or 24 h incubations may be subject to artifacts of containment, but it has the advantage of being much easier to standardize.

Cannot sample large, rare phytoplankton evenly Containment effects more likely More work Longer filtration times with possible artifacts

A reasonable choice if compromises are evaluated Can be used with caution for short-term P versus E measurements Used for special projects, e.g., effect of UV Used for work assessing effects of UV Used for P versus E with many replicates

Required for some types of analysis, e.g., 13C

pore size 0.7 mm, are commonly used and widely (although not universally) considered to capture all sizes of phytoplankton quantitatively. Perforated filters with uniform pore sizes ranging from 0.2 to 5 mm or more can be used for size-fractionation. Particles larger than the pores can squeeze through, especially when vacuum is applied. The filters are also subject to clogging, leading to retention of small particles. Labeled dissolved organic carbon, including excreted photosynthate and cell contents released through ‘sloppy feeding’ of grazers, is not collected on filters. These losses are generally several percent of total or less, but under some conditions, excretion can be much more. When 14C samples are processed with a more cumbersome acidification and bubbling technique, both dissolved and particulate organic carbon is measured. Interpretation of Carbon Uptake

Filtration or Acidification

Generally, an incubation with 14C or 13C is terminated by filtration. Labeled particles are collected on a filter for subsequent analysis. Residual dissolved inorganic carbon can be removed by careful rinsing with filtered sea water; exposure of the filter to acid purges both dissolved inorganic carbon and precipitated carbonate. The choice of filter can influence the result. Whatman GF/F glass-fiber filters, with nominal

Because the labeled carbon is initially only in the inorganic pool, short incubations with 14C (r1 h) characterize something close to gross production. As incubations proceed, cellular pools of organic carbon are labeled, and some 14C is respired. Also, some excreted 14C organic carbon is assimilated by heterotrophic microbes, and some of the phytoplankton are consumed by grazers. So, with time, the measurement comes closer to an estimate of the net primary production of the enclosed community (Table 4).


584


Table 4

Incubation times for the measurement of primary production

Incubation time

Advantages

Disadvantages

Comments

Short (r1 h)

Little time for unnatural physiological changes

Closer to Pg

1–6 h

Convenient Appropriate for some process studies Good for standardization of methodology

Usually requires artificial illumination Uncertain extrapolation to daily rates in nature Uncertain extrapolation to daily rates in nature Limits the number of stations that can be sampled Containment effects Vertical mixing is not simulated, leading to artifacts

A good choice for standard method using in situ incubation Closer to Pnc near the surface; close to Pg deep in the photic zone A good standard for SIS incubations. Close to Pnc near the surface; closer to Pg deep in the photic zone

Dawn–dusk

24 h

Good for standardization of methodology

Results may vary depending on start time. Longer time for containment effects to act

Used for P versus E, especially with larger samples

However, many factors, including the ratio of photosynthesis to respiration, influence the degree to which 14 C uptake resembles gross versus net production. Consequently, critical interpretation of 14C primary production measurements requires reference to models of carbon flow in the system.

See also

Conclusions

Further Reading

Primary production is not like temperature, salinity or the concentration of nitrate, which can in principle be measured exactly. It is a biological process that cannot proceed unaltered when phytoplankton are removed from their natural surroundings. Artifacts are unavoidable, but many insults to the sampled plankton can be minimized through the exercise of caution and skill. Still, the observed rates will be influenced by the methods chosen for making the measurements. Interpretation is also uncertain: the 14 C method is the standard operational technique for measuring marine primary production, yet there are no generally applicable rules for relating 14C measurements to either gross or net primary production. Fortunately, uncertainties in the measurements and their interpretation, although significant, are not large enough to mask important patterns of primary productivity in nature. Years of data on marine primary production have yielded information that has been centrally important to our understanding of marine ecology and biogeochemical cycling. Clearly, measurements of marine primary production are useful and important for understanding the ocean. It is nonetheless prudent to recognize that the measurements themselves require circumspect interpretation.

Geider RJ and Osborne BA (1992) Algal Photosynthesis: The Measurement of Algal Gas Exchange. New York: Chapman and Hall. Morris I (1981) Photosynthetic products, physiological state, and phytoplankton growth. In: Platt T (ed.) Physiological Bases of Phytoplankton Ecology. Canadian Bulletin of Fisheries and Aquatic Science 210: 83–102. Peterson BJ (1980) Aquatic primary productivity and the 14C–CO2 method: a history of the productivity problem. Annual Review of Ecology and Systematics 11: 359--385. Platt T and Sathyendranath S (1993) Fundamental issues in measurement of primary production. ICES Marine Science Symposium 197: 3--8. Sakshaug E, Bricaud A, Dandonneau Y, et al. (1997) Parameters of photosynthesis: definitions, theory and interpretation of results. Journal of Plankton Research 19: 1637--1670. Steemann Nielsen E (1963) Productivity, definition and measurement. In: Hill MW (ed.) The Sea, vol. 1, pp. 129--164. New York: John Wiley. Williams PJL (1993a) Chemical and tracer methods of measuring plankton production. ICES Marine Science Symposium 197: 20--36. Williams PJL (1993b) On the definition of plankton production terms. ICES Marine Science Symposium 197: 9--19.

Carbon Cycle. Fluorometry for Biological Sensing. Network Analysis of Food Webs. Ocean Carbon System, Modeling of. Primary Production Distribution. Primary Production Processes. Tracers of Ocean Productivity.


PRIMARY PRODUCTION PROCESSES J. A. Raven, Biological Sciences, University of Dundee, Dundee, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2284–2288, & 2001, Elsevier Ltd.

Introduction This article summarizes the information available on the magnitude of and the spatial and temporal variations in, marine plankton primary productivity. The causes of these variations are discussed in terms of the biological processes involved, the organisms which bring them about, and the relationships to oceanic physics and chemistry. The discussion begins with a definition of primary production. Primary producers are organisms that rely on external energy sources such as light energy (photolithotrophs) or inorganic chemical reactions (chemolithotrophs). These organisms are further characterized by obtaining their elemental requirements from inorganic sources, e.g. carbon from inorganic carbon such as carbon dioxide and bicarbonate, nitrogen from nitrate and ammonium (and, for some, dinitrogen), and phosphate from inorganic phosphate. These organisms form the basis of food webs, supporting all organisms at higher trophic levels. While chemolithotrophy may well have had a vital role in the origin and early evolution of life, the role of chemolithotrophs in the present ocean is minor in energy and carbon terms (Table 1), but is very important in biogeochemical element cycling, for example in the conversion of ammonium to nitrate. Quantitatively a much more important process in primary productivity on a global scale is photolithotrophy (Table 1). Essentially all photolithotrophs which contribute to net inorganic carbon removal from the atmosphere or the surface ocean are O2-evolvers, using water as electron donor for carbon dioxide reduction, according to eqn [1]: CO2 þ 2H2 O þ 8 photons -ðCH2 OÞ þ H2 O þ O2

½1

The contribution of O2-evolving photolithotrophy from terrestrial environments is greater than that in the oceans, despite the sea occupying more than two-thirds of the surface of the planet (Table 1). Sunlight is attenuated by sea water to an extent which limits primary productivity to, at most, the

top 300 m of the ocean. Since only a few percent of the ocean floor is within 300 m of the surface, the role of benthic primary producers (i.e. those attached to the ocean floor) is small in terms of the total marine primary production (Table 1). Despite the relatively small area of benthic habitat for photolithotrophs in the ocean as a percentage of the total sea area (2%), benthic primary productivity producers account for almost 10% of marine primary productivity (Table 1). Until very recently it has been assumed that the photosynthetic primary producers in the marine phytoplankton are the O2-evolvers with two photochemical reactions involved in moving each electron from water to carbon dioxide, although molecular genetic data from around 1990 indicated the presence of erythrobacteria in surface ocean waters. Recent work has shown that both rhodopsin-based and bacteriochlorophyll-based phototrophy is widespread in the surface ocean. This phototrophy does not involve O2 evolution and, while it does not necessarily involve net carbon dioxide fixation, it may impact on surface ocean carbon dioxide dynamics. Thus, growth of prokaryotes using dissolved organic carbon can occur with less carbon dioxide produced per unit dissolved organic carbon incorporated into organic carbon by the use of energy from photons to replace energy that would otherwise be transformed by oxidation of dissolved organic carbon. It is probable that these phototrophs which do not evolve O2 contribute o1% to gross carbon dioxide fixation by the surface ocean. This article investigates the reasons for this constrained planktonic primary production in the ocean in terms of the marine pelagic habitat and the diversity of the organisms involved in terms of their phylogeny and life-form.

The Habitat The surface ocean absorbs solar radiation via the properties of sea water, as well as of any dissolved organic material and of particles. A very small fraction (r1% in most areas) of the 400–700 nm component is converted to energy in organic matter in photosynthesis, while the rest is converted to thermal energy. In the absence of wind shear and ocean currents, themselves ultimately caused by solar energy input, the thermal expansion of the surface water would cause permanent stratification, except near the poles in winter.


585

586

PRIMARY PRODUCTION PROCESSES

Table 1

Net primary productivity of habitats, and area of the habitats, on a world basisa

Habitat

Total area (m2)

Organisms

Global net primary productivity (1015g C year 1)

Marine phytoplankton Marine planktonic chemolithotrophs converting to and to NH4þ to NO2 and NO2 to NO3 Marine benthicb,c

370 1012 370 1012

Cyanobacteria and microalgae Bacteria

46 r0.19

6.8 1012

Marine benthicb

0.35 1012

0.34 3.4 0.35

Inland waters

2 1012

Terrestriald

150 1012

Cyanobacteria and microalgae Macroalgae Angiosperms (salt marshes plus beds of seagrasses) Phytoplankton (cyanobacteria and microalgae), benthic algae and higher plants Mainly higher plants

0.58

54

a

From Raven (1991, 1996) and Falkowski et al. (2000). All values are for photolithotrophs unless otherwise indicated. Area of the habitat the marine benthic cyanobacteria, algae, and marine benthic angiosperm is in series with that of the overlying phytoplankton habitat. The benthic habitat area is included in the habitat area for marine phytoplankton. c The marine benthic cyanobacterial and algal category includes cyanobacteria and algae symbionts with protistans and invertebrates. d As well as higher plants the terrestrial productivity involves cyanobacteria and microalgae, both lichenized and free-living, although there seem to be no estimates of the magnitude of nonhigher plant productivity. b

Such an ocean is approximated by most parts of the tropical ocean, where ocean currents and wind are inadequate to cause breakdown of thermal stratification; the upper mixed layer shows very little seasonal variation. By contrast, at higher latitudes the varying solar energy inputs throughout the year, combined with wind shear and ocean current influences, lead to stratification with a relatively shallow upper mixed layer in the (local) summer and a much deeper one in the (local) winter, usually giving a winter mixing depth so great that net primary production is not possible as a result of the inadequate mean photon flux density (light-energy) incident on the cells. A very important impact of stratification is the isolation of the upper mixed layer, where inorganic nutrients are taken up by phytoplankton, from the lower, dark, ocean where nutrients are regenerated by heterotrophy. The movement of organic particles from the upper to lower zones is gravitational. While there is significant recycling of inorganic nutrients in the upper mixed layer via primary productivity and activities of other parts of the food web, ultimately there is loss of particles containing nutrient elements across the thermocline. Seasonal variations in mixing depth, and upwellings, are the main processes bringing nutrient solutes back to the euphotic zone. Global biogeochemical cycling considerations suggest that the nutrient element that limits the extent of global primary production each year is, over long time periods, phosphorus. This element has a shorter residence time than the other nutrients (such as iron) which are supplied solely from

terrestrial sources. Nitrogen, by contrast, is present in the atmosphere and dissolved in the ocean as dinitrogen in such large quantities that any limitation of marine phytoplankton primary productivity by the availability of such universally available nitrogen formed as ammonium and nitrate could be offset by diazotrophy, i.e. biologically dinitrogen fixation, which can only be brought about by certain Archea and Bacteria. In the ocean the phytoplanktonic cyanobacteria are the predominant diazotrophs, as the free-living Trichodesmium and as symbionts such as Richia in such diatoms as Hemiaulis and Rhizosolenia. Diazotrophy needs energy (ultimately from solar radiation) and trace elements such as iron (always), molybdenum (usually), and vanadium (sometimes). These trace elements all have longer oceanic residence times than phosphorus, and so are less likely to limit primary productivity than is phosphorus over geologically significant time intervals. However, the balance of evidence for the present ocean suggests that nitrogen is a limiting resource for the rate and extent of primary productivity over much of the world ocean, while iron seems to be the limiting nutrient in the ‘high nutrient (nitrogen, phosphorus), low chlorophyll’ areas of the ocean. Even where nitrogen does appear to be limiting, this could be a result of restricted iron supply which restricts the assimilation of combined nitrogen, and especially of nitrate.

Processes at the Cell Level The photosynthetic primary producers in the marine plankton show great variability in taxonomy, and in



size and shape. The taxonomic differences reflect phylogenetic differences, including the prokaryotic bacteria (cyanobacteria, embracing the chlorophyll b-containing chloroxybacteria) and a variety of phyla (divisions) of Eukaryotes. The Eukaryotes include members of the Chlorophyta (green algae), Cryptophyta (cryptophytes), Dinophyta (dinoflagellates), Haptophyta (Phaeocystis and coccolithophorids), and Heterokontophyta (of which the diatoms or Bacillariophyceae are the most common marine representatives). The phylogenetic differences determine pigmentation, with the ubiquitous chlorophyll a accompanied by phycobilins in cyanobacteria sensu stricto, chlorophyll b in Chloroxybacteria and green algae, chlorophyll(s) c together with significant quantities of light-harvesting carotenoids in dinoflagellates, haptophytes, and diatoms, and chlorophyll c with phycobilins in cryptophytes. These differences in pigmentation alter the capacity for a given total quantity of pigment per unit volume of cells for photon absorption in a given light field, noting that the deeper a cell lives in open-ocean water, the less longer wavelength (red-orange-yellow) light is available relative to blue-green light. This effect of different light-harvesting pigments on light absorption capacity is greatest in very small cells as a result of the package effect. Another phylogenetic difference among phytoplankton organisms is a dependence on Si (in diatoms) and on large quantities of Ca (in coccolithophorids) in those algae which have mineralized skeletons. Furthermore, some vegetative cells move relative to their immediate aqueous environment using flagella (almost all planktonic dinoflagellates, some green algae). Movement relative to the surrounding water occurs in any organism which is denser than the surrounding water sinking (e.g. by many mineralized cells) or less dense than the surrounding water (buoyancy engendered by cyanobacterial gas vacuoles or the ionic content of vacuoles in large vacuolate cells). The variation in cell size among marine phytoplankton organisms is also partly related to taxonomy. The smallest marine phytoplankton cells are prokaryotic, with cells of Prochlorococcus (cyanobacteria sensu lato) as small as 0.5 mm diameter, while cells of the largest diatoms (Ethmosdiscus spp.) and the green Halosphaera are at least 1 mm in diameter, i.e. a range of volume from 6.25 1020 m3 to more than 4.19 109 m3. This means a volume of the largest cells which is almost 1011 that of the smallest; allowing for the vacuolation of the large cells gives a ratio of almost 1010 for cytoplasmic volume. The size range for phytoplankton organisms is expanded by considering

587

colonial organisms (e.g. the cyanobacterium Trichodesmium, and the haptophyte Phaeocystis) to a range of cytoplasmic volumes up to almost 1012. At the level of the cell size cyanobacteria have a limited volume range, while in organisms size the range is at least 1012; for haptophytes it is at least 1011. Cell (or organism) size is, on physicochemical principles, very important for the effectiveness of light absorption per unit pigment, nutrient uptake as a result of surface area per unit volume, and of diffusion boundary layer thickness, and rate of vertical movement relative to the surrounding water for a given difference in density between the organisms and their environment. These physicochemical predictions are, to some extent, modified by the organisms by, for example, changed pigment per unit volume and light scattering, and modulation of density.

Determinants of Primary Productivity Despite these variations in phylogenetic origin and in the size of the organisms, that can be related to seasonal and spatial variations over the world ocean, it is not easy to find consistent spatial and temporal variations in the ‘major element’ ratios (C : N : P, 106 : 16 : 1 by atoms) or Redfield ratio in space or time. This means that we should not look to differences in the phylogeny or size of phytoplankton organisms to account for differences in the requirement for major nutrients (C, N, P) in supporting primary productivity. What is less clear is the possible variations in trace element (Fe, Mn, Zn, Mo, Cu, etc.) requirements in relation to the properties of different bodies of water in the world ocean. The trace metals are essential catalysts of primary productivity through their roles in photosynthesis, respiration, nitrogen assimilation, and protection against damaging active oxygen species. Geochemical evidence for limitation by some factor other than nitrogen and phosphorus is indicated for ‘high nutrient’ (i.e. available nitrogen and phosphorus) ‘low chlorophyll’ (i.e. photosynthetic biomass and hence productivity), or HNLC, regions of the ocean (north-eastern subAntarctic Pacific; eastern tropical Pacific; Southern Ocean). Before seeking other geochemical limitations on primary production to explain why these apparently available sources of nitrogen and phosphorus have not been used in primary productivity, we need to consider geophysical or ‘bottom up’ (mixing depth, surface photon flux density) and ecological or ‘top down’ factors (involvement of grazers or pathogens). While these nongeochemical ‘bottom up’ (control of production of biomass) and ‘top down’ (removal of


588


the product of primary production) constraints on the use of nitrate and phosphate are, in principle, causes of this HNLC phenomenon, in situ Fe enrichments show that addition of this trace element causes drawdown of nitrate and phosphate, increases in chlorophyll and primary productivity, and increased abundance, and contribution to primary productivity of large diatoms. These IRONEX and SOIREE experiments strongly support the notion that iron limits primary productivity in HNLC regions, as well as suggesting that Fe enrichment can impact differentially on primary producers as a function of their taxonomy and cell size. The increased importance of large diatoms as a result of Fe enrichment can be a result of the diffusion boundary layer thicknesses and surface area per unit volume, rather than of the biochemical demand for Fe to catalyze a given rate of metabolism per unit cell volume. While data are not abundant, theoretical considerations suggest that cyanobacteria should, other things being equal, have higher requirements for Fe for growth than do diatoms, haptophytes, or green algae. This prediction contrasts with observations (and production) for major nutrients such as organic C, N, and P, where cell quotas are much less variable phylogenetically than are those for micronutrients. There is, of course, much less elasticity possible for the C content of cells than for other, less abundant, nutrient elements, with the same applying to a lesser extent to N and P. It is clear that the cost of N, P, or Fe in fixing carbon dioxide is higher for growth at low (limiting) as opposed to high (saturating) photon flux densities. To broaden the issues of limitations on primary productivity, the ultimate limitation on primary productivity in the ocean is presumably the ‘geochemical’ limiting element P, i.e. the nutrient element with the shortest residence time in the ocean. It has been plausibly argued that, in the short term, nitrogen has become a limiting nutrient indirectly by the short-term (geologically speaking) Fe limitation. Thus, ‘new’ production, depending on nitrate upwelled or eddy-diffused from the deep ocean, has a greater Fe requirement than the NHþ 4 (or organic N) assimilation in ‘recycled’ production in which primary production is chemically fuelled by N, P, and Fe generated by zooplankton, and more importantly, by Fe limitation (at least in the geological short-term) is seen as restricting diazotrophy. In the context of the balance of diazotrophy plus atmospheric and riverine inputs of combined N, and denitrification and sedimentation loss of combined N, Fe limitation can reduce the combined N availability relative to that of P to below the 16:1 atomic ration of the Redfield ratio. This, then, restricts the N:P ration in upwelled

sea water. Even more immediate Fe limitation is seen in the HNLC ocean, as discussed above.

Conclusions Marine primary production accounts for almost half of the global primary production, and is carried out by a much greater phylogenetic range of organisms than is the case for terrestrial primary production. As on land, almost all marine primary production involves O2-evolving photolithotrophs. Marine phytoplankton has a volume of cells of 6.1020– 4.109 m3. While the primary production in the oceans is, on geological grounds, ultimately limited by P, proximal (shorter-term) limitation involves N or Fe.

Glossary HNLC high nutrient, low chlorophyll. Areas of the ocean in which combined nitrogen and phosphate are present at concentrations which might be expected to give higher rates of primary production and levels of biomass, than are observed unless some ‘top down’ or ‘bottom up’ limitation is involved. IRONEX iron enrichment experiment. Two releases of FeSO4, with SF6 as a tracer, south of the Galapagos in the Eastern Equatorial Pacific HNLC area. Photon flux density Units are mol photon m 2 s 1. Means of expressing incident irradiance in terms of photons, i.e. the aspect of the particle/wave duality of electromagnetic radiation which is appropriate for consideration of photochemical reactions such as photosyntheses. For O2-evolving photosynthetic organisms the appropriate wavelength range is 400–700 nm. SOIREE southern ocean iron release experiment. A Southern Ocean analogue of IRONEX, performed between Australasia and Antarctica.

See also Anthropogenic Trace Elements in the Ocean. Carbon Cycle. Carbon Dioxide (CO2) Cycle. Nitrogen Cycle. Primary Production Distribution. Primary Production Methods.

Further Reading Falkowski PG and Raven JA (1997) Photosynthesis. Malden: Blackwell Science.


Aquatic


Falkowski PG, et al. (2000) The global carbon cycle: a test of our knowledge of Earth as a system. Science 290: 291--296. Fuhrman JA (1999) Marine viruses and their biogeochemical and ecological effects. Nature 399: 541--548. Martin JH (1991) Iron, Liebig’s Law and the greenhouse. Oceanography 4: 52--55. Platt T and Li WKW (eds.) (1986) Photosynthetic Picoplankton. Canadian Bulletin of Fisheries and Aquatic Sciences 214. Raven JA (1991) Physiology of inorganic C acquisition and implications for resource use efficiency by marine phytoplankton: relation to increased CO2 and temperature. Plant Cell Environment 14: 779--794.

589

Raven JA (1996) The role of autotrophs in global CO2 cycling. In: Lidstrom ME and Tabita FR (eds.) Microbial Growth on C1 Compounds, pp. 351--358. Dordrecht: Kluwer Academic Publishers. Raven JA (1998) Small is beautiful: the picophytoplankton. Funct. Ecol. 12: 503--513. Redfield AC (1958) The biological control of chemical factors in the environment. American Scientist 46: 205--221. Stokes T (2000) The enlightened secrets across the ocean. Trends in Plant Science 5: 461. Van den Hoek C, Mann DG, and Jahns HM (1995) Algae. An Introduction to Phycology. Cambridge: Cambridge University Press.


PROCELLARIIFORMES K. C. Hamer, University of Durham, Durham, UK Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2289–2295, & 2001, Elsevier Ltd.

Introduction The procellariiformes (commonly referred to as petrels) are a monophyletic group of seabirds containing about 100 species in four families: the albatrosses (Diomedeidae, 13 species), the shearwaters, fulmars, prions, and gadfly petrels (Procellariidae, 65 species), the storm-petrels (Hydrobatidae, 21 species), and the diving-petrels (Pelecanoididae, 4 species). It has recently been suggested that the Diomedeidae and Procellariidae may in fact contain up to 21 species and 79 species, respectively (bringing the total to 125 species), but this remains open to debate and so the more conservative taxonomy has been retained here. Petrels range in body size from the least storm-petrel Halocyptena microsoma (wing span 32 cm, body mass 20 g) to the royal albatross Diomedea epomophora (wing span 300 cm, body mass 8700 g), but they can all be identified by a single diagnostic feature: the external nares open at the end of a prominent horny tube on the upper mandible (Figure 1). They are usually considered a separate and ancient order of birds, probably derived from a late Cretaceous ancestor, although DNA evidence has suggested a more recent origin, with the petrels joining the divers (loons), frigate-birds, and penguins within the Order Ciconiiformes. Petrels occur from tropical to polar regions in all oceans, but most species breed at cool temperate latitudes, with the highest species richness at 45– 601S for albatrosses and diving-petrels, and 30–451S for the Procellariidae. However, storm-petrels have highest species richness in warm temperate waters at 15–301N (Table 1). Overall, about 70% of petrel species and probably more than 80% of individuals breed in the southern hemisphere. They generally nest on islands, headlands, or mountains isolated from mammalian predators. About 20% of petrel species are surface-nesters while the rest breed below ground in burrows or beneath boulders and scree. Surface-nesters are either too large to burrow or are smaller, mainly tropical, species without predators at the colony. Burrow-nesters usually return to land only by night, probably as a defense against avian predators. Petrels nest colonially, often in high

590

numbers, with colonies of several species exceeding 2 million breeding pairs. Many species undertake regular annual migrations between breeding colonies and separate locations used outside the breeding season. These often involve transequatorial movements, although some species change latitude without crossing the equator and others move east or west without much change in latitude.

Population Ecology Demography

Petrels are very long-lived for their size, with some storm-petrels reaching 20 years of age and some albatrosses reaching 80 years. Sexual maturation occurs at 2–13 years of age (later in larger species; Figure 2) and all species have only a single-egg clutch. Breeding success is typically about 0.4–0.7 chicks fledged per nest, increasing with age and experience. Most breeding failures occur during incubation and early chick-rearing (especially at hatching), with a secondary peak in losses occurring around fledging in some species, mainly due to naive birds being captured by predators. Most petrels are annual breeders but some albatrosses, which have very long breeding seasons, are biennial, while female Audubon’s shearwaters Puffinus lherminieri at the Galapagos Islands lay at roughly 9-month intervals. In many species a substantial proportion of breeders (up to 30% in some cases) misses a year occasionally, probably to replenish body reserves depleted during the previous breeding season. Because petrels are long-lived and have low annual reproductive rates, their populations are more sensitive to changes in adult survival than to changes in juvenile survival or breeding success. For instance, in wandering albatrosses Diomedea exulans, a reduction of 1% in adult survival would require an increase of 6% in juvenile survival to prevent a decline in population size. Regulation of Population Size

Under normal circumstances, some petrel populations may be regulated in a density-dependent manner by intraspecific competition for food or breeding space. Although there is little direct evidence of prey depletion, the large size of many petrel colonies implies a strong potential for foraging success to be reduced by competition for food, and the notion of competitive exclusion is supported by


PROCELLARIIFORMES

591

Figure 1 Examples of Procellariiform Seabirds. (1) Wandering albatross (Diomedea exulans). Adult. Length: 115 cm; wingspan: 300 cm; approximate body mass: 8750 g. Range: Most ocean areas south of 30 S. (2) Mottled petrel (Pterodroma inexpectata). Other names: scaled petrel, Peale’s petrel. Length: 34 cm; wingspan: 74 cm; approximate body mass: 330 g. Range: Pacific Ocean. (3) Sooty shearwater (Puffinus griseus). Length: 44 cm; wingspan: 99 cm; approximate body mass: 780 g. Range: North and South Pacific Ocean, North and South Atlantic Ocean, Southern Ocean. (4) Common diving-petrel (Pelicanoides urinatrix). Other names: subantarctic diving-petrel. Length: 22 cm; wingspan: 35 cm; approximate body mass: 135 g. Range: Southern oceans. (5) Cape petrel (Daption capense). Other names: Cape pigeon, pintado petrel. Length: 39 cm; wingspan: 86 cm; approximate body mass: 450 g. Range: most ocean areas south of 30 S. (6) Wilson’s storm-petrel (Oceanites oceanicus). Length: 17 cm; wingspan: 40 cm; approximate body mass: 35 g. Range: North and South Atlantic Ocean, Southern Ocean.

segregation of feeding zones between sexes, age classes, and populations of some species, particularly southern albatrosses. However, the large foraging ranges of petrels and their wide dispersal at sea

reduce the likely importance of prey depletion in most cases, and direct interference competition is unlikely to have a major influence on foraging success except under particular circumstances such as


592

PROCELLARIIFORMES

Table 1

Species richness of breeding petrels by latitude

1825 1850

1900

1875

1925 1950 1975

Number of species Latitude (deg)

Diom

Procel

Hydro

Pelec

Total

75.1–90N 60.1–75N 45.1–60N 30.1–45N 15.1–30N 0–15N 0–15S 15.1–30S 30.1–45S 45.1–60S 60.1–75S 75.1–90S

0 0 0 2 2 0 1 0 6 9 0 0

1 2 2 10 15 3 9 17 32 21 6 2

0 2 3 7 9 1 4 5 4 4 2 1

1 0 0 0 0 0 1 1 2 3 0 0

1 4 5 20 27 4 15 23 44 37 8 3

Taxonomy follows Warham (1990) except that Levantine shearwater Puffinus yelkouan (previously regarded as a Mediterranean subspecies of P. puffinus) is given full specific status, raising the number of Procellariidae breeding at 30.1–451N from 9 to 10. Diom ¼ Diomedeidae, Procel ¼ Procellariidae, Hydro ¼ Hydrobatidae, Pelec ¼ Pelecanoididae. 14

1800

Iceland

1800 1825

Faroe

1850

Shetland

1875

Norway

1900

Ireland 1925

Britain

1950

1975

France

Mean age at first breeding (years)

12

10

Figure 3 The spread of northern fulmars in the eastern Atlantic Ocean. Contours are at 25-year intervals. The triangles show the position of the two preexpansion colonies at Grimsey, north of Iceland and St. Kilda, west of Scotland. (Redrawn from Williamson (1996) with permission.)

8

6

4

Changes in Distribution 2

0 1

2

3 log [body mass (g)]

4

Figure 2 The relationship between body mass and mean age at first breeding in 22 species of petrel. (Data from Table 15.4 in Warham (1990) and Table 1.3 in Warham (1996).)

when scavenging behind fishing vessels. Shortage of space may be an important regulatory factor for some species, especially burrow-nesters, when competition for nests may delay the age of first breeding and hence reduce recruitment. Predation at the colony seems to have little impact on petrel populations under natural conditions, although introduced alien predators can have large impacts (see Conservation below).

Most petrel species have fairly stable geographical distributions. However, one major exception to this is the northern fulmar Fulmarus glacialis, which has undergone a massive range expansion in the northeast Atlantic Ocean during the past 150 years (Figure 3). Historically, its breeding distribution was restricted to Arctic regions, with only two fulmar colonies in the temperate North Atlantic, one on the island of Grimsey about 40 km north of Iceland, the other at St Kilda, about 50 km west of NW Scotland. Fulmars spread to the Westman Islands off the south coast of Iceland sometime between 1713 and 1757, then to Faroe and Britain, colonizing the island of Foula, Shetland, in 1878. By 1975 their range included France, Germany (Helgoland), and Norway, and there are now colonies in Greenland, Labrador, and Newfoundland. The most likely cause of this spread was the appearance of a novel food supply in


PROCELLARIIFORMES

the form of offal and discards, first from Arctic whalers then from industrial and commercial trawlers. In keeping with this view, fulmars at high Arctic colonies appear less reliant on fishery discards than individuals at more southerly colonies. However, the latter also consume many other types of prey which they catch for themselves. These include fast-growing planktonivorous fish such as sand eels Ammodytidae and sprat Sprattus sprattus, which increased during the second half of the twentieth century following overfishing of predatory fish such as herring Clupea harengus and mackerel Scomber scombrus. Commercial fishing may thus have benefited fulmars in recent years almost as much by causing these dramatic changes in marine ecosystem structure as by the direct provision of discards. Another species currently expanding its geographical range is the Laysan albatross Diomedea immutabilis. Following increases in populations on small islands in Hawaii, this species has established (or possibly reestablished) colonies on the main Hawaiian Islands since 1970 and has recently founded a number of new colonies in western Mexico. Two species undergoing lesser range expansions are the Manx shearwater Puffinus puffinus, which spread from NW Europe to establish a breeding colony in Newfoundland during the 1970s, and the shy albatross D. cauta which recently colonized the Crozet Islands, far from its nearest colony in the Bass Strait, Australia.

Food and Foraging With the exception of diving-petrels, which forage close inshore, all petrels are pelagic feeders. During the breeding season, some albatrosses may travel up to 3000 km or more on a single foraging trip lasting several days, while trips by northern fulmars may be as short as 6 hours. Most species exploit a variety of epipelagic fish, squid, and crustacea, with a greater reliance on fish and squid by most albatrosses, shearwaters, fulmars, and gadfly petrels (most of the latter relying mainly on squid), and a greater reliance on crustacea by storm petrels, diving-petrels, and prions. In the Southern Ocean, a number of petrels prey extensively on the swarming crustacea Euphausia superba and E. chrystallorophias (krill). Most petrels exploit carrion when available, especially discards from industrial and commercial fishing vessels. Some species, notably northern fulmars, also include other prey such as polychaete worms and medusae in their diet at some colonies. Diets can differ markedly between breeding and nonbreeding seasons, especially in migratory species. For instance,

593

short-tailed shearwaters Puffinus tenuirostris consume a mixture of krill, fish, and squid during the breeding season in the Southern Ocean, but during the nonbreeding season in Alaska they prey almost exclusively on the euphausiid crustacean Thysanoessa raschii. In most petrels the diets of males and females are similar. However, in giant petrels, Macronectes giganteus and M. halli, males primarily scavenge for carrion from penguin and seal colonies while females feed primarily at sea. About 80% of petrel species obtain food by seizing it from the surface of the water, usually while in flight but sometimes while swimming. Other species dive below the surface and some pursue prey underwater, sometimes at great depth (up to 71 m in shorttailed shearwaters). Feeding often occurs in association with dolphins, tuna (Thunnus spp.), and other marine predators that force prey toward the surface. Some species also forage extensively at tidal fronts, or at polynyas in polar regions, where prey tend to aggregate. Stomach Oil

With the exception of diving-petrels, most petrels alter the chemical composition of captured prey during transport to the nest by differential retention of the aqueous and lipid fractions of the digesta within the proventriculus: following liquefaction of the food, the denser aqueous fraction passes into the duodenum first, leaving a lipid-rich liquid termed stomach oil in the proventriculus. Stomach oil has a much higher caloric density than the prey (up to 30 times higher in some cases) and may be essential in some species to allow adults to carry enough energy back to the chick. The extent to which adults form stomach oil increases with foraging range and trip duration.

Breeding Ecology Petrels are socially monogamous and form longlasting pair bonds, although they do switch partners on occasion, particularly after an unsuccessful breeding attempt. Pairing usually takes place at the colony and involves elaborate visual displays in diurnal species or auditory displays in nocturnal species. Young prebreeding birds return to the colony progressively earlier each year to display, pair off, and establish a breeding territory. This process takes several years and initial breeding attempts mostly fail. In established breeders, the period between arrival at the colony each year and egg-laying is generally 30–40% of the whole reproductive cycle, except in


PROCELLARIIFORMES

albatrosses where it comprises 7–14% of the total. Sedentary species, which occur in all families except the Diomedeidae, may also visit the colony outside the breeding period. Precopulatory behavior may be complex (as in albatrosses) or simple (as in diving petrels) but generally involves mutual preening of the head, nape, cheek, and throat. Copulation takes place on the ground, usually at the nest. Individuals may copulate many times and females sometimes solicit or accept copulations from males other than their partner, although these seldom result in extra-pair paternity. For instance, in a study of northern fulmars, females copulated up to 54 times with their male partner and up to 17 times with an extra-pair male. The final copulation was always with the pair male and there was no evidence of any extra-pair paternity. In all petrels, after copulating there is a period of several days to several weeks, termed the prelaying exodus, when some or all of the established breeders return to sea during egg formation. This may allow birds access to highly productive waters distant from the colony. Females stay away longer than males and in some species only the females leave. After returning to the colony, the female almost immediately lays a single egg weighing 5–30% of her own body mass (heavier than in other birds that lay a single-egg clutch) with a relatively large yolk. The embryo within the egg grows slowly, resulting in a long incubation period (about 39–79 days depending largely on body mass). The male is responsible for the first long incubation shift and thereafter the two parents alternate between incubating the egg and foraging at sea, at average intervals of about 1–20 days (generally longer in larger species but shortest in diving-petrels and longest in some gadfly petrels and albatrosses of intermediate body mass; Figure 4). The two sexes usually spend about the same amount of time incubating, although in some species the male spends longer. In both cases, incubation shifts by both sexes shorten toward the end of incubation, which increases the probability that the incubating bird has food for the chick when it hatches. If for whatever reason the foraging bird does not return sufficiently quickly, the incubating bird may head to sea before its partner returns, leaving the egg unattended. Such eggs are vulnerable to predators, especially in surface-nesters, but they can remain viable for many days without further incubation (up to 23 days in Madeiran storm-petrel Oceanodroma castro at the Galapagos Islands). This resistance to adverse effects of chilling may be a major adaptation of petrels to pelagic foraging, permitting them to nest far from their main food sources. However, cooling of the embryo does retard growth, resulting in a delay in hatching.

20

Mean duration of incubation shifts (days)

594

16

12

8

4

0 1

2

3 log [body mass (g)]

4

Figure 4 The relationship between body mass and incubation shift length in 32 species of petrel. (Data from Table 14.1 and 15.4 in Warham (1990).)

Like embryos before hatching, petrel chicks grow slowly for their size. This results in petrels having very long breeding seasons, with two extremes for the period from egg-laying to chick-fledging being 125 days in snow petrels Pagodroma nivea and 380 days in wandering albatrosses. Petrel chicks also accumulate very large quantities of nonstructural body fat during posthatching development (up to 30% of body mass in northern fulmars). This results in chicks attaining peak body masses far in excess of adult body mass (up to 170% of adult mass in yellow-nosed albatross Diomedea chlororhynchos) before losing mass prior to fledging. Newly hatched petrel chicks are unable to regulate their own body temperatures and they are brooded more or less continuously for the first few days after hatching. Small chicks also have limited gut capacity and so, during this period, they are fed small meals several times a day by the attending parent. Once chicks gain thermal independence, they are left unattended for most of the time while both parents forage simultaneously at sea. This change in parental attendance is accompanied by an increase in meal size and a decrease in feeding frequency. For most of the nestling period, parents deliver meals weighing 5–35% of adult body mass (proportionately smaller in heavier species) to the chick. Overall feeding frequency, resulting from provisioning by both parents, is generally between one meal every two days and two meals per day (higher in fulmars and diving petrels than in other species). Feeding frequency then


PROCELLARIIFORMES

Energy provisioning/requirement/accumulation (kJ per day)

declines at the end of the nestling period, so that in most cases total food delivery is insufficient to meet the chick’s nutritional requirements (Figure 5). Chicks thus enter negative energy balance and lose mass as they presumably deplete their fat stores. However, northern fulmar chicks continue to accumulate fat until fledging, and their prefledging drop in body mass is due entirely to the loss of water from embryonic tissues as they attain full size and functional maturity. Grey-headed albatrosses Diomedea chrysostoma show a similar pattern, and further data are required to determine how prefledging declines in body mass relate to changes in body fat in other species. Large fat stores were until recently thought to provide chicks with an energy reserve to tide them over long intervals between feeds. However, in many species, the intervals between feeds are too short to account for the quantities of fat stored, and the combined cumulative effects of variability in both feeding frequency and meal size are probably more important than long intervals per se. Fat stores may

0

0

20 40 60 80 Percent of nestling period completed

100

Figure 5 A generalized scheme for the relationship between food provisioning and growth of petrel chicks. Solid circles indicate mean energy provisioning rate (more or less independent of chick age except at the end of the nestling period): open circles indicate mean daily energy requirement (determined by body mass and structural (lipid-free) growth rate, and reaching a maximum after about 75% of the nestling period): solid squares indicate net rate of accumulation of energy within the body (about 70% of the difference between provisioning rate and requirement when growth is positive; 4100% of the difference when growth is negative). N.b. In some species, despite a drop in energy provisioning rate at the end of the nestling period, supply does not fall below energy requirement, and so chicks continue to accumulate energy, principally as body fat, until fledging.

595

also be important to chicks after fledging, while they learn to forage for themselves, and this view is supported by the fact that northern fulmars have larger fat stores at fledging than at any previous time. In addition to body fat, petrel chicks also store large quantities of stomach oil, which acts as a convenient energy store and is metabolically more efficient than converting the oil into body fat. Reproductive Effort

Petrels are long-lived and so they need to balance the requirements of current and future reproduction in order to maximize their lifetime reproductive output. During chick-rearing, they need to judge how much energy to invest in the chick without impairing their own ability to breed again. In species that make long foraging trips and feed their chick comparatively infrequently, this seems to be achieved by adults delivering food according to an intrinsic rhythm that is sensitive to changes in the parent’s body condition but not to changes in the chick’s nutritional requirements. Other species that make shorter trips on average appear to monitor the chick’s nutritional status and will reduce subsequent food delivery to a well-fed chick. However, experimental evidence indicates that they either cannot or do not increase food delivery to a chick in poor condition. Moreover, experimental manipulations that increase the work that adults need to perform to obtain food generally result in a reduction in the rate of food provisioning of the chick rather than a decline in the parent’s nutritional reserves. In some species, adults also adopt a dual foraging strategy, in which comparatively short trips undertaken to obtain food for the chick are interspersed with less frequent long trips on which adults obtain food for themselves. The switch between short and long trips seems to be determined by a threshold body condition, below which adults forage only for themselves. In some cases the areas of ocean visited on short and long trips are far apart. For instance, short-tailed shearwaters breeding in SE Australia forage comparatively close to the colony to obtain food for their chick but obtain food for themselves in the highly productive waters of the Polar Frontal Zone at least 1000 km away.

Conservation Petrels are adversely affected by a wide range of factors related to human activities, including ingestion of plastic particles and exposure to a wide variety of marine pollutants, but, historically, introduction of alien mammals to remote breeding colonies has been perhaps the greatest threat to petrel


596

PROCELLARIIFORMES

populations, with the greatest impacts being due to introductions of cats Felis catus and rats Rattus spp. Small species of petrels have been particularly affected, but even large species such as albatrosses have not been immune. Eradication programs have been developed to remove introduced mammals from a number of islands where petrels breed, especially in the southern hemisphere, and these have met with some success. For instance Marion Island, in the subAntarctic Indian ocean, was cat-free prior to 1949, when five cats were introduced as pets. By 1975 the cat population had risen to about 2000 individuals consuming 450 000 burrowing petrels annually. An eradication program was started in 1977, when the cat population was over 3000 individuals, and by 1993 there were no signs of live cats on the island. The elimination of cats resulted in a marked increase in the breeding success of several species of petrel, although a combination of small populations and delayed maturity has resulted in only slow population recoveries. It should also be noted that on islands that have been colonized by more than one species of predator, eradication of one species could lead to a disproportionate increase in other species, to the detriment of petrels. This is likely to be a particular problem with eradication of cats from islands that have also been colonized by rats. One of the major current threats to petrel populations, especially in the southern hemisphere, is longline fishing using baited hooks to catch demersal and pelagic fish, particularly tuna and Patagonian toothfish Dissostichus eleginoides. Petrels can be drowned after striking at baited hooks and, while the rate of by-catch is generally very low, the large number of hooks set means that total mortality can be sufficient to reduce population sizes of some species. For instance, the Japanese pelagic longline fishery for southern bluefin tuna T. maccoyi deployed up to 100 million hooks or more annually during the 1990s. In Australian waters, the average rate of sea bird bycatch from this fishery was 1.5 birds per 10 000 hooks set, but this resulted in an annual mortality of up to 3500 birds per year. Longline fishing operations have been implicated in serious population declines of a

number of albatrosses, including wandering albatross, grey-headed albatross, black-browed albatross Diomedea melanophrys, yellow-nosed albatross, and light-mantled sooty albatross Phoebetria palpebrata. Some populations have declined by as much as 90% and are facing imminent extinction if current catch rates are not reduced.

See also Ecosystem Effects of Fishing. Fisheries: Multispecies Dynamics. Fishery Management, Human Dimension. Fishing Methods and Fishing Fleets. Krill. Network Analysis of Food Webs. Seabird Foraging Ecology. Seabird Migration. Seabird Population Dynamics. Seabird Reproductive Ecology. Seabirds and Fisheries Interactions.

Further Reading Brooke M (1990) The Manx Shearwater. London: Academic Press. Croxall JP and Rothery P (1991) Population regulation of seabirds: implications of their demography for conservation. In: Perrins CM, Lebreton J-D, and Hirons GJM (eds.) Bird Population Studies. Relevance to Conservation and Management, pp. 272--296. Oxford: Oxford University Press. Fisher J (1952) The Fulmar. London: Collins. Robertson G and Gales R (eds.) (1998) Albatross Biology and Conservation. Chipping Norton, Australia: Surrey, Beatty & Sons. Tickell WLN (2000) Albatrosses. Yale: Yale University Press. Warham J (1990) The Petrels. Their Ecology and Breeding Systems. London: Academic Press. Warham J (1996) The Behaviour, Population Biology and Physiology of the Petrels. London: Academic Press. Weimerskirch H and Cherel Y (1998) Feeding ecology of short-tailed shearwaters: breeding in Tasmania and foraging in the Antarctic? Marine Ecology Progress Series 167: 261--274. Williamson M (1996) Biological Invasions. London: Chapman and Hall.


PROPAGATING RIFTS AND MICROPLATES Richard Hey, University of Hawaii at Manoa, Honolulu, HI, USA Copyright & 2001 Elsevier Ltd.

have similar magnitudes to local spreading rates. Figure 1 shows several variations of typical midocean ridge propagation geometry, in which a preexisting ‘doomed rift’ is replaced by the propagator.


Geometry

Introduction Propagating rifts appear to be the primary mechanism by which Earth’s accretional plate boundary geometry is reorganized. Many propagation episodes are caused by or accompany changes in direction of seafloor spreading. Propagating rifts, oriented at a more favorable angle to the new plate motion, gradually break through lithospheric plates. A propagator generally replaces a pre-existing spreading center, causing a sequence of spreading center jumps and leaving a failed rift system in its wake. This results in changes to the classic plate tectonic geometry. There is pervasive shear deformation in the overlap zone between the propagating and failing rifts, much of it accommodated by bookshelf faulting. Rigid plate tectonics breaks down in this zone. When the scale or strength of the overlap zone becomes large enough, it can stop deforming, and instead begin to rotate as a separate microplate between dual active spreading centers. This microplate tectonic behavior generally continues for several million years, until one of the spreading boundaries fails and the microplate is welded to one of the bounding major plates. Active microplates are thus modern analogs for how large-scale (hundreds of kilometers) spreading center jumps occur.

Propagating Rifts Propagating rifts are extensional plate boundaries that progressively break through mostly rigid lithosphere, transferring lithosphere from one plate to another. If the rifting advances to the seafloor spreading stage, propagating seafloor spreading centers follow, gradually extending through the rifted lithosphere. The orthogonal combination of seafloor spreading and propagation produces a characteristic V-shaped wedge of lithosphere formed at the propagating spreading center, with progressively younger and longer isochrons abutting the ‘pseudofaults’ that bound this wedge. Although propagation rates as high as 1000 km per million years have been discovered, propagation rates often

Figure 1A shows the discontinuous propagation model, in which periods of seafloor spreading alternate with periods of instantaneous propagation, producing en echelon failed rift segments, fossil transform faults and fracture zones, and blocks of progressively younger transferred lithosphere. Figure 1B shows the pattern produced if propagation, rift failure, and lithospheric transferral are all continuous. In this idealized model a transform fault migrates continuously with the propagator tip, never existing in one place long enough to form a fracture zone, and thus V-shaped pseudofaults are formed instead of fracture zones. Figure 1C shows a geologically more plausible model, in which the spreading rate accelerates from zero to the full rate over some finite time and distance on the propagating spreading center with concomitant decreases on the failing spreading center, so that lithospheric transferral is not instantaneous. Instead of a transform fault, a migrating broad ‘non-transform’ zone of distributed shear connects the overlapping propagating and failing ridges during the period of transitional spreading. Deformation occurring in this overlap zone is preserved in the zone of transferred lithosphere, cross-hatched in Figure 1. This zone is bounded by the failed rifts and inner (proximal) pseudofault. Even more complicated geometries occur in some places on Earth where the doomed rift, instead of failing monotonically as the propagator steadily advances, occasionally itself propagates in the opposite direction. In these ‘dueling propagator’ systems, both axes curve toward each other. Even in simple propagator systems, the failing rift often curves toward the propagating rift, resembling the smaller scale overlapping spreading center ‘69’ geometry (see Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology), but with about a 1:1 overlap length to width aspect ratio, in contrast to the 3:1 aspect ratio characteristic of overlappers. Classic plate tectonic geometry holds for the area outside the pseudofaults and zone of transferred lithosphere, but rigid plate tectonics breaks down in the overlap zone where some of the lithosphere formed on the doomed rift is progressively


597

598

PROPAGATING RIFTS AND MICROPLATES

Pseudofault Active transform fault Propagating rift

Fossil transform fault

Pseudofault

Doomed rift

Failed rift

(A) Pseudofault

Propagating rift

Instantaneous transform fault

Doomed rift Pseudofault Failed rift (B)

Pseudofault Tip of propagating rift Propagating rift

Overlap zone

Doomed rift Pseudofault

Failing rift

(C)

Figure 1 (A) Discontinuous, (B) continuous, and (C) non-transform zone oceanic propagating/failing rift models. Propagating rift lithosphere is marked by dark stipple, normal lithosphere created at the doomed rift is indicated by light stipple, and transferred lithosphere is cross-hatched. Heavy lines show active plate boundaries. In (C), active axes with full spreading rate are shown as heavy lines; active axes with transitional rates are shown as dashed lines. The overlap zone joins these transitional spreading axes. (Reproduced with permission from Hey et al., 1989.)

transferred to the other plate by the rift propagation and resulting migration of the overlap zone. Shear between the overlapping propagating and failing rifts appears to be accommodated by bookshelf faulting, in which, for example, right-lateral plate motion shear produces high angle left-lateral slip, apparently along the pre-existing abyssal hill faults. This

produces oblique seafloor fabric, with trends quite different from the ridge-parallel and -perpendicular structures expected on the basis of previous plate tectonic theory. Figure 2 is a shaded relief map of the type example propagating rift, at 95.51W along the Cocos-Nazca spreading center. This propagator is breaking westward away from the Galapagos hot



599

2° 40' N

2° 20'

95°40'W

95°20'

Figure 2 Shaded relief map of digital seabeam swath bathymetry at the Galapagos 95.51W propagating rift system. The relative plate motion is nearly north–south. Propagation is to the west. The oblique structures in the overlap zone and its wake, the zone of transferred lithosphere, are clearly evident. PR, propagating rift; PSC, propagating spreading center; OPF, IPF, outer and inner pseudofaults; OZ, overlap zone; ZTL, zone of transferred lithosphere; DR, doomed rift; F’R, failing rift; FR, failed rift grabens. (Adapted with permission from Hey et al., 1989.)

spot through 1 million-year-old Cocos lithosphere at a velocity of about 50 km per million years. Well organized seafloor spreading begins about 10 km behind the faulting, fissuring, and extension at the propagating rift tip. This 200 000 year time lag between initial rifting and the rise of asthenosphere through the lithospheric crack to form a steady-state spreading center suggests an asthenospheric viscosity of about 1018 Pa-s. The combination of seafloor spreading at about 60 km per million years and propagation produces a V-shaped wedge of young lithosphere surrounded by pre-existing lithosphere. The propagating rift lithosphere is characterized by unusually high amplitude magnetic anomalies and by unusual petrologic diversity, including highly fractionated ferrobasalts. This propagator is replacing a pre existing spreading system about 25 km to the

south, and thus spreading center jumps and failed rifts are being produced. Although propagation is continuous, segmented failed rift grabens seem to form episodically on a timescale of about 200 000 years. This has produced a very systematic pattern of spreading center jumps, in which each jump was younger and slightly longer than the preceding jump. The spreading center orientation is being changed clockwise by about 131, and more than 104 km3 per million years of Cocos lithosphere is being transferred to the Nazca plate. The active propagating and failing rift axes overlap by about 20 km, and are connected by a broad and anomalously deep zone of distributed shear deformation rather than by a classic transform fault. Most of the seismic activity occurs within this ‘non-transform’ zone, where the preexisting abyssal hill fabric originally created on the


600


doomed rift is sheared and tectonically rotated into new oblique trends. Simple equations accurately describe this geometry in terms of ratios of propagation and spreading rates, together with the observed propagating and doomed rift azimuths. For example, for the simplest continuous propagation geometry, if u is the spreading half rate and v is the propagation velocity, the pseudofaults form angles tan 1 (u/v) with the propagator axis, and the isochrons and abyssal hill fabric in the zone of transferred lithosphere have been rotated by an angle tan 1 (2u/v).

10

8

6 MgO (wt. %) 4 TiO2 (wt. %) 2

Thermal and Mechanical Consequences

The boundaries of the Galapagos high amplitude magnetic anomaly zone, of the ferrobasalt province, and of the spreading center jumps identified from magnetic anomalies, are all essentially coincident with the pseudofaults bounding the propagating rift lithosphere. All of these observations can be explained as mechanical and/or thermal consequences of a new rift and spreading center breaking through cold lithosphere, with increased viscous head loss and diminished magma supply on the propagating spreading center close to the propagator tip. This leads to an unusually deep axial graben, unusually extensive fractional crystallization, and unusually high petrologic diversity. Basalt glasses erupted along propagating rifts are generally more differentiated than those erupted on normal and failing rifts (Figure 3). Furthermore, the Galapagos 95.51W propagating rift lavas display a striking and unusual compositional diversity, in marked contrast to the narrow compositional range of basalt glasses from the adjacent doomed rift segment. With extremely rare exceptions, ferrobasalts (or FeTi basalts, enriched in iron and titanium) are confined to the wedge of propagating rift lithosphere. The compositional variation of the propagator lavas has been shown to be primarily attributable to shallow-level crystal fractionation of normal midocean ridge basalt parental magmas (MORB, see Mid-Ocean Ridge Geochemistry and Petrology), controlled by a balance between the magma cooling and supply rates. The observed variation in erupted lava compositions reflects the evolution of the rift from initial spreading toward the steady-state buffered magma chamber configuration of a normal spreading center. Lava compositions define gradients in mantle source geochemistry roughly centered about the Galapagos hot spot near 911W. These data indicate that the 95.51W propagator tip represents a boundary in mantle source geochemistry, implying that this rift propagation can be related to plumerelated subaxial asthenospheric flow away from the

1

150

Zr (ppm)

100 50

2000

Bathymetry (m)

3000 100 km W

50

50 PR Tip

100 E Propagation direction

Figure 3 Relationship between lava compositions and distance from the 95.51W propagating rift tip at the time of eruption. Bathymetric profile shows depth variation along the present-day spreading axis. (Reproduced with permission from Hey et al., 1989.)

hot spot. All six known active Galapagos propagators are propagating away from the Galapagos hot spot. Causes of Rift Propagation

One important observation is that many rifts and spreading centers propagate down topographic gradients away from hot spots or shallow ridge axis topography. Rift propagation in the Galapagos area is associated with and probably caused by plumerelated asthenospheric flow under and away from the Galapagos hot spot. This flow generates gravitational stresses on the shallow spreading center segments near the hot spot that promote crack propagation away from the hot spot. Flow of asthenosphere into these cracks produces new



lithosphere at propagating seafloor spreading centers. Regionally high deviatoric tensile stresses associated with regional uplift provide a quantitatively plausible driving mechanism. Crack growth occurs when the stress concentration at the tip, characterized in elastic fracture mechanics by a stress intensity factor, exceeds some threshold value that is a function of the strength of the lithosphere. Propagation can be driven by excess gravity sliding stresses due to the shallow lithosphere near the hot spot, which is almost, but not quite, counterbalanced by the resisting stress intensity contribution due to the local tip depression. The spreading center propagation rate is limited by the viscosity of the asthenosphere flowing into the rift. The overlap/offset ratio of propagating and failing rifts tends to be B1, close to the ratio at which the stress intensity factor is maximized (see Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology). Although many rifts appear to propagate in response to hot spot-related stresses, others appear to propagate because of stresses produced when changes in plate motion occur. A common mechanism in the Northeast Pacific producing such changes and propagation appears to be subduction-related stresses. This probably explains most of the massive reorganizations of the spreading geometry as the Pacific-Farallon ridge neared the Farallon-North America trench, although some propagation away from a hot spot has also occurred in the Juan de Fuca area. Propagation may be produced in many ways over a wide range of scales. Discussion

Propagating rifts form new extensional plate boundaries and rearrange the geometries of old ones, with the numerous consequences already discussed. These include the formation of anomalous structural, bathymetric, and petrologic provinces; the formation of pseudofaults instead of fracture zones; and the transfer of lithosphere from one plate to another, as well as spreading center reorientations, jumps, and failed rifts. Propagators that replace pre-existing spreading centers always seem to result in at least slight spreading center reorientation. Whether rifts propagate in response to changes in direction of seafloor spreading, or the spreading direction changes while the rift propagates, rift propagation is the primary mechanism by which many seafloor spreading systems have adjusted to changes in spreading direction. Spreading center jumps are a predictable consequence of rift propagation if there is a failing rift. The largest jumps sometimes involve transient

601

microplate formation, geometrically similar to the broad overlap zone model of Figure 1C but on a much larger scale. Rift Propagation on the Other Scales

The pervasively deformed zone of transferred lithosphere is characteristic of propagating rift systems in which the offset of the axes ranges from a few kilometers to more than 100 km. Below this offset width, small propagators are sometimes called migrating overlapping spreading centers (see Mid-Ocean Ridge Tectonics, Volcanism, and Geomorphology), the distinction being that this smaller scale reorganization takes place completely within the nonrigid plate boundary zone, before rigid plate tectonics begins, whereas propagators break through rigid lithosphere, thus reorganizing the plate boundary geometry. Sometimes the overlap zones between propagating and failing rifts are on very large scales, e.g., the 120 120 km pervasively deformed overlap zone between the giant dueling propagators southwest of Easter Island (Figure 4). Interestingly, both the Easter microplate just to the north and Juan Fernandez microplate just to the south of this area show evidence for pervasively deformed cores. This is interpreted as evidence for a two-stage evolution, with the deformed core formed during an early rift propagation stage, before later evolution as rapid rotation of a mostly rigid microplate about a pole near the center of the microplate. This suggests that largescale propagation or duelling propagation could sometimes initiate microplate formation. If an overlap zone becomes too big and strong to deform by pervasive bookshelf faulting, it could instead accommodate the boundary plate motion shear stresses by beginning to rotate as a separate microplate.

Microplates Microplates are small, mostly rigid areas of lithosphere, located at major plate boundaries but rotating as more or less independent plates. Although small microplates can form in many tectonic settings, and are common in broad continental deformation zones, this article addresses only the type formed along mid-ocean ridges. The two main subtypes – those formed at triple junctions and those formed along ridges away from triple junctions – share many similarities, and plate boundary reorganization by rift propagation is important in both settings. Although it was once thought that stable growing microplates could eventually grow into major oceanic plates, it now appears that these are transient


602


120˚W

108˚W 3

3A

2A 3A

2 J

2A

2 J

EPR

22˚S J

2

2A

2A

2

Nazca

Pito deep

3

2A 2

EIPF

J

East rift

3 2A 2

2 2A

J

West rift

PF EO

J

F IP W

FZ WOPF

Pacific

Central anomaly

Easter Island

J

Anomaly 3

3 3A 3

28˚S

J

2

2A

2

SA

FZ

2A

NAZ PAC

Magnetic isochrons

2

EMP

3

2A

3A

29˚S Duelling Propagators

EMP EPR

JFMP

J

2

3

2A

3A

2 J

ANT

2A

3

2A 2A

FZ

J

2 2 J

WIPF

36˚S

2

PF

PT 2A

J

J 2 2A

2A

3

2

4A 4 3A

East ridge

JFMP EIPF

WOPF 5

J

EO

Complex structures at JFMP-ANT boundary N 100 km

3

West rid g

?

Non-transform offset trace or overlapping spreading center trace Compressional ridge

Endeavor deep

e

Spreading ridge and pseudofault trace of rift propagation

3

2A 2 J J

116˚W

J

2A

3

Chile

3A

Tr.

Antarctic 106˚W

Figure 4 Tectonic boundaries, magnetic isochrons, and structures of Easter (EMP) and Juan Fernandez (JFMP) microplates. EPR, East Pacific Rise; FZ, fracture zone; WOPF, WIPF, EIPF, and EOPF are western outer and inner, and eastern inner and outer, pseudofaults, respectively. (Reproduced with permission from Bird and Naar, 1994.)

phenomena resulting from rift propagation on such a large scale that the overlap zone eventually begins to rotate between dual active spreading centers. The best studied oceanic microplates are the Easter microplate along the Pacific-Nazca ridge, and the Juan Fernandez microplate at the Pacific-NazcaAntarctica triple junction. Despite their different tectonic settings, they show many striking similarities (Figure 4). Geometry

The scales of the Easter (B500 km diameter) and the Juan Fernandez (B400 km diameter) microplates are similar. The eastern and western boundaries of both microplates are active spreading centers, propagating north and south respectively. Both microplates began

forming about 5 million years ago, and both East rifts have been propagating into roughly 3 millionyear-old Nazca lithosphere. Extremely deep axial valleys occur at their tips, B6000 m at Pito Deep at the northeastern Easter microplate boundary, and B5000 m at Endeavor Deep at the northeastern Juan Fernandez microplate boundary. The northern and southern boundaries are complicated deformation zones, with zones of shear, extension, and significant areas of compression. Both the Easter and Juan Fernandez microplates have large (B100 km 200 km) complex pervasively deformed cores, which could have formed by bookshelf faulting in overlap zones during an initial largescale propagating rift stage of evolution. The abyssal hill fabric and magnetic anomalies of the younger seafloor also show a later stage of growth as



independent microplates, by rift propagation and seafloor spreading on dual active spreading centers, with deformation during this microplate stage concentrated along the plate boundaries and resulting from the microplate rotation. At present, both microplates are rotating clockwise very rapidly about poles near the microplate centers, spinning like roller-bearings caught between the major bounding plates. The Easter microplate rotation velocity is about 151 per million years and the Juan Fernandez velocity is about 91 per million years. This roller-bearing analogy has been quantified in an idealized edge-driven model of microplate kinematics (Figure 5). If microplate rotation is indeed driven by shear between the microplate and surrounding major plates, the rotation velocity (in radians) is 2u/d, where 2u is the major plate relative velocity, and d is the microplate diameter. This follows because the total spreading on the microplate boundaries must be what the major plate motion would be if the microplate did not exist. The rotation (Euler) poles describing the motion of the microplate relative to the major plates will lie on the microplate boundaries, at the farthest extensions of the rifts,

Plate A MP/A pole

Microplate

MP/B pole

Plate B

603

which must lengthen to stay at the Euler poles because of the microplate rotation. Evolution

This idealized geometry requires a circular microplate shape, yet also requires seafloor spreading on the dual active ridges, which must constantly change this shape. The more the microplate grows, the more deformation must occur as it rotates, and the less successful the rigid plate model will be. Although it would appear that this inevitable plate growth would soon invalidate the model, numerous episodes of rift propagation helping to maintain the necessary geometry are observed to have occurred at the Easter and Juan Fernandez microplates. These propagators all propagated on the microplate interior side of the failing rifts, thus transferring microplate lithosphere to the major plates, shaving the new microplate growth at the edges and maintaining a circular enough shape for the edge-driven model to be very successful. Although instantaneous spreading rates may be symmetric, time-averaged accration is highly asymmetric, much faster on the major plates than the microplates because of the lithosphere transferred by rift propagation. Nevertheless, some deformation is clearly occurring along the northern boundaries of both microplates, forming large compressional ridges. According to the edge-driven model, a microplate may stop rotating if one of the bounding ridge axes propagates through to the opposite spreading boundary, eliminating coupling to one of the bounding plates. Dual spreading would no longer occur, spreading would continue on only one bounding ridge, and 106–107 km3 of microplate lithosphere would accrete to one of the neighboring major plates. There is evidence in the older seafloor record that this has happened many times before along the ancestral East Pacific Rise. Causes of Microplate Formation

Figure 5 Roller-bearing model of microplates based on a simple, concentrically rotating bearing. The microplate is approximated by a circular plate (white) which is caught between two major plates A and B. The main contacts between microplate and major plates are also the positions of the relative rotation poles (dots). Dark shading shows major spreading centers, overlapping about the microplate. Cross-hatched corners are areas of compression. Medium curved lines are predicted pseudofaults, and arrows show relative motions. This schematic model assumes growth from an infinitesimal point to a present circular shape – the model can be extended to take account of growth from a finite width, eccentric motions, and growth of the microplate. (Reproduced with permission from Searle et al., 1993.)

The occurrence of large, pervasively deformed cores in both microplates, probably formed during early propagating rift stages of evolution, suggests that these oceanic microplates may have originated as giant propagator or duelling propagator systems which formed overlap zones so big that their mechanical behavior had to change. This propagation could have been normal rift propagation, or from another documented type where a new propagator appears to start from near the center of a long transform fault, perhaps on a small intra-transform spreading center.


604


Both microplates appear to have originated at large left-stepping offsets of the East Pacific Rise. All large right-stepping offsets along the Pacific-Nazca spreading center are transform faults, while all large left-stepping offsets are microplates or the possible duelling propagator proto-microplate between the Easter and Juan Fernandez microplates (Figure 6). The Galapagos microplate at the Pacific-CocosNazca triple junction also fits this pattern. This suggests that a recent clockwise change in PacificNazca plate motion could have been an important

Cocos plate

Galapagos microplate _1

0°

114 km Ma

Pacific plate

10°S

Wilkes transform

factor triggering the formation of these microplates, although the right-stepping Wilkes transform also shows evidence for previous boundary reorganization, and the smaller scale 211S duelling propagators are also right-stepping. The dominant East rift of the Easter microplate, the dominant West rift of the Juan Fernandez microplate, and the dominant West rift of the duelling propagators between the microplates, are all propagating away from the Easter mantle plume (or the intersection of this plume with the ridge axis), suggesting that microplate formation as well as rift propagation can be driven by plume-related forces. Earth’s fastest active seafloor spreading occurs in this area, and every major mid-ocean ridge segment presently spreading faster than 142 km per million years is reorganizing by duelling propagators or microplates (Figure 6). The combination of thin lithosphere produced at these ‘superfast’ seafloor spreading rates, as well as the unusually hot asthenosphere produced by the Easter mantle plume, would reduce the forces resisting propagation and thus make these plate boundary reorganizations easier, perhaps explaining their common occurrence in this area.

_1

136 km Ma

Summary

Garrett transform

Nazca plate _1

20°S

142 km Ma

Pito deep Easter microplate

Easter Island

30°S

_1

149 km Ma

Endeavor deep Juan Fernandez microplate

_1

55 km Ma _1

95 km Ma

120°W

Antarctic plate 110°W

100°W

Figure 6 Plate tectonic geometry and relative plate motions along the southern East Pacific Rise. Light lines are ridges, those with arrows are propagating. Heavy straight lines are transform faults. (Adapted with permission from Hey et al., 1995.)

Rift propagation on scales ranging from overlapping spreading centers with a few kilometers offset up to several hundreds of kilometers at microplate tectonic scales, and indeed all the way up to several thousands of kilometers at continental rifting scales, appears to be the primary mechanism by which Earth’s accretional plate boundary geometry is reorganized. Although conceptually simple, the propagating rift hypothesis has important implications for plate tectonic evolution. It explains the existence of several classes of structures, including pseudofaults, failed rifts, and zones of transferred lithosphere, that are oblique to ridges and transform faults and thus previously seemed incompatible with plate tectonic theory. These are all quantitatively predictable consequences of rift propagation. It explains why passive continental margins are not parallel to the oldest seafloor isochrons, but instead are pseudofaults, bounding lithosphere created on propagating spreading centers and indicating the direction of the continental breakup propagators. It explains the large-scale reorganization of many seafloor spreading systems, including both the origination and termination of many fracture zones, as well as the formation of some transient microplates which appear to be the modern analogs of large-scale



spreading center jumps. This hypothesis provides a mechanistic explanation for the way in which many (if not all) spreading center jumps occur and why they occur in systematic patterns. It explains how spreading centers reorient when the direction of seafloor spreading changes, and the origin of large areas of petrologically diverse seafloor, including the major abyssal ferrobasalt provinces. The common occurrence of rift propagation over a wide range of spreading rates and tectonic environments indicates that it represents an efficient mechanism of adjustment of extensional plate boundaries to the forces driving plate motions.

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

Further Reading Bird RT and Naar DF (1994) Intratransform origins of mid-ocean ridge microplates. Geology 22: 987--990. Hey RN, Naar DF, Kleinrock MC, et al. (1985) Microplate tectonics along a superfast seafloor spreading system near Easter Island. Nature 317: 320--325.

605

Hey RN, Sinton JM, Duennebier FK, and Duennebier FK (1989) Propagating rifts and spreading centers. In: Winterer EL, Hussong DM, and Decker RW (eds.) Decade of North American Geology: The Eastern Pacific Ocean and Hawaii, pp. 161--176. Boulder, CO: Geological Society of America. Hey RN, Johnson PD, Martinez F, et al. (1995) Plate boundary reorganization at a large-offset, rapidly propagating rift. Nature 378: 167--170. Kleinrock MC and Hey RN (1989) Migrating transform zone and lithospheric transfer at the Galapagos 95.51W propagator. Journal of Geophysical Research 94: 13859--13878. Larson RL, Searle MC, Kleinrock MC, et al. (1992) Rollerbearing tectonic evolution of the Juan Fernandez microplate. Nature 356: 517--576. Naar DF and Hey RN (1991) Tectonic evolution of the Easter microplate. Journal of Geophysical Research 96: 7961--7993. Phipps Morgan J and Parmentier EM (1985) Causes and rate limiting mechanisms of ridge propagation: a fraction mechanics model. Journal of Geophysical Research 90: 8603--8612. Schouten H, Klitgord KD, and Gallo DG (1993) Edgedriven microplate kinematics. Journal of Geophysical Research 98: 6689--6701. Searle RC, Bird RT, Rusby RI, and Naar DF (1993) The development of two oceanic microplates: Easter and Juan Fernandez microplates, East Pacific Rise. Journal of the Geological Society 150: 965--976.


PROTOZOA, PLANKTONIC FORAMINIFERA R. Schiebel and C. Hemleben, Tu¨bingen University, Tu¨bingen, Germany

are specialized on planktonic foraminifers are not known.


Introduction Planktonic foraminifers are single celled organisms (protozoans) sheltered by a test (shell) made of calcite, with an average test diameter of 0.25 mm. They live in surface waters of all modern open oceans and deep marginal seas, e.g., Mediterranean, Caribbean Sea, Red Sea, and Japan Sea, and are almost absent from shelf areas including the North Sea and other shallow marginal seas. Planktonic foraminifers constitute a minor portion of the total zooplankton, but are the main producers of marine calcareous particles deposited on the ocean floor and form the socalled ‘Globigerina ooze.’ Planktonic foraminifers (Greek: foramen ¼ opening, ferre ¼ carry) first appeared in the middle Jurassic, about 170 million years ago (Ma), and spread since the mid-Cretaceous over all world oceans. Times of main appearance of new species in the Aptian (120 Ma), the Turonian (90 Ma), the Paleocene (55 Ma), and the Miocene (20 Ma), alternate with phases of main extinction in the Cenomanian (95 Ma), at the Cretaceous/Tertiary boundary (60 Ma), and in the Upper Eocene (40 Ma). Modern planktonic foraminifers have evolved since the early Tertiary, when first spinose species occurred directly after the Cretaceous/Tertiary boundary. Approximately 450 fossil and 50 Recent species are known, not including species based on molecular biology investigations. The appearance and radiation of new species seem to correlate with the development of new realms and niches, linked to plate tectonics and paleoceanographic changes. The geographical distribution and main events in planktonic foraminiferal evolution are associated in general with water mass properties, e.g., availability of food or temperature. The reproductive strategies depend highly on their life habitat in the photic zone or slightly below. The life span of planktonic foraminifers varies between 14 days and a year, mostly linked to the lunar cycle. Most living species bear symbionts requiring a habitat in the upper to middle photic zone. Their feeding habit depends on the spinosity (spinose versus nonspinose species) in respect to the size and class of prey. Predators that

606

History With the technological improvement of microscopes d’Orbigny in 1826 was able to describe the first planktonic foraminiferal species, Globigerina bulloides, from beach sands, and classified it as a cephalopod. In 1867 Owen described the planktonic life habit of these organisms. Following the Challenger Expedition (1872–1876) the surface-dwelling habitat of planktonic foraminifers was recognized. Rhumbler first described the biology of foraminifers in 1911. In the first half of the twentieth century, foraminifers were widely used for stratigraphic purposes, and many descriptions were published, mainly by Josef A. Cushman and co-workers. Studies on the geographic distribution of individual foraminiferal species are based on samples from the living plankton since the work of Schott in 1935. Planktonic foraminifers have been used since the beginning of the twentieth century to date marine sediments drilled by oil companies, and later on through the DeepSea Drilling and Ocean Drilling Programs. In addition, extensive studies on distribution, ecology of live and fossil faunas were carried out to understand the changing marine environment. The ecological significance has been applied in paleoecological and paleoceanographic settings and yielded subtle information on ancient oceans and the Earth’s climate. Recent investigation still focuses on evolution and population dynamics. Modern techniques, e.g., polymerase chain reaction (PCR), are being used to reveal the genetic code, and their relation to morphological classification tests needs to be checked.

Methods Planktonic foraminifers are sampled from the water column by plankton nets of various design, with a mesh size of 0.063–0.2 mm, by employing plankton recorders, water samplers, pumping systems, or collection by SCUBA divers. To study faunas from sediment samples or consolidated rock, the surrounding sediment has to be disaggregated by hydrogen peroxide, tensids, acetic acid (pure), or physical methods, and washed over a sieve (0.03– 0.063 mm). Shells may be studied under a binocular microscope, or with a scanning electron microscope


PROTOZOA, PLANKTONIC FORAMINIFERA

Cellular Structure Planktonic foraminifers have a single cell that builds calcareous shells and forms chambered tests. Chamber formation, resulting from deposition of calcite, takes place within a cytoplasmic envelope produced by rhizopodia that also secrete a primary organic membrane. A calcitic bilamellar wall is formed at the primary organic membrane. The only exception is the monolamellar genus Hastigerina (Figure 1). The proximal side of the POM consists of two to three calcite layers whereas the distal (outer) layer reveals as many layers as there are chambers. Layered pustules are built within the outer layers of the wall, concurrent with successive stages of calcite

OCL POM IOL (A)

PP MP CY

P OL OCL

PP

POM ICL IOL (B)

MP

PP

for more detail. Transmission electron microscopy is suited to the study of cytoplasm at high resolution. Some species have already been cultivated under laboratory conditions. For faunistic analysis live and dead specimens are distinguished by their content of cytoplasm. For statistical significance on average 300 individuals have to be classified and counted. According to the distribution and ecology of modern planktonic foraminifers, and due to the fact that their calcitic tests contribute substantially to the microfossil faunal record of marine sediments, planktonic foraminifers are used in reconstructing the climatic, ecological, and geological history of the Earth. Physical factors that determine the modern faunal composition are related to the fossil assemblages by multiple regression statistical techniques (transfer functions) to yield a confident estimate on ancient environmental parameters. Stable isotopic (18/16O and 13/12C) and trace element ratios of the calcareous (calcite) shell display mostly the composition of the ambient water. These so-called proxies of the physical, chemical, and biological state of modern and ancient oceans are used to reconstruct productivity, temperature, and salinity of paleo-water masses, and to determine the relative age of marine sediments. Laboratory experiments and field calibration are carried out for synoptical evaluation of physical and chemical controls over the geochemical composition of foraminiferal calcite. The metabolic fractionation of isotopes (vital effect) that are included in the foraminiferal shell, varies between species, and depends on water temperature and carbonate (CO32) concentration. The radioactive 14C isotope gives an absolute age of the shell, limited to the last approximately 40 000 years. Molecular biology methods have recently been used to investigate foraminiferal rRNA genes (rDNA) after DNA extraction, amplification by PCR and normally automated sequencing.

607

(C)

MP

Figure 1 Schematic diagrams of wall structures and pores in bilamellar spinose (A), bilamellar nonspinose (B), and monolamellar (C) planktonic foraminifera (courtesy Cushman Foundation). CY, foraminiferal cytoplasm; ICL, inner calcite layer; IOL, inner organic lining; MP, micropore; OCL, outer calcite layer; OL, outer organic layer; POM, primary organic membrane; P, pustule; PP, pore plate.

lamination. Spines are not layered and are lodged as plugs within the wall. Intrashell cytoplasm is differentiated from a reticulate or rhizopodial type on the outer shell. Planktonic foraminiferal cell organelles, e.g., nucleus, mitochondria, peroxisomes, Golgi complex, endoplasmic reticulum, annulate lamellae, vacuolar system (Figure 2), are typical of those observed in other eukaryotic cells. A fibrillar system seems to be unique among known protozoa, and is suspected to be a floating device or calcifying organelle. Food in the form of lipids and starch is stored in special vacuoles. Chambers are connected by openings (foramen) between them and have sealed pores in the chamber wall which faces the external environment. Gas exchange between cell and the ambient sea water takes place through these pores; the aperture(s) serves for cytoplasmic contact with the surrounding water, mainly to exchange food particles and waste products. Different types of spines, pores, wall structures, and test morphology may adapt the species to certain


608


VAC f

f2 go

mi

at f1

f

S

li

S

VAC

Figure 2 Transmission electron microscopical sections of Globigerinella siphonifera cytoplasm including cell organelles. at, animal tissue; f, food vacuole; f1, food vacuole including fresh green algae; f2, food vacuole including partly digested green algae; go, Golgi complex; li, lipid droplet; mi, mitochondria; s, symbiont; vac, empty vacuole. Scalebar ¼ 3 mm.

environments and are of taxonomic significance. Spines allow anastomosing cytoplasm to stretch far out of the test, to form rhizopodial nets for capturing prey, and to carry prey and symbionts as on a conveyor belt to support the cell.

Reproduction and Ontogeny Planktonic foraminifers probably display only sexual reproduction without the diploid generation. Shallow-dwelling species have been shown to reproduce once per month (Globigerina bulloides), or within two weeks (Globigerinoides ruber), triggered by the synodic lunar cycle. During reproduction adults release gametes (several hundred thousand) to form offspring with a first calcified chamber (proloculus), which is 8–34 mm in size. The prolocular ontogenetic stage consists of a first (protoconch) and second chamber (deuteroconch). First juvenile chambers are formed on a subdaily rate. During ontogeny the rate of chamber formation gradually decreases. The neanic stage is marked by substantial changes in morphology, and occasional changes in selection of diet and depth habitat, which might explain the relative enrichment of d13C with increasing test size. Maturity is reached when the adult stage is reached and tests consist of 10–20 chambers, with a size of 0.1–2 mm (about 0.25 mm on average). The terminal stage is related to reproduction and marked by chamber alterations such as shedding of spines and partial wall thickening. The empty adult test sinks towards the seafloor forming the ‘Globigerina Ooze.’

Symbionts, Commensals, and Parasites Species that bear symbionts (mostly spinose species) are bound to light and live in the euphotic zone of the

ocean. Some species without symbionts live in the deep ocean, and only ascend to the sea surface once a year to reproduce (e.g., Globorotalia truncatulinoides). Symbionts associated with spinose (Figure 3) and occasionally with nonspinose species are dinoflagellates and chrysophycophytes, which may contribute photosynthetic compounds to the host and provide energy to drive the calcification process. This is especially important for the use of d13CSHELL in paleo-reconstructions because the d13C, among other parameters, is determined by the symbiont activity, which is directly correlated to the light level in the water column. Commensalistic dinophytes acquire nutrients as metabolic by-products from the host. Parasites, such as sporozoans, may dart in and feed on foraminiferal cytoplasm.

Molecular Biology Most recently, methods widely used in molecular biology have also been applied to planktonic foraminifers. A general question is based on the bipolar distribution of the group, and the genetic relationships of species in both hemispheres, especially those clearly distributed in both arctic and antarctic cold waters. Species diversity, based on molecular genetics, is greater than would be expected by applying the traditional concept of morphotaxa. This points towards a larger variety of genetically defined populations, which may permit these to be classified as cryptic sibling species. In addition, the most recent results indicate a polyphyletic origin linked to benthic foraminifers. The molecular methods being used explore the small and large subunits of ribosomal DNA (SSU and LSU rDNA) sequence variability. The results achieved by these molecular methods are manyfold and as yet cannot be explained consistently. However, the large potential of this new field



609

tests have been found in pteropods, salps, shrimp, and other metazooplankton. Species are spatially and temporally distributed according to diet and temperature and are sensitive to environmental impacts.

Ecology and Distribution

Figure 3 Spinose planktonic foraminifer Orbulina universa. Spines allow cytoplasm to stretch far out of the test, to form a rhizopodial net. Symbionts are carried out by cytoplasm streaming during the light period, and are withdrawn into the test during darkness. Diameter of the test is about 0.5 mm (without spines).

will certainly aid in unraveling the distribution pattern, ecological adaption, speciation, and phylogeny of planktonic foraminifers.

Trophic Demands Planktonic foraminifers are basically omnivorous. Spinose species prefer a wide variety of animal prey, including larger metazoans such as copepods, pteropods, and ostracods. Cannibalism has also been reported and bacteria are suspected to form part of the diet. Nonspinose species are largely herbivorous. However, in addition to diatoms, which seem to be the major diet, dinoflagellates, thecate algae, and eukaryotic algae, and also muscle tissue and other animal tissue has been found in food vacuoles. The position of planktonic foraminifers in the marine food web is, therefore, different compared to other protozoans, and occasionally ranges above the basic level of heterotrophic consumers. Predators specialized on planktonic foraminifers are not known, but

Different faunal groups are characteristic of various oceanic realms. Species are bound to their typical depth habitat in the water column, permitting separation of potentially competing species, and faunal composition changes on a temporal and spatial scale. The vertical separation of the habitats of different species is more evident in warmer than in colder waters; physical and biotic conditions between the sea surface and bathyal depth vary more in subtropical and tropical regions than at high latitudes. Faunal provinces roughly follow a latitudinal pattern, displaying the water temperature and salinity. However, on a finer scale the amount and quality of light, turbidity of the ambient water, trophic state, and distribution of predators play an important role. Only two Recent species (Neogloboquadrina pachyderma and Turborotalita quinqueloba) are frequent in polar regions. In general, assemblages of planktonic foraminifers occur in five major faunal provinces: (1) polar, (2) subpolar, (3) transition, (4) subtropical, and (5) tropical. Faunal mixing occurs due to hydrodynamic features (e.g., upwelling or current systems) and additional provinces are (6) upwelling, (7) subtropic/tropic, and (8) transitional/subpolar (Figure 4). Special environments like the upwelling of nutrientrich water masses are characterized by high numbers of Globigerina bulloides. Typical faunas exist along the margins of the subtropical gyres and at hydrographic frontal systems. The highest diversity is recorded from temperate to subtropical waters (Figure 5). A seasonal distribution pattern of planktonic foraminifers is most pronounced in high and mid-latitudes, displaying the phytoplankton succession and associated food chain. Due to meso-scale and local features and a certain reproduction pattern, the distribution of planktonic foraminifers is patchy on various temporal and spatial scales. The highest numbers (41000 specimens per m3) of adult tests (40.1 mm) are recorded in areas and during times of highest primary production, which are upwelling areas and seasonal blooms in the temperate and polar oceans (Figure 6). High numbers of individuals correlate with maximum amounts of chlorophyll in the upper ocean and to the deep chlorophyll maximum at the base of the surface mixed layer of the ocean. In the mesotrophic to oligotrophic ocean 1–50 specimens per m3 occur, and


610


80

40

0

40

80

120

160

160

120

80

80

2

60 3

40

4

4 67

5

0

6

8 6

6

5 5

4 40

1

1

1

7

3 2

4 3

7

4

2 2

3

1

1

1

6

70 Figure 4 Foraminiferal provinces. 1, polar; 2, subpolar; 3, transitional; 4, subtropic; 5, tropic; 6, upwelling; 7, subtropic/tropic; 8, transitional/subpolar (after Hemleben et al., 1989).

0˚C

Neogloboquadrina Polar 5˚C

Globorotalia

Subpolar 10˚C

Orbulina Transitional

Globigerina

Globigerinoides

18˚C

ruber

Subtropical

sacculifer 24˚C Tropical

Globigerinella

30˚C

Figure 5 Schematic distribution pattern of modern planktonic foraminifers exemplified for the North Atlantic (according to Hemleben et al., 1989). Species from the lower left to the upper right are Globigerinoides sacculifer, Globigerinoides ruber, Globigerinella siphonifera, Orbulina universa (spherical stage cracked open to show the interior preadult test), Globorotalia truncatulinoides, Globorotalia inflata, Globigerina bulloides, and Neogloboquadrina pachyderma.

from blue waters (e.g., eastern Mediterranean) less than one specimen per m3 is reported.

Sedimentation Global calcite production of planktonic foraminifers amounts to about two Gigatons per year, from which

only 1–2% reaches the deep sea floor. Planktonic foraminiferal shells dissolve while settling through the water column. Preservation of tests depends on the biogeochemistry of the ambient water and on the resting time of tests in the water column. Due to the low sinking velocity and long time of exposition, small and thin-walled tests (about 100 m per day) are


Water depth (m)


100 300 500 700 900 1100 1300 ? 1500 1700 1900 2100 2300 2500 January Feb. March April

611

? ?

?

May

June July August Sept. Oct.

Nov.

Dec.

Figure 6 The calcite flux of planktonic foraminiferal shells from the upper water column to the deep sea is determined by population dynamics and settling velocity of empty tests. In the eastern North Atlantic maximum standing stocks of more than 400 specimens per m3 occur in the upper 300 m during the phytoplankton spring bloom (March through May). A minor maximum in abundance during fall (September to October) is due to redistribution of chlorophyll and entrainment of nutrients and resulting phytoplankton growth in surface waters. During summer and winter the number of specimens may not exceed 10–50 per m3. Most species live in surface waters. Only a few species live at a depth of 100–500 m. Abyssal species are rare in the transitional North Atlantic and more frequent in the subtropical realm. As a result of population dynamics and differential settling velocity of tests (100–1500 m day1), the test flux occurs in pulses (arrows), being highest during spring and fall exceeding 60 mg m2 d1 (red; orange ¼ 30–60; yellow ¼ 10–30; dark green ¼ 3–10; light green ¼ 1–3; blue ¼ o1 mg m2 d1). Remineralization of tests is highest between 200 and 700 m depth. Below 700 m major CaCO3 flux is restricted to mass sinking events during high-productivity periods. November and December have so far not been sampled.

preferentially removed from the settling assemblage, and mainly large and fast sinking tests (up to 1500 m per day) are deposited at the seafloor. Mass sinking of aggregates (marine snow) during seasons of enhanced biological productivity includes planktonic foraminiferal test, which balances the fossil faunal record towards species assemblages that reflect high productivity, e.g., seasonal upwelling and spring blooms (Figure 6). A substantial amount of planktonic foraminiferal shells is remineralized far above the calcite lysocline, between 200 and 700 m water depth. Below the calcite compensation depth virtually no calcareous particles are preserved.

rates of CaCO3 of planktonic foraminiferal and other marine calcite-sequestrating organisms (mainly coccolithophorids and pteropods). Their role in the marine and global carbon budget which still needs to be quantified, provides great potential information on marine biogeochemistry.

See also Benthic Foraminifera. Calcium Carbonates. Conservative Elements. Geophysical Heat Flow. Large Marine Ecosystems. Marine Snow. Pelagic Biogeography. Plankton and Climate. Upwelling Ecosystems.

Application As major contributors to the vertical CaCO3 flux, planktonic foraminiferal shells cause a substantial portion of CaCO3 burial in deep-sea sediments. As a component of the marine carbon turnover and vertical flux, planktonic foraminifers are of major interest for paleoclimatologists, because their tests carry fossil information on climates since the midCretaceous. Their faunal composition, details of the test morphology, and their stable isotope and element ratios, provide detailed information on paleotemperature (d18O, Mg/Ca, Sr/Ca, d44Ca), paleoproductivity (d13C), paleo-pH (d11B), nitrate (NO3) concentration of seawater (d15N), and paleoCO2 levels by estimating the vertical flux and burial

Further Reading Be´ AHW (1977) An ecological, zoogeographic and taxonomic review of Recent planktonic Foraminifera. In: Ramsay ATS (ed.) Oceanic Micropaleontology, vol. 1, pp. 1--100. London, New York, San Francisco: Academic Press. Bolli HM, Saunders JB, and Perch-Nielsen K (1985) Plankton Stratigraphy. Cambridge: Cambridge University Press. Darling K, Wade CM, and Stewart IA (2000) Molecular evidence for genetic mixing of Arctic and Antarctic subpolar populations of planktonic foraminifers. Nature 405: 43--47. Fischer G and Wefer G (1999) Use of Proxies in Paleoceanography. Berlin: Springer-Verlag.


612


Hemleben Ch, Spindler M, and Anderson OR (1989) Modern Planktonic Foraminifera. New York: SpringerVerlag. Kennet JP and Shrinivasan MS (1983) Neogene Planktonic Foraminifera. A Phylogenetic Atlas. Stroudsburg, PA: Hutchinson Ross. Loeblich AR Jr and Tappan H (1987) Foraminiferal Genera and Their Classification. New York: Van Nostrand Reinhold.

Olsson RK, Hemleben Ch, Berggren WA, and Huber B (eds.) (1999) Atlas of Paleocene Planktonic Foraminifera. Smithsonian Contributions to Paleobiology 85. Washington, DC: Smithsonian Institution Press. Spero HJ, Lea DW, and Young JR (1996) Experimental determination of stable isotope variability in Globigerina bulloides: implications for paleoceanographic reconstructions. Marine Micropaleontology 28: 231--246.


PROTOZOA, RADIOLARIANS O. R. Anderson, Columbia University, Palisades, NY, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2315–2320, & 2001, Elsevier Ltd.

Introduction Radiolarians are exclusively open ocean, silica-secreting, zooplankton. They occur abundantly in major oceanic sites worldwide. However, some species are limited to certain regions and serve as indicators of water mass properties such as temperature, salinity, and total biological productivity. Abundances of total radiolarian species vary across geographic regions. For example, maximum densities reach 10 000 per m3 in some regions such as the subtropical Pacific. By contrast, densities range about 3–5 per m3 in the Sargasso Sea. Radiolarians are classified among the Protista, a large and eclectic group of eukaryotic microbiota including the algae and protozoa. Algae are photosynthetic, single-celled protists, while the protozoa obtain food by feeding on other organisms or absorbing dissolved organic matter from their environment. Radiolarians are single-celled or colonial protozoa. The single-celled species vary in size from

(A)

o100 mm to very large species with diameters of 1–2 mm. The larger species are taxonomically less numerous and include mainly gelatinous species found commonly in surface waters. The smaller species typically secrete siliceous skeletons of remarkably complex design (Figure 1). The skeletal morphology is species-specific and used in taxonomic identification. Larger, noncolonial species are either skeletonless, being enclosed only by a gelatinous coat, or produce scattered siliceous spicules within the peripheral cytoplasm and surrounding gelatinous layer. Colonial species contain numerous radiolarian cells interconnected by a network of cytoplasmic strands and enclosed within a clear, gelatinous envelope secreted by the radiolarian. The colonies vary in size from several centimeters to nearly a meter in length. The shape of the colonies is highly variable among species. Some are spherical, others ellipsoidal, and some are elongate ribbon-shaped or cylindrical forms. These larger species of radiolarians are arguably, the most diverse and largest of all known protozoa. Many of the surface-dwelling species contain algal symbionts in the peripheral cytoplasm that surrounds the central cell body. The algal symbionts provide some nutrition to the radiolarian host by secretion of photosynthetically produced organic products. The food resources are absorbed by the radiolarian and,

(B)

Figure 1 Morphology of polycystine radiolaria. (A) Spumellarian with spherical central capsule and halo of radiating axopodia emerging from the fusules in the capsular wall and surrounded by concentric, latticed, siliceous shells. (B) Nassellarian showing the ovate central capsule with conical array of microtubules that extend into the basally located fusules and external axopodia protruding from the opening of the helmet-shaped shell. Reproduced with permission from Grell K (1973) Protozoology. Berlin: Springer-Verlag.


613

614

PROTOZOA, RADIOLARIANS

combined with food gathered from the environment, are used to support metabolism and growth. Radiolarians that dwell at great depths in the water column where light is limited or absent typically lack algal symbionts. The siliceous skeletons of radiolarians settle into the ocean sediments where they form a stable and substantial fossil record. These microfossils are an important source of data in biostratigraphic and paleoclimatic studies. Variations in the number and kind of radiolarian species (based on skeletal form) in relation to depth in the sediment provide information about climatic and environmental conditions in the overlying water mass at the time the radiolarian skeletons were deposited at that geographic location. The radiolarians are second only to diatoms as a major source of biogenic opal (silicate) deposited in the ocean sediments.

N

CW

PV

DV.

SK

Cellular Morphology The radiolarian cell body contains a dense mass of central cytoplasm known as the central capsule (Figure 2). Among the organelles included in the central capsule are the nucleus, or nuclei in species with more than one nucleus, most of the food reserves, major respiratory organelles, i.e., mitochondria, Golgi bodies for intracellular secretion, protein-synthesizing organelles, and vacuoles. The central capsule is surrounded by a nonliving capsular wall secreted by the radiolarian cytoplasm. The thickness of the capsular wall varies among species. It may be thin or in some species very reduced, consisting of only a sparse deposit of organic matter contained within the surrounding cytoplasmic envelope. In others, the wall is quite thick and opalescent with a pearl-like appearance. The capsular wall contains numerous pores through which cytoplasmic strands (fusules) connect to the extracapsular cytoplasm. The extracapsular cytoplasm usually forms a network of cytoplasmic strands attached to stiffened strands of cytoplasm known as axopodia that extend outward from the fusules in the capsular wall. The central capsular wall and axopodia are major defining taxonomic attributes of radiolarians. A frothy or gelatinous coat typically surrounds the central capsule and supports the extracapsular cytoplasm. Algal symbionts, when present, are enclosed within perialgal vacuoles produced by the extracapsulum. In most species, the algal symbionts are exclusively located in the extracapsulum. Thus far, symbionts have been observed within the central capsular cytoplasm in only a few species. Food particles, including small algae and protozoa or larger invertebrates such as copepods, larvacea, and crustacean larvae, are captured by the

Figure 2 Cytoplasmic organization of a spumellarian radiolarian showing the central capsule with nucleus (N), capsular wall (CW) and peripheral extracapsulum containing digestive vacuoles (DV) and algal symbionts in perialgal vacuoles (PV). The skeletal matter (SK) is enclosed within the cytokalymma, an extension of the cytoplasm, that acts as a living mold to dictate the shape of the siliceous skeleton deposited within it. Reproduced with permission from Anderson OR (1983) Radiolaria. New York: Springer-Verlag.

sticky rhizopodia of the extracapsulum. The cytoplasm moves by cytoplasmic streaming to coat and enclose the captured prey. Eventually, the prey is engulfed by the extracapsular cytoplasm and digested in digestive vacuoles (lysosomes). These typically accumulate in the extracapsulum near the capsular wall. Large prey such as copepods are invaded by flowing strands of cytoplasm and the more nutritious soft parts such as muscle and organ tissues are broken apart, engulfed within the flowing cytoplasm and carried back into the extracapsulum where digestion takes place. The siliceous skeleton, when present, is deposited within cytoplasmic spaces formed by extensions of the rhizopodia. This elaborate framework of skeletal-depositing cytoplasm is known as the cytokalymma. Thus, the form of the skeleton is dictated by the dynamic streaming and molding action of the cytokalymma during the silica deposition process. Consequently, the very elaborate and species-specific form of the skeleton is determined by the dynamic activity of the radiolarian and is not simply a consequence of passive physical chemical processes taking place at interfaces among the frothy components of the cytoplasm as was previously proposed by some researchers.



Taxonomy Radiolarians are included in some modern classification schemes in the kingdom Protista. However, the category of radiolaria as such is considered an artificial grouping. Instead of the group ‘Radiolaria’, two major subgroups previously included in ‘Radiolaria’ are placed in the kingdom Protista. These are the Polycystina and the Phaeodaria. Polycystina are radiolarians that contain a central capsule with pores that are rather uniform in shape and either uniformly distributed across the surface of capsular wall, or grouped at one location. The Phaeodaria have capsular walls with two distinctive types of openings. One is much larger and is known as the astropyle with an elaborately organized mass of cytoplasm extending into the extracapsulum. The other type is composed of smaller pores known as parapylae with thin strands of emergent cytoplasm. Some Phaeodaria also have skeletons that are enriched in organic matter compared with the skeletons of the Polycystina. Among the Polycystina, there are two major taxonomic groups, the Spumellaria and Nassellaria, assigned as orders in some taxonomic schemes. Spumellaria have central capsules that are usually spherical or nearly so at some stage of development and have pores distributed uniformly over the entire surface of the capsular wall. All known colonial species are members of the Spumellaria. Although expert opinion varies, there are two families and about 10 genera of colonial radiolarians. There are seven widely recognized families of solitary Spumellaria with scores of genera. Nassellaria have central capsules that are more ovate or elongated and the fusules are located only at one pole of the elongated capsular wall. This pore field is called a porochora and the fusules tend to be robust with axopodia that emerge through outward-directed collar-like thickenings surrounding the pore rim. Moreover, the skeleton of the Nassellaria, when present, tends to be elongated and forms a helmet-shaped structure, often with an internal set of rods forming a tripod to which the external skeleton is attached. Current systematics include seven major families with numerous genera. Spumellarian skeletons are typically more spherical, or based on a form that is not derived from a basic tripodal or helmet-like architectural plan. The shells of the Phaeodaria are varied in shape. Some species have ornately decorated open lattices resembling geodesic structures composed of interconnected, hollow tubes of silica. Other species have thickened skeletons resembling small clam shells with closely spaced pores on the surface. There are 17 major families with scores of genera. Since many species of radiolarians were first identified from sediments based solely on their mineralized skeletons, much of the key taxonomic characteristics include skeletal

615

morphology. Increasingly, evidence of cytoplasmic fine structure obtained by electron microscopy and molecular genetic analyses is being used to augment skeletal morphology in making species discriminations and constructing more natural evolutionary relationships. It is estimated that there are several hundred valid living species of radiolarians.

Biomineralization Biomineralization is a biological process of secreting mineral matter as a skeleton or other hardened product. The skeleton of radiolaria is composed of hydrated opal, an oxidized compound of silicon (nominally SiO2 nH2O) highly polymerized to form a space-filling, glassy mass incorporating a variable number (n) of water molecules within the molecular structure of the solid. Electron microscopic evidence indicates that some organic matter is incorporated in the skeleton during early stages of deposition, but on the whole, the skeleton is composed mostly of pure silica. Electron microscopic, X-ray dispersive analysis shows that a small amount of divalent cations such as Ca2þ may be incorporated in the final veneer deposited on the surface to enhance the hardness of the skeleton. During deposition of the skeleton, the cytoplasm forms the living cytokalymma, i.e., the cytoplasmic silica-depositing mold, by extension of the surface of the rhizopodia. The cytokalymma enlarges as silica is deposited within it, gradually assuming a final form that dictates the morphology of the internally secreted skeleton. Small vesicles are observed streaming outward from the cell body into the cytoplasm of the cytokalymma and these may bring silica to be deposited within the skeletal spaces inside the cytokalymma. The cytoplasmic membrane surrounding the developing skeleton appears to act as a silicalemma or active membrane that deposits the molecular silica into the skeletal space. During deposition, the dynamic molding process is clearly evident as the living cytoplasm continuously undergoes transformations in form, gradually approximating the ultimate geometry of the species-specific skeleton being deposited by the radiolarian cell. In general, species with multiple, concentric, lattice shells surrounding the central capsule appear to lay down the lattices successively, progressing outward from the innermost shell. The process of skeletal construction has been documented in fair detail for a few species, most notably among the colonial radiolaria. Two forms of growth have been identified. Bar growth is a process of depositing silica as rodlets within a thin tubular network of cytoplasm formed by the cytokalymma. The rodlets become connected during silicogenesis and further augmented with silica to form a porous lattice with typically large polygonal pores. The


616


pores, once formed, may be further subdivided into smaller pores by additional bar growth that spans the opening of the pore. Rim growth occurs by deposition of silica as curved plates that are differentially deposited at places to form rounded pores. At maturity, these are typically spherical skeletons with rather regular, rounded pores scattered across the surface. For both types of skeletons, in some species, the ratio of the bar width between the pores to the pore diameter is a taxonomic diagnostic feature. The rate of silica biomineralization in some species has been determined by daily observation of growth of individuals in laboratory culture using light microscopy. The amount of silica in the skeleton of a living radiolarian is mathematically related to the size of the skeleton. For example, in the spumellarian species Spongaster tetras with a rectangular, spongiose skeleton, the amount of silica (W) in micrograms (mg) as related to the length of the major diagonal axis of the quadrangular shell (L) in micrometers (mm) is approximated as follows: W ¼ 3:338 106 L2:205

½1

The average daily growth in cultures of an S. tetras is 3 mm with an average daily gain in weight of c. 8 ng. The total weight gain for one individual radiolarian during maturation is about 0.1 mg. Silica deposition during maturation appears to be sporadic and irregular, varying from one individual to another, with periods of rapid deposition followed by plateaus in growth. The amount of skeletal opal produced by S. tetras alone in the Caribbean Sea, for example, is c. 42 mg per m3 of sea water, with a range of 8–61 mg per m3. Peak production occurred in mid-summer (June to July). The rate of total radiolarian-produced biogenic opal settling into the ocean sediments at varying oceanic locations has been estimated in the range of 1–10 mg per m2 per day.

Reproduction Protozoa reproduce by either asexual or sexual reproduction. Asexual reproduction occurs by cell division during mitosis to produce two or more genetically identical offspring. Sexual reproduction occurs by the release of haploid gametes (e.g., sperm and egg cells) that fuse to produce a zygote with genetic characteristics contributed by both of the parent organisms. Thus, sexual reproduction permits new combinations of genetic material and the offspring are usually genetically different from the parents. There is evidence that some colonial radiolaria have asexual reproduction. The central capsules within the colony have been observed to divide by fission. This increases the number of central

capsules and allows the colony to grow in size. The colony may also break into parts, thus increasing the total numbers of colonies at a given location. In most species of radiolaria, reproduction occurs by release of numerous flagellated swarmer cells that are believed to be gametes. The nucleus of the parent radiolarian undergoes multiple division and the entire mass of the parent cell is converted into uninucleated flagellated swarmers. These are released nearly simultaneously in a burst of activity, and presumably after dispersal fuse to form a zygote. The details of gamete fusion and the early ontogenetic development of radiolaria are poorly understood and require additional investigation. Ontogenetic development of individuals from very early stages to maturity has been documented in laboratory cultures and the stages of skeletal deposition are well understood for several species, as explained above in the section on biomineralization.

Physiological Ecology and Zoogeography The physiological ecology of radiolaria has been studied by collecting samples of radiolaria and other biota at varying geographical locations in the world oceans to determine what abiotic and biotic factors are correlated with and predict their abundances, and by experimental studies of the physical and biological factors that promote reproduction, growth, and survival of different species under carefully controlled laboratory conditions. Temperature appears to be a major variable in determining abundances of some species of radiolaria. For example, high latitude species that occur abundantly at the North or South Poles are also found at increasing depths in the oceans toward the equator. Since the water temperature in general decreases with depth, these organisms populate broad depth regions within the water column that match their physiological requirements. Species that occur in subtropical locations, where the water is intermediate in temperature based on a global range, are found at the equator at intermediate water depths that are cooler than the warm surface water. Some species are characteristically most abundant in only warm, highly productive water masses. For example, some species of colonial radiolaria occur typically in surface water near the equator in the Atlantic Ocean, while others are most abundant at higher latitudes in the Sargasso Sea where usually the water is also less productive. Upwelling regions where deep, nutrientenriched sea water is brought to the surface are typically highly productive regions for radiolaria, as occurs for example along the Arabian, Chilean, and California coast lines. Shallow-water dwelling species



have been categorized into seven zoogeographic zones based on water mass properties: (1) SubArctic at high northern latitudes; (2) transition region as occurs in the North Pacific drift waters; (3) north central region, typical of waters within the large anticyclonic circulation of the North Pacific; (4) equatorial region in locations occupied by the North and South Equatorial Current systems; (5) south central water mass, as in the South Pacific anticyclonic circulation pattern; (6) subAntarctic, a water regime bounded on the north by the Subtropical Convergence and on the south by the Polar Convergence; and (7) Antarctic, bounded by the Polar Convergence on the north and the Antarctic Continent on the south. The growth requirements of some species have been studied extensively in laboratory cultures. For example, the following three surface- to near-surface-dwelling species exhibit a range of optimal growth conditions. Didymocyrtis tetrathalamus, with a somewhat hourglass-shaped skeleton (150 mm), prefers cooler water (21–271C) and salinities in the range of 30–35 ppm. Dictyocoryne truncatum, a spongiose triangular-shaped species (300 mm), is more intermediate in habitat requirements with optimal temperature of 281C and salinity of 35 ppm. Spongaster tetras, a quadrangular, spongiose species (300 mm), prefers warmer, more saline water (c. 281C and 35–40 ppm). The temperature tolerance ranges (in 1C) for the three species also show a similar pattern of increasing preference for warmer water, i.e., 10–34, 15–28, and 21–31, respectively. The prey consumed by radiolarians varies substantially among species, but many of the polycystine species appear to be omnivorous, consuming both phytoplankton and zooplankton prey. The smaller species consume microplankton and bacteria. Larger species are capable of capturing copepods and small invertebrates. Phaeodaria, especially those species dwelling at great depths in the water column, appear to consume detrital matter in addition to preying on plankton in the water column. The broad range of prey accepted by many of the radiolarians studied thus far suggests that they are opportunistic feeders and are capable of adapting to a broad range of trophic conditions. The role of algal symbionts, when present, has been debated for some time – beginning with their discovery in the mid-nineteenth century. At first, it was supposed that the green symbionts may largely provide oxygen to the host. However, most radiolaria dwell in fairly well-oxygenated habitats and it is unlikely that photosynthetically derived oxygen is necessary. The

617

other competing hypothesis was that the symbionts provide organic nourishment to the host. Modern physiological studies have confirmed that the algal symbionts provide photosynthetically produced nutrition for the host. Biochemical analyses combined with 14C isotopic tracer studies have shown that stores of lipids (fats) and carbohydrates in the host cytoplasm contain carbon derived from algal photosynthetic activity. Well-illuminated, laboratory cultures of symbiont-bearing radiolaria survive for weeks without addition of prey organisms. Some of the algal symbionts are digested as food and can be replaced by asexual reproduction of the algae, but it appears that much of the nutrition of the host comes from organic nutrients secreted into the host cytoplasm by the algal symbionts. This readily available, ‘internal’ supply of autotrophic nutrition makes symbiont-bearing radiolaria much less dependent on external food sources and may account in part for their widespread geographic distribution, including some oligotrophic water masses such as the Sargasso Sea.

See also Marine Silica Cycle.

Further Reading Anderson OR (1983) Radiolaria. New York: SpringerVerlag. Anderson OR (1983) The radiolarian symbiosis. In: Goff LJ (ed.) Algal Symbiosis: A Continuum of Interaction Strategies, pp. 69--89. Cambridge: Cambridge University Press. Anderson OR (1996) The physiological ecology of planktonic sarcodines with applications to paleoecology: Patterns in space and time. Journal of Eukaryotic Microbiology 43: 261--274. Cachon J, Cachon M, and Estep KW (1990) Phylum Anctinopoda Classes Polycystina ( ¼ Radiolaria) and Phaeodaria. In: Margulis L, Corliss JO, Melkonian M, and Chapman DJ (eds.) Handbook of Protoctista, pp. 334--379. Boston: Jones and Barlett. Casey RE (1971) Radiolarians as indicators of past and present water masses. In: Funnell BM and Riedel WR (eds.) The Micropaleontology of Oceans, pp. 331--349. Cambridge: Cambridge University Press. Steineck PL and Casey RE (1990) Ecology and paleobiology of foraminifera and radiolaria. In: Capriulo GM (ed.) Ecology of Marine Protozoa, pp. 46--138. New York: Oxford University Press.


RADIATIVE TRANSFER IN THE OCEAN C. D. Mobley, Sequoia Scientific, Inc., WA, USA Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2321–2330, & 2001, Elsevier Ltd.

Introduction Understanding how light interacts with sea water is a fascinating problem in itself, as well as being fundamental to fields as diverse as biological primary production, mixed-layer thermodynamics, photochemistry, lidar bathymetry, ocean-color remote sensing, and visual searching for submerged objects. For these reasons, optics is one of the fastest growing oceanographic research areas. Radiative transfer theory provides the theoretical framework for understanding light propagation in the ocean, just as hydrodynamics provides the framework for physical oceanography. The article begins with an overview of the definitions and terminology of radiative transfer as used in oceanography. Various ways of quantifying the optical properties of a water body and the light within the water are described. The chapter closes with examples of the absorption and scattering properties of two hypothetical water bodies, which are characteristic of the open ocean and a turbid estuary, and a comparison of their underwater light fields.

Terminology The optical properties of sea water are sometimes grouped into inherent and apparent properties.

•

•

Inherent optical properties (IOPs) are those properties that depend only upon the medium and therefore are independent of the ambient light field. The two fundamental IOPs are the absorption coefficient and the volume scattering function. (These quantities are defined below.) Apparent optical properties (AOPs) are those properties that depend both on the medium (the IOPs) and on the directional structure of the ambient light field, and that display enough regular features and stability to be useful descriptors of a water body. Commonly used AOPs are the irradiance reflectance, the remote-sensing reflectance, and various diffuse attenuation functions.

‘Case 1 waters’ are those in which the contribution by phytoplankton to the total absorption and

scattering is high compared to that by other substances. Absorption by chlorophyll and related pigments therefore plays the dominant role in determining the total absorption in such waters, although covarying detritus and dissolved organic matter derived from the phytoplankton also contribute to absorption and scattering in case 1 waters. Case 1 water can range from very clear (oligotrophic) to very productive (eutrophic) water, depending on the phytoplankton concentration. ‘Case 2 waters’ are ‘everything else,’ namely, waters where inorganic particles or dissolved organic matter from land drainage contribute significantly to the IOPs, so that absorption by pigments is relatively less important in determining the total absorption. Roughly 98% of the world’s open ocean and coastal waters fall into the case 1 category, but near-shore and estuarine case 2 waters are disproportionately important to human interests such as recreation, fisheries, and military operations. Table 1 summarizes the terms, units, and symbols for various quantities frequently used in optical oceanography.

Radiometric Quantities Consider an amount DQ of radiant energy incident in a time interval Dt centered on time t, onto a surface of area DA located at position (x,y,z), and arriving through a set of directions contained in a solid angle DO about the direction (y, j) normal to the area DA, as produced by photons in a wavelength interval Dl centered on wavelength l. The geometry of this situation is illustrated in Figure 1. Then an operational definition of the spectral radiance is Lðx; y; z; t; y; j; lÞ

DQ Dt DA DO Dl ½Js1 m2 sr1 nm1

½1

In the conceptual limit of infinitesimal parameter intervals, the spectral radiance is defined as Lðx; y; z; t; y; j; lÞ

@4Q @t @A @O @l

½2

Spectral radiance is the fundamental radiometric quantity of interest in optical oceanography: it completely specifies the positional (x,y,z), temporal (t), directional (y, j), and spectral (l) structure of the light field. In many oceanic environments, horizontal


619

620

RADIATIVE TRANSFER IN THE OCEAN

Table 1

Quantities commonly used in optical oceanography

Quantity

SI units

Symbol

Radiometric quantities Quantity of radiant energy Power Intensity Radiance Downwelling plane irradiance Upwelling plane irradiance Net irradiance Scalar irradiance Downwelling scalar irradiance Upwelling scalar irradiance Photosynthetic available radiation

J nm1 W nm1 W sr1 nm1 W m2 sr1nm1 W m2 nm1 W m2 nm1 W m2 nm1 W m2 nm1 W m2 nm1 W m2 nm1 Photonss1 m2

Q F I L Ed Eu E Eo Eou Eou PAR

Inherent optical properties Absorption coefficient Volume scattering function Scattering phase function Scattering coefficient Backscatter coefficient Beam attenuation coefficient Single-scattering albedo

m1 m1 sr1 sr1 m1 m1 m1 –

a b b˜ b bb c oo

Apparent optical properties Irradiance reflectance (ratio) Remote-sensing reflectance Attenuation coefficients of radiance L(z, y, j) of downwelling irradiance Ed(z) of upwelling irradiance Eu(z) of PAR

– sr1 m1 m1 m1 m1 m1

R Rrs

ΔQ

ΔΩ

ΔA

x

y

z Figure 1 Geometry used to define radiance.

variations (on a scale of tens to thousands of meters) of the IOPs and the radiance are much less than variations with depth, in which case it can be assumed that these quantities vary only with depth z.

K(y, j) Kd Ku KPAR

(An exception would be the light field due to a single light source imbedded in the ocean; such a radiance distribution is inherently three-dimensional.) Moreover, since the timescales for changes in IOPs or in the environment (seconds to seasons) are much greater than the time required for the radiance to reach steady state (microseconds) after a change in IOPs or boundary conditions, time-independent radiative transfer theory is adequate for most oceanographic studies. (An exception is time-of-flight lidar bathymetry.) When the assumptions of horizontal homogeneity and time independence are valid, the spectral radiance can be written as L(z, y, j, l). Although the spectral radiance completely specifies the light field, it is seldom measured in all directions, both because of instrumental difficulties and because such complete information often is not needed. The most commonly measured radiometric quantities are various irradiances. Suppose the light detector is equally sensitive to photons of a given wavelength l traveling in any direction (y, j) within a hemisphere of directions. If the detector is located at depth z and is oriented facing upward, so as to collect photons traveling downward, then the detector output is a measure of the spectral



downwelling scalar irradiance at depth z, Eod(z, l). Such an instrument is summing radiance over all the directions (elements of solid angle) in the downward hemisphere; thus Eod(z, l) is related to L(z, y, j, l) by Eod ðz; lÞ ¼

ð

Lðz; y; j; lÞ dO

½Wm2 nm1 ½3

2pd

Here 2pd denotes the hemisphere of downward directions (i.e., the set of directions (y, j) such that 0ryrp/2 and 0rjo2p, if y is measured from the þ z or nadir direction). The integral over 2pd can be evaluated as a double integral over y and j after a specific coordinate system is chosen. If the same instrument is oriented facing downward, so as to detect photons traveling upward, then the quantity measured is the spectral upwelling scalar irradiance Eou(z, l). The spectral scalar irradiance Eo(z, l) is the sum of the downwelling and upwelling components: EO ðz; lÞ Eod ðz; lÞ þ Eou ðz; lÞ ð Lðz; y; j; lÞ dO ¼

½4

4p

Eo(z, l) is proportional to the spectral radiant energy density (J m3 nm1) and therefore quantifies how much radiant energy is available for photosynthesis or heating the water. Now consider a detector designed so that its sensitivity is proportional to |cos y|, where y is the angle between the photon direction and the normal to the surface of the detector. This is the ideal response of a ‘flat plate’ collector of area DA, which when viewed at an angle y to its normal appears to have an area of DA|cos y|. If such a detector is located at depth z and is oriented facing upward, so as to detect photons traveling downward, then its output is proportional to the spectral downwelling plane irradiance Ed(z, l). This instrument is summing the downwelling radiance weighted by the cosine of the photon direction, thus Ed ðz; lÞ ¼

ð

Lðz; y; j; lÞjcosyj dO 2pd

½Wm2 nm1

½5

Turning this instrument upside down gives the spectral upwelling plane irradiance Eu(z, l). Ed and Eu are useful because they give the energy flux (power per unit area) across the horizontal surface at depth z owing to downwelling and upwelling photons, respectively. The difference Ed Eu is called the net (or vector) irradiance.

621

Photosynthesis is a quantum phenomenon, i.e., it is the number of available photons rather than the amount of radiant energy that is relevant to the chemical transformations. This is because a photon of, say, l ¼ 400 nm, if absorbed by a chlorophyll molecule, induces the same chemical change as does a photon of l ¼ 600 nm, even though the 400 nm photon has 50% more energy than the 600 nm photon. Only a part of the photon energy goes into photosynthesis; the excess is converted to heat or is re-radiated. Moreover, chlorophyll is equally able to absorb and utilize a photon regardless of the photon’s direction of travel. Therefore, in studies of phytoplankton biology, the relevant measure of the light field is the photosynthetic available radiation, PAR, defined by ð 700 nm lZ Eo ðz; lÞ dl PARðzÞ 350 nm hc ½6 ½photons s1 m2 where h ¼ 6.6255 1034 J s is the Planck constant and c ¼ 3.0 1017 nm s1 is the speed of light. The factor l/hc converts the energy units of Eo to quantum units (photons per second). Bio-optical literature often states PAR values in units of mol photons s1 m2 or einst s1 m2 (where one einstein is one mole of photons).

Inherent Optical Properties Consider a small volume DV of water, of thickness Dr as illuminated by a collimated beam of monochromatic light of wavelength l and spectral radiant power Fi(l) (W nm1), as schematically illustrated in Figure 2. Some part Fa(l) of the incident power Fi(l) is absorbed within the volume of water. Some part Fi(c, l) is scattered out of the beam at an angle c, and the remaining power Ft(l) is transmitted through the volume with no change in direction. Let Fs(l) be the total power that is scattered into all directions. The inherent optical properties usually employed in radiative transfer theory are the absorption and scattering coefficients. In the geometry of Figure 2, the absorption coefficient a(l) is defined as the limit of the fraction of the incident power that is absorbed within the volume, as the thickness becomes small: lim aðlÞ Dr-0

1 Fa ðlÞ 1 ½m Fi ðlÞ Dr

½7

The scattering coefficient b(l) has a corresponding definition using Fs(l). The beam attenuation coefficient c(l) is defined as c(l) ¼ a(l) þ b(l).


622

RADIATIVE TRANSFER IN THE OCEAN ΔΩ

4

Φs ()

10

ΔV _1 _1

Φa

Φt

id

Co

Cle

VSF, (m sr )

Φi

Tu rb

2

10

ha

rbo

r

as

tal

ar

0

10

oc

oc

ea

n

ea

n

_2

10

Δr

Pure sea water Figure 2 Geometry used to define inherent optical properties.

Now take into account the angular distribution of the scattered power, with Fs(c, l)/Fi(l) being the fraction of incident power scattered out of the beam through an angle c into a solid angle DO centered on c, as shown in Figure 2. Then the fraction of scattered power per unit distance and unit solid angle, b(c, l), is lim lim Fs ðc; lÞ bðc; lÞ Dr-0 DO-0 Fi ðlÞDrDO

½m1 sr1 ½8

The spectral power scattered into the given solid angle DO is just the spectral radiant intensity scattered into direction c times the solid angle: Fs(c, l) ¼ Is(c, l) DO. Moreover, if the incident power Fi(l) falls on an area DA, then the corresponding incident irradiance is Ei(l) ¼ Fi(l)/DA. Noting that DV ¼ DrDA is the volume of water that is illuminated by the incident beam gives lim IS ðc; lÞ bðc; lÞ ¼DV-0 Ei ðlÞDV

½9

This form of b(c, l) suggests the name volume scattering function (VSF) and the physical interpretation of scattered intensity per unit incident irradiance per unit volume of water. Figure 3 shows measured VSFs (at 514 nm) from three greatly different water bodies; the VSF of pure water is shown for comparison. VSFs of sea water typically increase by five or six orders of magnitude in going from c ¼ 901 to c ¼ 0.11 for a given water sample, and scattering at a given angle c can vary by two orders of magnitude among water samples. Integrating b(c, l) over all directions (solid angles) gives the total scattered power per unit incident irradiance and unit volume of water, in other words the spectral scattering coefficient: bðlÞ ¼

ð

bðc; lÞdO ¼ 2p 4p

ðp

bðc; lÞ sin c dc

½10

10

_4

0.1

1.0

10.0

100.0

Scattering angle, (deg) Figure 3 Volume scattering functions (VSF) measured in three different oceanic waters. The VSF of pure sea water is shown for comparison.

Eqn. [10] follows because scattering in natural waters is azimuthally symmetric about the incident direction (for unpolarized light sources and randomly oriented scatterers). This integration is often divided into forward scattering, 0rcrp/2, and backward scattering, p/2rcrp, parts. Thus the backscatter coefficient is bb ðlÞ 2p

ðp

bðc; lÞ sin c dc

½11

p=2

The VSFs of Figure 3 have b values ranging from 0.037 to 1.824 m1 and backscatter fractions bb/b of 0.013 to 0.044. The preceding discussion assumed that no inelastic-scattering processes are present. However, inelastic scattering does occur owing to fluorescence by dissolved matter or chlorophyll, and to Raman scattering by the water molecules themselves. Power lost from wavelength l by scattering into wavelength l0 al appears as an increase in the absorption a(l). The gain in power at l0 appears as a source term in the radiative transfer equation. Two more inherent optical properties are commonly used in optical oceanography. The single-scattering albedo is oo(l) ¼ b(l)/c(l). The single-scattering albedo is the probability that a photon will be scattered (rather than absorbed) in any given interaction, hence oo(l) is also known as the probability of photon survival. The volume scattering phase function, ˜ bðc; lÞ is defined by bðc; lÞ

0


bðc; lÞ ½sr1 bðlÞ

½12


Writing the volume scattering function b(c, l) as the product of the scattering coefficient b(l) and the phase ˜ function bðc; lÞ partitions b(c, l) into a factor giving the strength of the scattering, b(l) with units of m1, and a factor giving the angular distribution of the ˜ scattered photons, bðc; lÞ with units of sr1. A striking feature of the sea water VSFs of Figure 3 is that their phase functions are all similar in shape, with the main differences being in the detailed shape of the functions in the backscatter directions (c4901). The IOPs are additive. This means, for example, that the total absorption coefficient of a water body is the sum of the absorption coefficients of water, phytoplankton, dissolved substances, mineral particles, etc. This additivity allows the development of separate models for the absorption and scattering properties of the various constituents of sea water.

The Radiative Transfer Equation The equation that connects the IOPs and the radiance is called the radiative transfer equation (RTE). Even in the simplest situation of horizontally homogeneous water and time independence, the RTE is a formidable integro-differential equation:

dLðz; y; j; lÞ ¼ cðz; lÞLðz; y; j; lÞ cos y dz ð Lðz; y0 ; j0 ; lÞ þ 4p

bðz; y0; j0 -y; j; lÞ dO0 þ Sðz; y; j; lÞ

½13

The scattering angle c in the VSF is the angle between the incident direction (y0 , j0 ) and the scattered direction (y, j). The source term S(z, y, j, l) can describe either an internal light source such as bioluminescence, or inelastically scattered light from other wavelengths. The physical environment of a water body – waves on its surface, the character of its bottom, the incident radiance from the sky – enters the theory via the boundary conditions necessary to solve the RTE. Given the IOPs and suitable boundary conditions, the RTE can be solved numerically for the radiance distribution L(z, y, j, l). Unfortunately, there are no shortcuts to computing other radiometric quantities. For example, it is not possible to write down an equation that can be solved directly for the irradiance Ed; one must first solve the RTE for the radiance and then compute Ed by integrating the radiance over direction.

623

Apparent Optical Properties Apparent optical properties are always a ratio of two radiometric variables. This ratioing removes effects of the magnitude of the incident sky radiance onto the sea surface. For example, if the sun goes behind a cloud, the downwelling and upwelling irradiances within the water can change by an order of magnitude within a few seconds, but their ratio will be almost unchanged. (There will still be some change because the directional structure of the underwater radiance will change when the sun’s direct beam is removed from the radiance incident onto the sea surface.) The ratio just mentioned,

Rðz; lÞ

Eu ðz; lÞ Ed ðz; lÞ

½14

is called the irradiance reflectance (or irradiance ratio). The remote-sensing reflectance Rrs(y, j, l) is defined as

Rrs ðy; j; lÞ

Lw ðy; j; lÞ Ed ðlÞ

½sr1

½15

where Lw is the water-leaving radiance, i.e., the total upward radiance minus the sky and solar radiance that was reflected upward by the sea surface. Lw and Ed are evaluated just above the sea surface. Both Rrs(y, j, l) and R(z, l) just beneath the sea surface are of great importance in remote sensing, and both can be regarded as a measure of ‘ocean color.’ R and Rrs are proportional (to a first-order approximation) to bb/(a þ bb), and measurements of Rrs above the surface or of R within the water can be used to estimate water quality parameters such as the chlorophyll concentration. Under typical oceanic conditions, for which the incident lighting is provided by the sun and sky, the radiance and various irradiances all decrease approximately exponentially with depth, at least when far enough below the surface (and far enough above the bottom, in shallow water) to be free of boundary effects. It is therefore convenient to write the depth dependence of, say, Ed(z, l) as ðz ½16 Ed ðz; lÞ Ed ð0; lÞ exp Kd ðz0 ; lÞdz0 0

where Kd(z, l) is the spectral diffuse attenuation coefficient for spectral downwelling plane irradiance. Solving for Kd(z, l) gives


624


Kd ðz; lÞ ¼

dlnEd ðz; lÞ dz

Organic Particles

Biogenic particles occur in many forms.

1 dEd ðz; lÞ ¼ Ed ðz; lÞ dz

½m1

½17

The beam attenuation coefficient c(l) is defined in terms of the radiant power lost from a collimated beam of photons. The diffuse attenuation coefficient Kd(z, l) is defined in terms of the decrease with depth of the ambient downwelling irradiance Ed(z, l), which comprises photons heading in all downward directions (a diffuse, or uncollimated, light field). Kd(z, l) clearly depends on the directional structure of the ambient light field, hence its classification as an apparent optical property. Other diffuse attenuation coefficients, e.g., Ku, Kod, or KPAR, are defined in an analogous manner, using the corresponding radiometric quantities. In most waters, these K functions are strongly correlated with the absorption coefficient a and therefore can serve as convenient, if imperfect, descriptors of a water body. However, AOPs are not additive, which complicates their interpretation in terms of water constituents.

Optical Constituents of Seawater Oceanic waters are a witch’s brew of dissolved and particulate matter whose concentrations and optical properties vary by many orders of magnitude, so that ocean waters vary in color from the deep blue of the open ocean, where sunlight can penetrate to depths of several hundred meters, to yellowish-brown in a turbid estuary, where sunlight may penetrate less than a meter. The most important optical constituents of sea water can be briefly described as follows. Sea Water

Water itself is highly absorbing at wavelengths below 250 nm and above 700 nm, which limits the wavelength range of interest in optical oceanography to the near-ultraviolet to the near infrared. Dissolved Organic Compounds

These compounds are produced during the decay of plant matter. In sufficient concentrations these compounds can color the water yellowish brown; they are therefore generally called yellow matter or colored dissolved organic matter (CDOM). CDOM absorbs very little in the red, but absorption increases rapidly with decreasing wavelength, and CDOM can be the dominant absorber at the blue end of the spectrum, especially in coastal waters influenced by river runoff.

Bacteria Living bacteria in the size range 0.2– 1.0 mm can be significant scatterers and absorbers of light, especially at blue wavelengths and in clean oceanic waters, where the larger phytoplankton are relatively scarce. Phytoplankton These ubiquitous microscopic plants occur with incredible diversity of species, size (from less than 1 mm to more than 200 mm), shape, and concentration. Phytoplankton are responsible for determining the optical properties of most oceanic waters. Their chlorophyll and related pigments strongly absorb light in the blue and red and thus, when concentrations are high, determine the spectral absorption of sea water. Phytoplankton are generally much larger than the wavelength of visible light and can scatter light strongly. Detritus Nonliving organic particles of various sizes are produced, for example, when phytoplankton die and their cells break apart, and when zooplankton graze on phytoplankton and leave cell fragments and fecal pellets. Detritus can be rapidly photooxidized and lose the characteristic absorption spectrum of living phytoplankton, leaving significant absorption only at blue wavelengths. However, detritus can contribute significantly to scattering, especially in the open ocean. Inorganic Particles

Particles created by weathering of terrestrial rocks can enter the water as wind-blown dust settles on the sea surface, as rivers carry eroded soil to the sea, or as currents resuspend bottom sediments. Such particles range in size from less than 0.1 mm to tens of micrometers and can dominate water optical properties when present in sufficient concentrations. Particulate matter is usually the major determinant of the absorption and scattering properties of sea water and is responsible for most of the temporal and spatial variability in these optical properties. A central goal of research in optical oceanography is to understand how the absorption and scattering properties of these various constituents relate to the particle type (e.g., microbial species or mineral composition), present conditions (e.g., the physiological state of a living microbe, which in turn depends on nutrient supply and ambient lighting), and history (e.g., photo-oxidation of pigments in dead cells). Biogeo-optical models have been developed that attempt (with varying degrees of success) to predict the IOPs



in terms of the chlorophyll concentration or other simplified measures of the composition of a water body.

Examples of Underwater Light Fields Solving the radiative transfer equation requires mathematically sophisticated and computationally intensive numerical methods. Hydrolight is a widely used software package for numerical solution of oceanographic radiative transfer problems. The input to Hydrolight consists of the absorption and scattering coefficients of each constituent of the water body (microbial particles, dissolved substances, mineral particles, etc.) as functions of depth and wavelength, the corresponding scattering phase functions, the sea state, the sky radiance incident onto the sea surface, and the reflectance properties of the bottom boundary (if the water is not assumed infinitely deep). Hydrolight solves the onedimensional, time-independent radiative transfer equation, including inelastic scattering effects, to obtain the radiance distribution L(z, y, j, l). Other quantities of interest such as irradiances or reflectances are then computed using their definitions and the solution radiance distribution. To illustrate the range of behavior of underwater light fields, Hydrolight was run for two greatly different water bodies. The first simulation used a chlorophyll profile measured in the Atlantic Ocean north of the Azores in winter. The water was well mixed to a depth of over 100 m. The chlorophyll concentration Chl varied between 0.2 and 0.3 mg m3 between the surface and 116 m depth; it

then dropped to less than 0.05 mg m3 below 150 m depth. The water was oligotrophic, case 1 water, and commonly used bio-optical models for case 1 water were used to convert the chlorophyll concentration to absorption and scattering coefficients (which were not measured). A scattering phase function similar in shape to those seen in Figure 3 was used for the particles; this phase function had a backscatter fraction of bb/b ¼ 0.018. The second simulation was for an idealized, case 2 coastal water body containing 5 mg m3 of chlorophyll and 2 g m3 of brown-colored mineral particles representing resuspended sediments. Bio-optical models and measured massspecific absorption and scattering coefficients were used to convert the chlorophyll and mineral concentrations to absorption and scattering coefficients. The large microbial particles of low index of refraction were assumed to have a phase function with bb/ b ¼ 0.005, and the small mineral particles of high index of refraction had bb/b ¼ 0.03. The water was assumed to be well mixed and to have a brown mud bottom at a depth of 10 m. Both simulations used a clear sky radiance distribution appropriate for midday in January at the Azores location. The sea surface was covered by capillary waves corresponding to a 5 m s1 wind speed. Figure 4 shows the component and total absorption coefficients just beneath the sea surface for these two hypothetical water bodies, and Figure 5 shows the corresponding scattering coefficients. For the case 1 water, the total absorption is dominated by chlorophyll at blue wavelengths and by the water itself at wavelengths greater than 500 nm. However, the water makes only a small contribution to the

Case 2

Case 1 0.08

0.8

0.06

0.6

_1

Absorption coefficient, a (m )

625

Total 0.04

0.4 Total

0.02

Chl

Min

Water

Water

0.2

Chl CDOM

CDOM 0 400

500

600

Wavelength, (nm)

700

0 400

500

600

700

Wavelength, (nm)

Figure 4 Absorption coefficients for the case 1 and case 2 water bodies. The contributions by the various components are labeled.


626


Case 2

Case 1

0.30

3.0 Total

_1

Scattering coefficient, b (m )

0.25 2.0

0.20 Total

Minerals

0.15

Chl

Chl

0.10

1.0

0.05 Water

Water

0

0 400

500

600

700

500

400

Wavelength, (nm)

600

700

Wavelength, (nm)

Figure 5 Scattering coefficients for the case 1 and case 2 water bodies. The contributions by the various components are labeled (CDOM is nonscattering).

log [radiance, L (W m sr nm )]

th 0

_1

_1


Dep

_2

_1 _2

_1 _2 _3 _4 0

_ 90

90 0 )

0

700 00 6 nm) 500 ,( h t 400 g elen Wav

)

eg

eg

(d

(d

v

v

90

n,

n,

700 600 ) 0 50 (nm th, 400 g n e el Wav

io

io

1

ct

ct

0

re

80

re

18

di

di

_8 0

g

g

_7

in

_ 90

in

_6

ew

ew

_5

Vi

Vi

_4

_1

0

00 m

th 1

Dep

Figure 6 The case 1 water radiance distribution in the azimuthal plane of the sun at depth 0 (just below the sea surface) and at 100 m.

total scattering. In the case 2 water, absorption by the mineral particles is comparable to or greater than that by the chlorophyll-bearing particles, and water dominates only in the red. The mineral particles are the primary scatterers. Figures 6 and 7 show the radiance in the azimuthal plane of the sun as a function of polar viewing direction and wavelength, for selected depths. For the case 1 simulation (Figure 6), the depths shown are zero, just beneath the sea surface, and 100 m; for the

case 2 simulation (Figure 7), the depths are zero and 10 m, which is at the bottom. Note that the radiance axis is logarithmic. A viewing direction of yv ¼ 0 corresponds to looking straight down and seeing the upwelling radiance (photons traveling straight up). Near the sea surface, the angular dependence of the radiance distribution is complicated because of boundary effects such as internal reflection (the bumps near yv ¼ 901, which is radiance traveling horizontally) and refraction of the sun’s direct beam



th 0

_1


_1


Dep

_1

m

_3

_2

_2

_1

1

th 10

Dep

_2

627

0 _1 _2

g in ew Vi

_3 0 _ 90

Vi

_4 _5 _6 0 _ 90

ew

in

g

c re di

180

re

700 600 m) 0 50 (n 400 length, e v a W

0

,

n tio

90

180

di

ct

90

io

n,

v

0

v

(d

(d

eg

eg

700 600 m) 0 n 0 ( 5 400 length, e Wav

)

) Figure 7 The case 2 water radiance distribution in the azimuthal plane of the sun at depth 0 (just below the sea surface) and at 10 m.

10

Case 1

2

0

10

Case 2

1

Depth = 0

Depth = 0 10

0

5m

_

_1

Scalar irradance Eo (W m 2 nm )

10

_2

100 m

_4

200 m

10

_1

10

10 m 10

_2

10

_3

_6

10

10

_4

_8

10 10

10

_5

_ 10

400

500

600 700 Wavelength, (nm)

10

400

500 600 700 Wavelength, (nm)

Figure 8 The scalar irradiance Eo at selected depths for the case 1 and case 2 waters.

(the large spike near yv ¼ 1401). As the depth increases, the angular shape of the radiance distribution smooths out as a result of multiple scattering. By 100 m in the case 1 simulation, the shape of the radiance distribution is approaching its asymptotic shape, which is determined only by the IOPs. In the case 2 simulation, the upwelling radiance ( 901ryvr901) at the bottom is isotropic; this is a consequence of having assumed the mud bottom to be a Lambertian reflecting surface. As the depth increases, the color of the radiance becomes blue for the case 1 water and greenish-yellow for the case 2

water. In the case 1 simulation at 100 m, there is a prominent peak in the radiance near 685 nm, even though the solar radiance has been filtered out by the strong absorption by water at red wavelengths. This peak is due to chlorophyll fluorescence, which is transferring energy from blue to red wavelengths, where it is emitted isotropically. As already noted, the extensive information contained in the full radiance distribution is seldom needed. A biologist would probably be interested only in the scalar irradiance Eo, which is shown at selected depths in Figure 8. This irradiance was computed by


628


radiance distributions given the IOPs and boundary conditions. The difficult science lies in learning how to predict the IOPs for the incredible variety of water constituents and environmental conditions found in the world’s oceans, and in learning how to interpret measurements such as Rrs. The development of biogeo-optical models for case 2 waters, in particular, is a research topic for the next decades.

_1

Remote-sensing reflectance, Rrs (sr )

0.015 Case 2 Case 1 0.010

0.005

See also

0 400

500

600

700

Wavelength, (nm)

Figure 9 The remote-sensing reflectance Rrs for the case 1 and case 2 waters.

integrating the radiance over all directions. Although the irradiances near the surface are almost identical, the decay of these irradiances with depth is much different in the case 1 and case 2 waters. The remote-sensing reflectance Rrs, the quantity of interest for ‘ocean color’ remote sensing, is shown in Figure 9 for the two water bodies. The shaded bars at the bottom of the figure show the nominal SeaWiFS sensor bands. The SeaWiFS algorithm for retrieval of the chlorophyll concentration uses a function of the ratio Rrs(490 nm)/Rrs (555 nm). When applied to these Rrs spectra, the SeaWiFS algorithm retrieves a value of Chl ¼ 0.24 mg m3 for the case 1 water, which is close to the average value of the measured profile over the upper few tens of meters of the water column. However, when applied to the case 2 spectrum, the SeaWiFS algorithm gives Chl ¼ 8.88 mg m3, which is almost twice the value of 5.0 mg m3 used in the simulation. This error results from the presence of the mineral particles, which are not accounted for in the SeaWiFS chlorophyll retrieval algorithm. These Hydrolight simulations highlight the fact that it is now possible to compute accurate underwater

Bacterioplankton. Bio-Optical Models.

Further Reading Bukata RP, Jerome JH, Kondratyev KY, and Pozdnyakov DV (1995) Optical Properties and Remote Sensing of Inland and Coastal Waters. New York: CRC Press. Caimi FM (ed.) (1995) Selected Papers on Underwater Optics. SPIE Milestone Series, vol. MS 118. Bellingham, WA: SPIE Optical Engineering Press. Jerlov NG (1976) Marine Optics. Amsterdam: Elsevier. Kirk JTO (1994) Light and Photosynthesis in Aquatic Ecosystems, 2nd edn. New York: Cambridge University Press. Mobley CD (1994) Light and Water: Radiative Transfer in Natural Waters. San Diego: Academic Press. Mobley CD (1995) The optical properties of water. In Bass M (ed.) Handbook of Optics, 2nd edn, vol. I. New York: McGraw Hill. Mobley CD and Sundman LK (2000) Hydrolight 4.1 Users’ Guide. Redmond, WA: Sequoia Scientific. [See also www.sequoiasci.com/hydrolight.html] Shifrin KS (1988) Physical Optics of Ocean Water. AIP Translation Series. New York: American Institute of Physics. Spinrad RW, Carder KL, and Perry MJ (1994) Ocean Optics. New York: Oxford University Press. Walker RE (1994) Marine Light Field Statistics. New York: Wiley.


RADIOACTIVE WASTES Copyright & 2001 Elsevier Ltd.

local marine environment through diffuse outlets like muncipal sewage systems and rivers.


Disposal at Sea

L. Føyn, Institute of Marine Research, Bergen, Norway

Introduction The discovery and the history of radioactivity is closely connected to that of modern science. In 1896 Antoine Henri Becquerel observed and described the spontaneous emission of radiation by uranium and its compounds. Two years later, in 1898, the chemical research of Marie and Pierre Curie led to the discovery of polonium and radium. In 1934 Fre´de´ric Joliot and Ire`ne Curie discovered artificial radioactivity. This discovery was soon followed by the discovery of fission and the enormous amounts of energy released by this process. However, few in the then limited community of scientists working with radioactivity believed that it would be possible within a fairly near future to establish enough resources to develop the fission process for commercial production of energy or even think about the development of mass-destruction weapons. World War II made a dramatic change to this. The race that began in order to be the first to develop mass-destruction weapons based on nuclear energy is well known. Following this came the development of nuclear reactors for commercial production of electricity. From the rapidly growing nuclear industry, both military and commercial, radioactive waste was produced and became a problem. As with many other waste problems, discharges to the sea or ocean dumping were looked upon as the simplest and thereby the best and final solution.

The Sources Anthropogenic radioactive contamination of the marine environment has several sources: disposal at sea, discharges to the sea, accidental releases and fallout from nuclear weapon tests and nuclear accidents. In addition, discharge of naturally occurring radioactive materials (NORM) from offshore oil and gas production is a considerable source for contamination. The marine environment receives in addition various forms of radioactive components from medical, scientific and industrial use. These contributions are mostly short-lived radionuclides and enter the

The first ocean dumping of radioactive waste was conducted by the USA in 1946 some 80 km off the coast of California. The International Atomic Energy Agency (IAEA) published in August 1999 an ‘Inventory of radioactive waste disposal at sea’ according to which the disposal areas and the radioactivity can be listed as shown in Table 1. Figure 1 shows the worldwide distribution of disposal-points for radioactive waste. The majority of the waste disposed consists of solid waste in various forms and origin, only 1.44% of the total activity is contributed by low-level liquid waste. The disposal areas in the north-east Atlantic and the Arctic contain about 95% of the total radioactive waste disposed at sea. Most disposal of radioactive waste was performed in accordance with national or international regulations. Since 1967 the disposals in the northeast Atlantic were for the most part conducted in accordance with a consultative mechanism of the Organization for Economic Co-operation and Development/Nuclear Energy Agency (OECD/NEA). The majority of the north-east Atlantic disposals were of low-level solid waste at depths of 1500– 5000 m, but the Arctic Sea disposals consist of various types of waste from reactors with spent fuel to containers with low-level solid waste dumped in fairly shallow waters ranging from about 300 m depth in the Kara Sea to less than 20 m depth in some fiords on the east coast of Novaya Zemlya. Most of the disposals in the Arctic were carried out by the former Soviet Union and were not reported internationally. An inventory of the USSR disposals was presented by the Russian government in 1993. Already before this, the good collaboration between Russian and Norwegian authorities had led Table 1

Worldwide disposal at sea of radioactive wastea

North-west Atlantic Ocean North-west Atlantic Ocean Arctic Ocean North-east Pacific Ocean West Pacific Ocean

2.94 PBq 2.94 PBq 38.37 PBq 0.55 PBq 0.89 PBq

a PBq (petaBq) ¼ 1015 Bq (1 Bq1 ¼ disintegration s 1). The old unit for radioactivity was Curie (Ci); 1 Ci3.7 1010 Bq.


629

630

RADIOACTIVE WASTES

North-east Pacific 0.55 PBq

North-west Atlantic 2.94 PBq

North-east Atlantic 42.32 PBq

Arctic 38.37 PBq

West Pacific 0.89 Pbq

Figure 1 The worldwide location for disposal of radioactive waste at sea.

to a joint Norwegian–Russian expedition to the Kara Sea in 1992 followed by two more joint expeditions, in 1993 and 1994. The main purpose of these expeditions was to locate and inspect the most important dumped objects and to collect samples for assessing the present environmental impact and to assess the possibility for potential future leakage and environmental impacts. These Arctic disposals differ significantly from the rest of the reported sea disposals in other parts of the world oceans as most of the dumped objects are found in shallow waters and some must be characterized as high-level radioactive waste, i.e. nuclear reactors with fuel. Possible releases from these sources may be expected to enter the surface circulation of the Kara Sea and from there be transported to important fisheries areas in the Barents and Norwegian Seas. Figures 2 and 3 give examples of some of the radioactive waste dumped in the Arctic. Pictures were taken with a video camera mounted on a ROV (remote operated vehicle). The ROV was also equipped with a NaI-detector for gamma-radiation measurements and a device for sediment sampling close to the actual objects. The Global Convention on the Prevention of Marine Pollution by Dumping of Wastes and Other Matter was adopted by an Intergovernmental Conference in London in 1972. The convention named the London Convention 1972, formerly the London Dumping Convention (LDC), addressed from the

very beginning the problem of radioactive waste. But it was not until 20 February 1994 that a total prohibition on radioactive waste disposal at sea came into force. Discharges to the Sea

Of the total world production of electricity about 16% is produced in nuclear power plants. In some countries nuclear energy counts for the majority of the electricity produced, France 75% and Lithuania 77%, and in the USA with the largest production of nuclear energy of more than 96 000 MWh this accounts for about 18% of the total energy production. Routine operations of nuclear reactors and other installations in the nuclear fuel cycle release small amounts of radioactive material to the air and as liquid effluents. However, the estimated total releases of 90Sr, 131I and 137Cs over the entire periods of operation are negligible compared to the amounts released to the environment due to nuclear weapon tests. Some of the first reactors that were constructed used a single-pass cooling system. The eight reactors constructed for plutonium production at Hanford, USA, between 1943 and 1956, pumped water from Columbia River through the reactor cores then delayed it in cooling ponds before returning it to the river. The river water and its contents of particles and components were thereby exposed to a great neutron


RADIOACTIVE WASTES

(A)

(B)

(C)

(D)

631

Figure 2 (A)–(D) Pictures of a disposed submarine at a depth of c. 30 m in the Stepovogo Fiord, east coast of Novaya Zemyla. The submarine contains a sodium-cooled reactor with spent fuel. Some of the hatches of the submarine are open which allows for ‘free’ circulation of water inside the vessel.

(A)

(B)

Figure 3 (A) Containers of low level solid waste at the bottom of the Abrosimov Fiord at a depth of c. 15 m and (B) a similar container found washed ashore.


632

RADIOACTIVE WASTES

flux and various radioactive isotopes were created. In addition corrosion of neutron-activated metal within the reactor structure contributed to the radioactive contamination of the cooling water. Only a limited number of these radionuclides reached the river mouth and only 32P, 51Cr, 54Mn and 65Zn were detected regularly in water, sediments and marine organisms in the near-shore coastal waters of the US Pacific Northwest. Reactors operating today all have closed primary cooling systems that do not allow for this type of contamination. Therefore, under normal conditions production of electricity from nuclear reactors does not create significant amounts of operational discharges of radionuclides. However, the 434 energyproducing nuclear plants of the world in 1998 created radioactive waste in the form of utilized fuel. Utilized fuel is either stored or reprocessed. Only 4–5% of the utilized nuclear fuel worldwide is reprocessed. Commercial, nonmilitary, reprocessing of nuclear fuel takes place in France, Japan, India and the United Kingdom. Other reprocessing plants defined as defense-related are in operation and producing waste but without discharges. For example in the USA, at the Savannah River Plant and the Hanford complex, about 83 000 m3 and 190 000 m3, respectively, of high-level liquid waste was in storage in 1985. Reprocessing plants and the nuclear industry in the former Soviet Union have discharged to the Ob and Yenisey river systems ending up in the Arctic ocean. In 1950–51 about 77 106 m3 liquid waste of 100 PBq was discharged to the River Techa. The Techa River is connected to the River Ob as is the Tomsk River where the Tomsk-7, a major production site for nuclear weapons plutonium, is situated. Other nuclear plants, such as the Krasnoyarsk industrial complex, have discharged to the Yenisey river. Large amounts of radioactive waste are also stored at the sites. Radioactive waste stored close to rivers has the potential of contaminating the oceans should an accident happen to the various storage facilities. The commercial reprocessing plants in France at Cap de la Hague and in the UK at Sellafield have for many years, and still do, contributed to the radioactive contamination of the marine environment. They both discharge low-level liquid radioactive effluents to the sea. Most important, however, these discharges and their behavior in the marine environment have been and are still thoroughly studied and the results are published in the open literature. The importance of these discharges is extensive as radionuclides from Sellafield and la Hague are traced throughout the whole North

Table 2 Total discharges of some radionuclides from Sellafield 1952–92 3

H Sr 134 Cs 137 Cs 238 Pu 239 Pu 241 Pu 241 Am 90

39 PBq 6.3 PBq 5.8 PBq 41.2 PBq 0.12 PBq 0.6 PBq 21.5 PBq 0.5 PBq

Atlantic. Most important is Sellafield; Table 2 summarizes the reported discharge of some important radionuclides. In addition a range of other radionuclides have been discharged from Sellafield, but prior to 1978 the determination of radionuclides was, for many components, not specific. Technetium (99Tc) for instance was included in the ‘total beta’ determinations with an estimated annual discharge from 1952 to 1970 below 5 TBq and from 1970 to 1977 below 50 TBq. Specific determination of 99Tc in the effluents became part of the routine in 1978 when about 180 TBq was discharged followed by about 50 TBq in 1979 and 1980 and then an almost negligible amount until 1994. The reason for mentioning 99Tc is that this radionuclide, in an oceanographic context, represents an almost ideal tracer in the oceans. Technetium is most likely to be present as pertechnetate, TcO4 , totally dissolved in seawater; it acts conservatively and moves as a part of the water masses. In addition the main discharges of technetium originate from point sources with good documentation of time for and amount of the release. The discharges from Sellafield are a good example of this. From 1994 the UK authorities have allowed for a yearly 99 Tc discharge of up to 200 TBq. Based on surveys before and after the discharges in 1994, 30 TBq (March–April) and 32 TBq (September–October), the transit time for technetium from the Irish Sea to the North Sea was calculated to be considerably faster than previous estimations of transit times for released radionuclides. This faster transport is demonstrated by measurements indicating that the first discharge plume of 99Tc had reached the south coast of Norway before November 1996 in about 2.5 years compared to the previously estimated transit time of 3–4 years. Other reprocessing plants may have discharges to the sea, but without a particular impact in the world oceans. The reprocessing plant at Trombay, India, may, for example, be a source for marine contamination.


RADIOACTIVE WASTES

Accidental Releases

Accidents resulting in direct radioactive releases to the sea are not well known as most of them are connected to wreckage of submarines. Eight nuclear submarines with nuclear weapons have been reported lost at sea, two US and six former USSR. The last known USSR wreck was the submarine Komsomolets which sank in the Norwegian Sea southwest of Bear Island, on 7 April 1989. The activity content in the wreck is estimated by Russian authorities to be 1.55–2.8 PBq 90Sr and 2.03–3 PBq 137 Cs and the two nuclear warheads on board may contain about 16 TBq 239,240Pu equivalent to 6–7 kg plutonium. Other estimates indicate that each warhead may contain 10 kg of highly enriched uranium or 4–5 kg plutonium. On August 12th 2000, the Russian nuclear submarine Kursk sank at a depth of 108 meters in the Barents Sea north of the Kola peninsula. Vigorous explosions in the submarine’s torpedo-chambers caused the wreckage where 118 crew-members were entrapped and lost their lives. Kursk, and Oscar II attack submarine, was commissioned in 1995 and was powered by two pressurized water reactors. Kursk had no nuclear weapons on board. Measurements close to the wreck in the weeks after the wreckage showed no radioactive contamination indicating that the primary cooling-systems were not damaged in the accident. A rough inventory calculation estimates that the reactors at present contain about 56 000 TBq. Russian authorities are planning for a salvage operation where the submarine or part of the submarine will be lifted from the water and transported to land. Both a possible salvage operation or to leave the wreck where it is will be create a demand for monitoring as the location of the wreck is within important fishing grounds. The wreckage of Komsomolets in 1989 and the attempts to raise money for an internationally financed Russian led salvage operation became very public. The Russian explanation for the intensive attempts of financing the salvage was said to be the potential for radioactive pollution. The wreck of the submarine is, however, located at a depth of 1658 m and possible leaching of radionuclides from the wreck will, due to the hydrography of the area, hardly have any vertical migration and radioactive components will spread along the isopycnic surfaces gradually dispersing the released radioactivity in the deep water masses of the Nordic Seas. An explanation for the extensive work laid down for a salvage operation and for what became the final solution, coverage of the torpedo-part of the hull, may be that this submarine was said to be able to fire its torpedo

633

missiles with nuclear warheads from a depth of 1000 m. In 1990, the Institute of Marine Research, Bergen, Norway, started regular sampling of sediments and water close to the wreck of Komsomolets. Values of 137 Cs were in the range 1–10 Bq per kg dry weight sediment and 1–30 Bq per m3 water. No trends were found in the contamination as the variation between samples taken at the same date were equal to the variation observed from year to year. Detectable amounts of 134Cs in the sediment samples indicate that there is some leaching of radioactivity from the reactor. Accidents with submarines and their possible impact on the marine environment are seldom noticed in the open literature and there is therefore little common knowledge available. An accident, however, that is well known is the crash of a US B-52 aircraft, carrying four nuclear bombs, on the ice off Thule air base on the northwest coast of Greenland in January 1968. Approximately 0.4 kg plutonium ended up on the sea floor at a depth of 100–300 m. The marine environment became contaminated by about 1 TBq 239,240 Pu which led to enhanced levels of plutonium in benthic animals, such as bivalves, sea-stars and shrimps after the accident. This contamination has decreased rapidly to the present level of one order of magnitude below the initial levels. Fallout from Nuclear Weapon Tests and Nuclear Accidents

Nuclear weapon tests in the atmosphere from 1945 to 1980 have caused the greatest man-made release of radioactive material to the environment. The most intensive nuclear weapon tests took place before 1963 when a test-ban treaty signed by the UK, USA and USSR came into force. France and China did not sign the treaty and continued some atmospheric tests, but after 1980 no atmospheric tests have taken place. It is estimated that 60% of the total fallout has initially entered the oceans, i.e. 370 PBq 90Sr, 600 PBq 137Cs and 12 PBq 239,240Pu. Runoff from land will slightly increase this number. As the majority of the weapon tests took place in the northern hemisphere the deposition there was about three times as high as in the southern hemisphere. Results from the GEOSECS expeditions, 1972–74, show a considerable discrepancy between the measured inventories in the ocean of 900 PBq 137Cs, 600 PBq 90Sr and 16 PBq 239,240Pu and the estimated input from fallout. The measured values are far higher than would be expected from the assumed fallout data. Thus the exact input of anthropogenic radionuclides may be partly unknown or the geographical


634

RADIOACTIVE WASTES

coverage of the measurements in the oceans were for some areas not dense enough for accurate calculations. Another known accident contributing to marine contamination was the burn-up of a US satellite (SNAP 9A) above the Mozambique channel in 1964 which released 0.63 PBq 238Pu and 0.48 TBq 239 Pu; 73% was eventually deposited in the southern hemisphere. The Chernobyl accident in 1986 in the former USSR is the latest major event creating fallout to the oceans. Two-thirds of the c.100 PBq 137Cs released was deposited outside the Soviet Union. The total input to the world oceans of 137Cs from Chernobyl is estimated to be from 15–20 Pbq, i.e. 4.5 PBq in the Baltic Sea; 3–5 PBq in the Mediterranean Sea, 1.2 PBq in the North Sea and about 5 PBq in the northeast Atlantic. Natural Occurring Radioactive Material

Oil and gas production mobilize naturally occurring radioactive material (NORM) from the deep underground reservoir rock. The radionuclides are primarily 226Ra, 228Ra and 210Pb and appear in sludge and scales and in the produced water. Scales and sludge containing NORM represent an increasing amount of waste. There are different national regulations for handling this type of waste. In Norway, for example, waste containing radioactivity above 10 Bq g1 is stored on land in a place specially designed for this purpose. However, there are reasons to believe that a major part of radioactive contaminated scales and sludge from the worldwide offshore oil and gas production are discharged to the sea. Reported NORM values in scales are in the ranges of 0.6–57.2 Bq g1 226Ra þ 228Ra (Norway), 0.4–3700 Bq g1 (USA) and 1–1000 Bq g1 226Ra (UK). Scales are an operational hindrance in oil and gas production. Frequent use of scale-inhibitors reduce the scaling process but radioactive components are released to the production water adding to its already elevated radioactivity. More than 90% of the radioactivity in produced water is due to 226Ra and 228 Ra having a concentration 100–1000 times higher than normal for seawater. The discharge of produced water is a continuous process and the amount of water discharged is considerable and increases with the age of the production wells. As an example, the estimated amount of produced water discharged to the North Sea in 1998 was 340 million m3 and multiplying by an average value of 5 Bq1 of 226Ra in produced water, the total input of 226Ra to the North Sea in 1998 was 1.7 TBq.

Discussion The total input of anthropogenic radioactivity to the world’s oceans is not known exactly, but a very rough estimate gives the following amounts: 85 PBq dumped, 100 PBq discharged from reprocessing and 1500 PBq from fallout. Some of the radionuclides have very long half-lives and will persist in the ocean, for example 99Tc has a half-life of 2.1 105 years, 239,240 Pu, 2.4 104 years and 226Ra, 1600 years. 137 Cs, 90Sr and 228Ra with half-lives of 30 years, 29 years and 5.75 years, respectively, will slowly decrease depending on the amount of new releases. In an oceanographic context it is worth mentioning the differences in denomination between radioactivity and other elements in the ocean. The old denomination for radioactivity was named after Curie (Ci) and 1 g radium was defined to have a radioactivity of 1 Ci; 1 Ci 3.7 1010 Bq and 1 PBq 27 000 Ci. Therefore released radioactivity of 1 PBq can be compared to the radioactivity of 27 kg radium. The common denominations for major and minor elements in seawater are given in weight per volume. For comparison if 1 PBq or 27 kg radium were diluted in 1 km3 of seawater, this would give a radium concentration of 0.027 mg l1 or 1000 Bq l1. Calculations like this clearly visualize the sensitivity of the analytical methods used for measuring radioactivity. In the Atlantic Ocean for example radium (228Ra) has a concentration of 0.017–3.40 mBq l1, whereas 99Tc measured in surface waters off the southwest coast of Norway is in the range of 0.9–6.5 mBq l1. Measured in weight the total amount of radionuclides do not represent a huge amount compared to the presence of nonradioactive components in seawater. The radioisotopes of cesium and strontium are both important in a radioecological context since they have chemical behavior resembling potassium and calcium, respectively. Cesium follows potassium in and out of the soft tissue cells whereas strontium follows calcium into bone cells and stays. Since uptake and release in organisms is due to the chemical characteristics and rarely if the element is radioactive or not, radionuclides such as 137Cs and 90 Sr have to compete with the nonradioactive isotopes of cesium and strontium. Oceanic water has a cesium content of about 0.5lmg1 and a strontium content of about 8000 mg l1. Uptake in a marine organism is most likely to be in proportion to the abundance of the radioactive and the nonradioactive isotopes of the actual element. This can be illustrated by the following example. The sunken nuclear submarine Komsomolets contained an estimated (lowest) amount of 1.55 PBq 90Sr (about 300 g) and 2.03 PBq


RADIOACTIVE WASTES 137

Cs (about 630 g). If all this was released at once and diluted in the immediate surrounding 1 km3 of water the radioactive concentration would have been 1550 Bq l1 for 90Sr and 2030 Bq l1 for 137Cs, the concentration in weight per volume would have been 0.000 3 mg l1 90Sr and 0.000 63 mg l1 137Cs. This means that even if the radioactive material was kept in the extremely small volume of 1 km3, compared to the volume of the deep water of the Norwegian Sea available for a primary dilution, the proportion of radioactive to nonradioactive isotopes of strontium and cesium, available for uptake in marine organisms, would have been about 2.7 106 and 7.9 105, respectively. From the examples above it can be seen that if uptake, and thereby impact, in marine organisms follows regular chemical–physiological rules there is a ‘competition’ in seawater in favor of the nonradioactive isotopes for elements normally present in seawater. Measurable amount of radionuclides of cesium and strontium are detected in marine organisms but at levels far below the concentrations in freshwater fish. Average concentrations of 137Cs in fish from the Barents Sea during the period with the most intensive nuclear weapon tests in that area, 1962–63, never exceeded 90 Bq kg1 fresh weight, whereas fallout from Chernobyl resulted in concentrations in freshwater fish in some mountain lakes in Norway far exceeding 10 000 Bq kg1. For radionuclides like technetium and plutonium, which will persist in the marine environment, uptake will be based only on the actual concentrations in seawater of radionuclide. The levels of 99Tc, for example, increased in seaweed (Fucus vesiculosus) from 70 Bq per kg dry weight (December 1997) to 124 Bq kg1 in January 1998 in northern Norway which reflected the increased concentration in the water as the peak of the technetium plume from Sellafield reached this area. Previously the effects of anthropogenic radioactivity have been based on the possible dose effect to humans. Most of the modeling work has been concentrated on assessing the dose to critical population groups eating fish and other marine organisms. But even if the radiation from anthropogenic radionuclides to marine organisms is small compared to natural radiation from radionuclides like potassium, 40 K, the presence of additional radiation may give a chronic exposure with possible effects, at least on individual marine organisms. The input of radioactivity, NORM, from the offshore oil and gas production may also give reason for concern. The input will increase as it is a continuous part of the production. Even if radium as the main radionuclide is not likely to be taken up by marine

635

organisms the use of chemicals like scale inhibitors may change this making radium more available for marine organisms.

Conclusion The sea began receiving radioactive waste from anthropogenic sources in 1946, in a rather unregulated way in the first decades. Both national and international regulations controlling disposals have now slowly come into force. Considerable amounts are still discharged regularly from nuclear industries and the practice of using the sea as a suitable wastebasket is likely to continue for ever. In 1994 an international total prohibition on radioactive waste disposal at sea came into force, but the approximately 85 PBq of solid radioactive waste that has already been dumped will sooner or later be gradually released to the water masses. Compared to other wastes disposed of at sea the amount of radioactive waste by weight is rather diminutive. However, contrary to most of the ‘ordinary’ wastes in the sea, detectable amounts of anthropogenic radioactivity are found in all parts of the world oceans and will continue to contaminate the sea for many thousands of years to come. This means that anthropogenic radioactive material has become an extra chronic radiation burden for marine organisms. In addition, the release of natural occurring radionuclides from offshore oil and gas production will gradually increase the levels of radium, in particular, with a possible, at present unknown, effect. However, marine food is not, and probably never will be, contaminated at a level that represents any danger to consumers. The ocean has always received debris from human activities and has a potential for receiving much more and thereby help to solve the waste disposal problems of humans. But as soon as a waste product is released and diluted in the sea it is almost impossible to retrieve. Therefore, in principal, no waste should be disposed of in the sea without clear documentation that it will never create any damage to the marine environment and its living resources. This means that with present knowledge no radioactive wastes should be allowed to be released into the sea.

See also International Organizations. Nuclear Fuel Reprocessing and Related Discharges. Single Compound Radiocarbon Measurements. UraniumThorium Decay Series in the Oceans Overview.


636

RADIOACTIVE WASTES

Further Reading Guary JC, Guegueniat P, and Pentreath RJ (eds.) (1988) Radionuclides: A Tool for Oceanography. London, New York: Elsevier Applied Science. Hunt GJ, Kershaw PJ, and Swift DJ (eds.) (1998) Radionuclides in the oceans (RADOC 96–97). Distribution, Models and Impacts. Radiation Protection Dosimetry 75: 1--4.

IAEA (1995) Environmental impact of radioactive releases; Proceedings of an International Symposium on Environmental Impact of Radioactive Releases. Vienna: International Atomic Energy Agency. IAEA (1999) Inventory of Radioactive Waste Disposals at Sea. IAEA-TECDOC-1105 Vienna: International Atomic Energy Agency. pp. 24 A.1–A.22.


RADIOCARBON R. M. Key, Princeton University, Princeton, NJ, USA Copyright & 2001 Elsevier Ltd.

collision of cosmic ray produced neutrons with nitrogen according to the reaction [I].


14 6 C

½I

)14 ¯ þQ 7 Nþb þv

½II

where v¯ is an antineutrino and Q is the decay energy. The atmospheric production rate varies somewhat and is influenced by changes in the solar wind and in the earth’s geomagnetic field intensity. A mean of 1.57 atom cm2 s1 is estimated based on the longterm record preserved in tree rings and a carbon reservoir model. This long-term production rate yields a global natural 14C inventory of approximately 50 t (1t ¼ 106 g). Production estimates based on the more recent record of neutron flux measurements tend to be higher, with values approaching 2 atom cm2 s1. Figure 1 shows the atmospheric history of 14C from AD 1511 to AD 1954 measured by Minze Stuiver (University of Washington) using tree growth rings. The strong decrease that occurs after about AD 1880 is due to dilution by anthropogenic addition of CO2 during the industrial revolution by the burning of fossil fuels (coal, gas, oil). This dilution has come to be known as the Suess effect (after Hans E. Suess). 20 10

14

In 1934 F.N.D. Kurie at Yale University obtained the first evidence for existence of radiocarbon (carbon14, 14C). Over the next 20 years most of the details for measuring 14C and for its application to dating were worked out by W.F. Libby and co-workers. Libby received the 1960 Nobel Prize in chemistry for this research. The primary application of 14C is to date objects or to determine various environmental process rates. The 14 C method is based on the assumption of a constant atmospheric formation rate. Once produced, atmospheric 14C reacts to form 14CO2, which participates in the global carbon cycle processes of photosynthesis and respiration as well as the physical processes of dissolution, particulate deposition, evaporation, precipitation, transport, etc. Atmospheric radiocarbon is transferred to the ocean primarily by air–sea gas exchange of 14CO2. Once in the ocean, 14CO2 is subject to the same physical, chemical, and biological processes that affect CO2. While alive, biota establish an equilibrium concentration of radiocarbon with their surroundings; that is, 14C lost by decay is replaced by uptake from the environment. Once the tissue dies or is removed from an environment that contains 14C, the decay is no longer compensated. The loss of 14C by decay can then be used to determine the time of death or removal from the original 14C source. After death or removal of the organism, it is generally assumed that no exchange occurs between the tissue and its surroundings; that is, the system is assumed to be closed. As a result of the 14C decay rate, the various reservoir sizes involved in the carbon cycle, and exchange rates between the reservoirs, the ocean contains approximately 50 times as much natural radiocarbon as does the atmosphere. Carbon-14 is one of three naturally occurring carbon isotopes; 14C is radioactive, has a half-life of 5730 years and decays by emitting a b-particle with an energy of about 156 keV. On the surface of the earth, the abundance of natural 14C relative to the two stable naturally occurring carbon isotopes is 12C : 13C : 14 C ¼ 98.9% : 1.1% : 1.2 1010 %. Natural radiocarbon is produced in the atmosphere, primarily by the

14 1 þ14 7 N )6 C þ 1 H

where n is a neutron and H is the proton emitted by the product nucleus. Similarly, the decay of 14C takes place by emission of a b-particle and leads to stable nitrogen according to reaction (II),

Δ C (ppt)

Introduction

1 0n

0 −10 −20

1500

1600

1700

1800

1900

Year Figure 1 Atmospheric history of D14C measured by M. Stuiver in tree rings covering AD 1511 to AC 1954. Most of the decrease over the last hundred years is due to the addition of anthropogenic CO2 to the atmosphere during the industrial revolution by the burning of fossil fuels.


637

RADIOCARBON

d14 C D C ¼ d C 2ðd C þ 25Þ 1 þ 1000 14

14

!

13

½1

d14C is given by eqn [2] and the definition of d13C is analogous to that for d14C. " 14

d C¼

14

# C=Csmp 14 C=Cstd 1000 14 C=Cj std

13

C and the additive constant 25 is a normalization factor conventionally applied to all samples and based on the mean value of terrestrial wood. The details of 14C calculations can be significantly more involved than expressed in the above equations; however, there is a general consensus that the calculations and reporting of results be done as described by Minze Stuiver and Henry Polach in a paper specifically written to eliminate differences that existed previously. D14C has units of parts per thousand (ppt). That is, 1 ppt means that 14C/12C for the sample is greater than 14C/12C for the standard by 0.001. In these units the radioactive decay rate of 14 C is approximately 1 ppt per 8.1 years. The number of surface ocean measurements made before any bomb-derived contamination are insufficient to provide the global distribution before input from explosions. It is now possible to measure D14C values in the annual growth rings of corals. By establishment of the exact year associated with each ring, reconstruction of the surface ocean D14C history is possible. Applying the same procedure to long-lived mollusk shells extends the method to higher latitudes than is possible with corals. Whether corals or shells are used, it must be demonstrated that the coral or shell incorporates 14C in the same ratio as the water in which it grew or at least that the fractionation is known. This method works only over the depth range at which the animal lived. Figure 2 shows the D14C record from two Pacific coral reefs measured by Ellen Druffel. Vertical lines indicate the period of atmospheric nuclear tests (1945–1963). The relatively small variability over the first B300 years of the record includes variations due to weather events, climate change, ocean circulation, atmospheric production, etc. The last 50 years of the sequence records the

100 50

14

Prior to 16 July 1945 all radiocarbon on the surface of the earth was produced naturally. On that date, US scientists carried out the first atmospheric atomic bomb test, known as the Trinity Test. Between 1945 and 1963, when the Partial Test Ban Treaty was signed and atmospheric nuclear testing was banned, approximately 500 atmospheric nuclear explosions were carried out by the United States (215), the former Soviet Union (219), the United Kingdom (21) and France (50). After the signing, a few additional atmospheric tests were carried out by China (23) and other countries not participating in the treaty. The net effect of the testing was to significantly increase 14C levels in the atmosphere and subsequently in the ocean. Anthropogenic 14C has also been added to the environment from some nuclear power plants, but this input is generally only detectable near the reactor. It is unusual to think of any type of atmospheric contamination – especially by a radioactive species – as beneficial; however, bomb-produced radiocarbon (and tritium) has proven to be extremely valuable to oceanographers. The majority of the atmospheric testing, in terms of number of tests and 14C production, occurred over a short time interval, between 1958 and 1963, relative to many ocean circulation processes. This time history, coupled with the level of contamination and the fact that 14C becomes intimately involved in the oceanic carbon cycle, allows bomb-produced radiocarbon to be valuable as a tracer for several ocean processes including biological activity, air–sea gas exchange, thermocline ventilation, upper ocean circulation, and upwelling. Oceanographic radiocarbon results are generally reported as D14C, the activity ratio relative to a standard (NBS oxalic acid, 13.56 dpm per g of carbon) with a correction applied for dilution of the radiocarbon by anthropogenic CO2 with age corrections of the standard material to AD1950. D14C is defined by eqn [1].

Δ C (ppt)

638

0 −50 1700

1900

2000

Year

½2

The first part of the second term in the right side of eqn [1] 2(d13C þ 25), corrects for fractionation effects. The factor of 2 accounts for the fact that 14C fractionation is expected to be twice as much as for

1800

Figure 2 Long-term history of D14C in the surface Pacific Ocean measured by E. Druffel in two coral reefs. The vertical lines surround the period of atmospheric nuclear weapons testing. The oceanic response to bomb contamination is delayed relative to the atmosphere because of the relatively long equilibration time between the ocean and atmosphere for 14CO2.


RADIOCARBON

Sampling and Measurement Techniques The radiocarbon measurement technique has existed for only 50 years. The first 14C measurement was made in W.F. Libby’s Chicago laboratory in 1949

1000

New Zealand • • Germany • • Tree Ring • •

Δ C (ppt)

800

• • •

600

14

400 200 0 1940

1950

1960

1970

1980

1990

Year

(A) 200 150

Druffel Toggweiler

100 50

14

Δ C (ppt)

invasion of bomb-produced D14C. Worth noting is the fact that the coral record of the bomb signal is lagged. That is, the coral values did not start to increase immediately testing began nor did they cease to increase when atmospheric testing ended. The lag is due to the time for the northern and southern hemisphere atmospheres to mix (B1 year) and to the relatively long time required for the surface ocean to equilibrate with the atmosphere with respect to D14C (B10 years). Because of the slow equilibration, the surface ocean is frequently not at equilibrium with the atmosphere. This disequilibrium is one of the reasons why pre-bomb surface ocean results, when expressed as ages rather than ppt units, are generally ‘old’ rather than ‘zero’ as might be expected. Figure 3A shows measured atmospheric D14C levels from 1955 to the present in New Zealand (data from T.A. Rafter, M.A. Manning, and co-workers) and Germany (data from K.O. Munnich and co-workers) as well as older estimates based on tree ring measurements (data from M. Stuiver). The beginning of the significant increase in the mid-1950s marks the atmospheric testing of hydrogen bombs. Atmospheric levels increased rapidly from that point until the mid-1960s. Soon after the ban on atmospheric testing, levels began a decrease that continues up to the present. The rate of decrease in the atmosphere is about 0.055 y1. Also clearly evident in the figure is that the German measurements were significantly higher than those from New Zealand between approximately 1962 and 1970. The difference reflects the facts that most of the atmospheric tests were carried out in the Northern Hemisphere and that approximately 1 year is required for atmospheric mixing across the Equator. During that interval some of the atmospheric 14CO2 is removed. Once atmospheric testing ceased, the two hemispheres equilibrated to the same radiocarbon level. Figure 3B shows detailed Pacific Ocean D14C coral ring data (J.R. Toggwelier and E. Druffel). This surface ocean record shows an increase during the 1960s; however, the peak occurs somewhat later than in the atmosphere and is significantly less pronounced. Careful investigation of coral data also demonstrates the north–south difference evidenced in the atmospheric record.

639

0 − 50

North Pacific = Filled South Pacific = Empty

− 100 1940 (B)

1950

1960

1970

1980

1990

Year

Figure 3 (A) Detailed atmospheric D14C history as recorded in tree rings for times prior to 1955 and in atmospheric gas samples from both New Zealand and Germany subsequently. The large increase in the late 1950s and 1960s was due to the atmospheric testing of nuclear weapons (primarily fusion devices). The hemispheric difference during the 1960s is because most atmospheric bomb tests were carried out north of the Equator and there is a resistance to atmospheric mixing across the equator. Atmospheric levels began to decline shortly after the ban on atmospheric bomb testing. (B) D14C in the surface Pacific Ocean as recorded in the annual growth rings of corals. The same general trend seen in the atmosphere is present. The bomb contamination peak is broadened and time-lagged relative to the atmosphere due to both mixing and to the time required for transfer from the atmosphere to the ocean.

and the first list of ages was published in 1951. A necessary prerequisite to the age determination was accurate measurement of the radiocarbon half-life. This was done in 1949 in Antonia Engelkeimer’s laboratory at the Argonne National Laboratory. Between 1952 and 1955 several additional radiocarbon dating laboratories opened. By the early 1960s several important advances had occurred including the following.

• •

Significantly improved counting efficiency and lower counting backgrounds, resulting in much greater measurement precision and longer timescale over which the technique was applicable. Development of the extraction and concentration technique for sea water samples.


640

• •

•

RADIOCARBON

More precise determination of the half-life by three different laboratories. Recognition by Hans Suess, while at the USGS and Scripps Institution of Oceanography, that radiocarbon in modern samples (since the beginning of the industrial revolution) was being diluted by anthropogenic CO2 addition to the atmosphere and biosphere. Recognition that atmospheric and oceanic D14C levels were increasing as a result of atmospheric testing of nuclear weapons.

During the 1970s and 1980s incremental changes in technique and equipment further increased the precision and lowered the counting background. With respect to the ocean, this was a period of sample collection, analysis, and interpretation. The next significant change occurred during the 1990s with application of the accelerator mass spectrometry (AMS) technique to oceanic samples. This technique counts 14C atoms rather than detecting the energy released when a 14C atom decays. The AMS technique allowed reduction of the sample size required for oceanic D14C determination from approximately 250 liters of water to 250 milliliters! By 1995 the AMS technique was yielding results that were as good as the best prior techniques using large samples and decay counting. This size reduction and concurrent automation procedures had a profound effect on sea water D14C determination. Many of the AMS techniques were developed and most of the oceanographic AMS D14C measurements have been made at the National Ocean Sciences AMS facility in Woods Hole, Massachusetts, by Ann McNichol, Robert Schneider, and Karl von Reden under the initial direction of Glenn Jones and more recently John Hayes. The natural concentration of 14C in sea water is extremely low (B1 4109 atoms kg1). Prior to AMS, the only available technique to measure this low concentration was radioactive counting using either gas proportional or liquid scintillation detectors. Large sample were needed to obtain high precision and to keep counting times reasonable. Between about 1960 and 1995 most subsurface open-ocean radiocarbon water samples were collected using a Gerard–Ewing sampler commonly known as a Gerard barrel. The final design of the Gerard barrel consisted of a stainless steel cylinder with a volume of approximately 270 liters. An external scoop and an internal divider running the length of the cylinder resulted in efficient flushing while the barrel was lowered through the water on wire rope. When the barrel was returned to the ship deck, the water was transferred to a gas-tight container and acidified to convert carbonate species to

CO2. The CO2 was swept from the water with a stream of inert gas and absorbed in a solution of sodium hydroxide. The solution was returned to shore where the CO2 was extracted, purified, and counted. When carefully executed, the procedure produced results which were accurate to 2–4 ppt based on counting errors alone. Because of the expense, time, and difficulty, samples for replicate analyses were almost never collected. With the AMS technique only 0.25 liter of sea water is required. Generally a 0.5 liter water sample is collected at sea and poisoned with HgCl2 to halt all biological activity. The water is returned to the laboratory and acidified, and the CO2 is extracted and purified. An aliquot of the CO2 is analyzed to determine d13C and the remainder is converted to carbide and counted by AMS. Counting error for the AMS technique can be o2 ppt, however, replicate analysis shows the total sample error to be approximately 4.5 ppt.

Sampling History Soon after the radiocarbon dating method was developed, it was applied to oceanic and atmospheric samples. During the 1950s and 1960s most of the oceanographic samples were limited to the shallow waters owing to the difficulty of deep water sampling combined with the limited analytical precision. The majority of the early samples were collected in the Atlantic Ocean and the South-west Pacific Ocean. Early sample coverage was insufficient to give a good description of the global surface ocean radiocarbon content prior to the onset of atmospheric testing of thermonuclear weapons; however, repeated sampling at the same location was sufficient to record the surface water increase due to bomb-produced fallout. A very good history of radiocarbon activity, including the increase due to bomb tests and subsequent decrease, exists primarily as a result of the work of R. Nydal and co-workers (Trondheim) and K. Munnich and co-workers (Heidelberg). The primary application of early radiocarbon results was to estimate the flux of CO2 between the atmosphere and ocean and the average residence time in the ocean. Sufficient subsurface ocean measurements were made, primarily by W. Broecker (Lamont–Doherty Earth Observatory LDEO) and H. Craig (Scripps Institution of Oceanography SIO), to recognize that radiocarbon had the potential to be an important tracer of deep ocean circulation and mixing rates. During the 1970s the Geochemical Ocean Sections (GEOSECS) program provided the first full water


RADIOCARBON

column global survey of the oceanic radiocarbon distribution. The GEOSECS cruise tracks were approximately meridional through the center of the major ocean basins. Radiocarbon was sampled with a station spacing of approximately 500 km and an average of 20 samples per station. All of the GEO¨ stlund SECS D14C measurements were made by G. O (University of Miami) and M. Stuiver (University of Washington) using traditional b counting of largevolume water samples with a counting accuracy of B4 ppt. GEOSECS results revolutionized what was

known about the oceanic D14C distribution and the applications for which radiocarbon is used. During the early 1980s the Atlantic Ocean was again surveyed for radiocarbon as part of the Transient Tracers in the Ocean (TTO) North Atlantic Study (NAS) and Tropical Atlantic Study (TAS) programs and the South Atlantic Ventilation Experiment (SAVE). Sampling for these programs was designed to enable mapping of property distributions on constant pressure or density surfaces with reasonable gridding uncertainty. The radiocarbon

14

Δ C (ppt)

14

Δ C (ppt)

Pressure (dB)

− 200

− 100

0

− 200

100

0

0

1000

1000

2000

2000

3000

4000

4000

5000

5000

6000

6000

Pressure (dB)

_ 200

100

0

0

1000

1000

2000

2000

3000

Southern

100

14

14

0

0

Atlantic

Δ C (ppt) − 100

− 100

3000

Pacific

− 200

641

3000

4000

4000

5000

5000

6000

6000

Δ C (ppt) _ 100 0

100

Indian

Figure 4 Average vertical D14C profiles for the major ocean basins. Except for the Southern Ocean the dotted line is for the Southern Hemisphere and the solid line for the Northern Hemisphere. The Pacific and Southern Ocean profiles were compiled from WOCE data; the Atlantic profiles from TTO and SAVE data; and the Indian Ocean profiles from GEOSECS data. In approximately the upper 1000 m( ¼ 1000 dB) of each profile, the natural D14C is contaminated with bomb-produced radiocarbon.


642

RADIOCARBON

portion of these programs was directed by W. ¨ stlund made the D14C measurements Broecker. O 13 with d C provided by Stuiver using the GEOSECS procedures. Comparison of TTO results to GEOSECS gave the first clear evidence of the penetration of the bomb-produced radiocarbon signal into the subsurface North Atlantic waters. The French carried out a smaller scale (INDIGO) 14C program in ¨ stlund and the Indian Ocean during this time with O P. Quay (University of Washington) collaborating. These data also quantified upper ocean changes since GEOSECS and relied on the same techniques. The most recent oceanic survey was carried out during the 1990s as part of the World Ocean Circulation Experiment (WOCE). This program was a multinational effort. The US 14C sampling effort was heavily focused on the Pacific (1991–1993) and Indian oceans (1995–1996) since TTO and SAVE had provided reasonable Atlantic coverage. R. Key (Princeton University) directed the US radiocarbon effort with collaboration from P. Schlosser (LDEO) and Quay. In the deep Pacific where gradients were known to be small, most radiocarbon sampling was by the proven large-volume b technique. The Pacific thermocline, however, was sampled using the AMS technique. Shifting techniques allowed thermocline waters to be sampled at approximately 2–3 times the horizontal density used for large volume sampling. ¨ stlund and Stuiver again measured the large-volO ume samples while the AMS samples were measured at the National Ocean Sciences AMS facility (NOSAMS) at Woods Hole Oceanographic Institution. By 1994 the analytical precision at NOSAMS had improved to the point that all US Indian Ocean WOCE 14C sampling used this technique. WOCE sampling increased the total number of 14C results for the Pacific and Indian Oceans by approximately an order of magnitude. Analysis of the Pacific Ocean samples was completed in 1998. US WOCE 14C sampling in the Atlantic was restricted to two zonal sections in the north-west basin using the AMS technique. Analysis of the Atlantic and Indian Ocean samples is expected to be finished during 2000–2001.

the dotted line being southern basin and solid line northern basin. All of the profiles have higher D14C in shallow waters, reflecting proximity to the atmospheric source. The different collection times combined with the penetration of the bomb-produced signal into the upper thermocline negate the possibility of detailed comparison for the upper 600– 800 dB (deeper for the North Atlantic). Detailed comparison is justified for deeper levels. The strongest signal in deep and bottom waters is that the North Atlantic is significantly younger (higher D14C) than the South Atlantic, while the opposite holds for the Pacific. Second, the average age of deep water increases (D14C decreases) from Atlantic to Indian to Pacific. Third, the Southern Ocean D14C is very uniform below approximately 1800 dB at a level (B 160 ppt). This is similar to the near bottom water values for all three southern ocean basins. All three differences are directly attributable to the largescale thermohaline circulation. Figure 5 shows meridional sections for the Atlantic, Indian and Pacific oceans using subsets of the data from Figure 4. As with Figure 4, the D14C values in the upper water column have been increased by invasion of the bomb signal. The pattern of these contours, however, is generally representative of the natural D14C signal. The D14C ¼ 100% contour can be taken as the approximate demarcation between the bomb-contaminated waters and those having only natural radiocarbon. Comparison of the major features in each section shows that the meridional D14C distributions in the Pacific and Indian Oceans are quite similar. The greatest difference between these two is that the Indian Ocean deep water (1500–3500 m) is significantly younger than Pacific deep waters. In both oceans:

• • • •

D14C Distribution and Implications for Large-scale Circulation The distribution of radiocarbon in the ocean is controlled by the production rate in the atmosphere, the spatial variability and magnitude of 14CO2 flux across the air–sea interface, oceanic circulation and mixing, and the carbon cycle in the ocean. Figure 4 shows average vertical radiocarbon profiles for the Pacific, Atlantic, Southern, and Indian oceans with

• •

The near bottom water has higher D14C than the overlying deep water. The deep and bottom waters have higher D14C at the south than the north. The lowest D14C values are found as a tongue extending southward from the north end of the section at a depth of B2500 m. Deep and bottom water at the south end of each section is relatively uniform with D14CB 160 ppt. The D14C gradient with latitude from south to north is approximately the same for both deep waters and for bottom waters. The D14C contours in the thermocline shoal both at the equator and high latitudes. (This feature is suppressed in the North Indian Ocean owing to the limited geographic extent and the influence of flows through the Indonesian Seas region and from the Arabian Sea.)


RADIOCARBON

0

100

150 100

1000

Depth (m)

50

0 − 50

2000 _ 150

3000

−100

−160

4000 5000

20°S

40°S

20°N 0 Latitude

(A) 0

100

50 0 _ 100

Depth (m)

1000

40°N

_ 50

_ 150

2000

_ 160

3000

−170

−180

−190

4000 − 160

5000 60°S

40°S

20°S Latitude

(B) 0

Depth (m)

1000 2000

50

0

20°N

100

100

_ 50 0 _ 100 _150 _160

3000

_ 220

_ 210 _ 190 _ 180 _ 200

4000

_ 240 _ 230

_170

5000

60°S (C)

40°S

20°S

0 Latitude

20°N

40°N

Figure 5 Typical meridional sections for each ocean compiled from a subset of the data used for Figure 4. The deep water contour patterns are primarily due to the large-scale thermohaline circulation. The highest deep water D14C values are found in the North Atlantic and the lowest in the North Pacific. The natural D14C in the upper ocean is contaminated by the influx of bombproduced radiocarbon.

In the Atlantic Ocean the pattern in the shallow water down through the upper thermocline is similar to that in the other oceans. The D14C distribution in the deep and bottom waters of the Atlantic is, however, radically different. The only similarities to the other oceans are (1) the D14C value for deep and

643

bottom water at the southern end of the section, (2) a southward-pointing tongue in deep water, and (3) the apparent northward flow indicated by the near-bottom tongue-shaped contour. Atlantic deep water has higher D14C than the bottom water, and the deep and bottom waters at the north end of the section have higher rather that lower D14C as found in the Indian and Pacific. Additionally, the far North Atlantic deep and bottom waters have relatively uniform values rather than a strong vertical gradient. The reversal of the Atlantic deep and bottom water D14C gradients with latitude relative to those in the Indian and Pacific is due to the fact that only the Atlantic has the conditions of temperature and salinity at the surface (in the Greenland–Norwegian Sea and Labrador Sea areas) that allow formation of a deep water mass (commonly referred to as North Atlantic Deep Water, NADW). Newly formed NADW flows down slope from the formation region until it reaches a level of neutral buoyancy. Flow is then southward, primarily as a deep western boundary current constrained by the topography of the North American slope. In its southward journey, NADW encounters and overrides northward-flowing denser waters of circumpolar origin. This general circulation pattern can be very clearly demonstrated by comparing the invasion of the bomb-produced tritium and radiocarbon signals obtained during GEOSECS to those from the TTO programs. This large circulation pattern leads to the observed D14C distribution in the deep Atlantic. Since neither the Pacific nor the Indian Ocean has a northern hemisphere source of deep water, the large-scale circulation is simpler. The densest Pacific waters originate in the Southern Ocean and flow northward along the sea floor (Circumpolar Deep Water, CDW). In the Southern Ocean, CDW is partially ventilated, either by direct contact with the atmosphere or by mixing with waters that have contacted the atmosphere, resulting in somewhat elevated D14C. As CDW flows northward, it ages, warms, mixes with overlying water, and slowly upwells. This upwelling, combined with mixing with overlying lower thermocline waters, results in the water mass commonly known as Pacific Deep Water (PDW). PDW has the lowest D14C values found anywhere in the oceans. The long-term mean flow pattern for PDW is somewhat controversial; however, the radiocarbon distribution supports a southward flow with the core of the flow centered around 2500 m. WOCE results further imply that if there is a mean southward flow of PDW, it may be concentrated toward the eastward and westward boundaries rather than uniformly distributed zonally. Figure 6 shows a zonal Pacific WOCE D14C section


644

RADIOCARBON

at 321S contoured at the same intervals as the previous sections. PDW is identified by the minimum layer between 2000 and 3000 m. The PDW core appears segregated into two channels, one against the South American slope and the other over the Kermadec Trench. The actual minimum values in the latter were found at B1701W, essentially abutting the western wall of the trench. The northwardflowing CDW is also clearly indicated in this section by the relatively high D14C values near the bottom between 1401W and the Date Line. Little has been said about the natural D14C values found in the upper ocean where bomb-produced radiocarbon is prevalent. GEOSECS samples were collected only B10 years after the maximum in atmospheric D14C. GEOSECS surface water measurements almost always had the highest D14C values. Twenty years later during WOCE, the maximum D14C was generally below the surface. Broecker and Peng (1982, p. 415, Figures 8–19) assembled the few surface ocean D14C measurements made prior to bomb contamination for comparison to the GEOSECS surface ocean data. For the Atlantic and Pacific Oceans, their plot of D14C versus latitude shows a characteristic ‘M’ shape with maximum D14C values of approximately 50 ppt centered in the main ocean gyres between latitudes 201 and 401. Each ocean had a relative minimum D14C value of approximately 70 ppt in the equatorial latitudes, 201 S to 201 N and minima at high latitudes ranging from 70 ppt for the far North Atlantic to

South Pacific WOCE Data at 32°S

0

Depth (m)

1000 2000 3000 4000 5000 6000

.. .. .. .. ... .. . . .. . . .. . .. .. ... . . . . . . . . . . .

.. .. .. . .. .. .. .. . .. .. .. . .. . .. .. .. .. . . . . . . . . . . . . . .

.. .. . . . . . . .

... ... .. . . . .

.. ... . .. .. . .. . .. .. . .. .. .

.. . ... .. . . . .

.. .. .. .. .. . . . . . . ... . . . . ... ... ... . .. .

. . .

... .. .. . .. . .. .. .. . . . . . .

.. . .. .. .. .. . .. . .. ... ... .. .. .. ... .. .. .. .. ... ..... .. ..... .. .. .. . .. .. . .. .. .... .. .... ... ... ... ... .. .. .... .. .... . ... ... .... .. ... ...... .. ......... .. . . .. ... . . .. . . . . .. . . ... .. .. .. .. .. .. ...100 . .. .. . .. .. . . . . . . . .. . . . .. .. . .. .. . .. 50.. .. . . .. . .. .. ... .. .. .. .... .. .... .. . ... .. .. .... . ... ..... . . .. . 0. .. ... . .. . . . . .. . . . . . . .. . . . . .. . .. _ .50. .. . .. .. . .. . .. . .. . . .. ... .. .. . ...... .. . . . . . . . . _ . . . . . . .. . . . _. . . . .. .. .. . .... .. . . . .. . .. .. 100 . . . . . 150 . .. .. .. . . . . _ . .. . . . . . .. . . . . . . . ... . .... . . 160 . . . .. . . . ._170. _ . . .. . . . . .. . .. . . 180 . . . _.190 .. .. . . . . . . . ... . . . . . . . . . .. . . .. . .... . . . . . . . . . _. 200 . . . . . . .. . .. . . . . . . . . _..210 . . . .. .. . . . _ 220 . . . . . . . . .. . . . . . . . . . . . . .. . . . ._ 210 . . . . .. . . . . . . . ... . . . _ .200. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . .. . . . . . . . . . . . _ 190 . .. . . . .. . . . _ 180 . . .. . . . . . .. . _ . _ 170 . . . .160 . . . . . . . . . . . . . . . . . .. . _. . . . .160 . . .. . . . . .. . .

160°E 180° 160°W 140°W 120°W 100°W 80°W Longitude Figure 6 Zonal section of D14C in the South Pacific collected during the WOCE program. The two minima at 2000–2500 m depth are thought to be the core of southward-flowing North Pacific Deep Water. Northward-flowing Circumpolar Deep Water is identified by the relatively high values in the Kermadec Trench area at the bottom between 1401W and the Date Line.

150 ppt for the other high latitudes. Pre-bomb measurements in the Indian Ocean are extremely sparse; however, the few data that exist imply a similar distribution. The GEOSECS surface ocean data had the same ‘M’ shape; however, all of the values were significantly elevated owing to bombderived contamination and the pattern was slightly asymmetric about the equator with the Northern Hemisphere having higher values since most of the atmospheric bomb tests were carried out there. The ‘M’ shape of D14C with latitude is due to circulation patterns, the residence time of surface water in an ocean region, and air–sea gas exchange rates. At mid-latitudes the water column is relatively stable and surface waters reside sufficiently long to absorb a significant amount of 14C from the atmosphere. In the equatorial zone, upwelling of deeper (and therefore lower D14C) waters lowers the surface ocean value. At high latitudes, particularly in the Southern Ocean, the near-surface water is relatively unstable, resulting in a short residence time. In these regions D14C acquired from the atmosphere is more than compensated by upwelling, mixing, and convection. Figure 7 shows a comparison for GEOSECS and WOCE surface data from the Pacific Ocean. The GEOSECS D14C values are higher than WOCE everywhere except for the Equator. The difference is due to two factors. First, GEOSECS sampling occurred shortly after the atmospheric maximum. At that time the air–sea D14C gradient was large and the surface ocean D14C values were dominated by air–sea gas exchange processes. Second, by the 1990s, atmospheric D14C levels had declined significantly and sufficient time had occurred for ocean mixing to compete with air–sea exchange in terms of controlling the surface ocean values. During the 1990s, the maximum oceanic D14C values were frequently below the surface. Near the Equator the situation is different. Significant upwelling occurs in this zone. During GEOSECS, waters upwelling at low latitude in the Pacific were not yet contaminated with bomb radiocarbon. Twenty years later, the upwelling waters had acquired a bomb radiocarbon component. While surface ocean D14C generally decreased between GEOSECS and WOCE, values throughout the upper kilometer of the water column generally increased as mixing and advection carried bombproduced radiocarbon into the upper thermocline. The result of these processes on the bomb-produced D14C signal can be visualized by comparing GEOSECS and WOCE depth distributions. Figure 8 shows such a comparison. To produce this figure the WOCE data from section P16 (1521W) were gridded (center panel). GEOSECS data collected east of the data line were then gridded to the same grid (top


645

RADIOCARBON

0

Depth (m)

100

0

80°S 60°S 40°S 20°S

0°

20°N 40°N 60°N

Latitude Figure 7 Distribution of D14C in the surface Pacific Ocean as recorded by the GEOSECS program in the early 1970s and the WOCE program in the early 1990s. From Figure 3B it follows that GEOSECS recorded the maximum bomb-contamination. Over the 20 years separating the programs, mixing and advection dispersed the signal. By the time of WOCE the maximum contamination level was found below the surface at many locations. The asymmetry about the Equator in the GEOSECS data is a result of most atmospheric bomb tests being executed in the Northern Hemisphere.

panel). Once prepared, the two sections were simply subtracted grid box by grid box (bottom panel). One feature of Figure 8 is the asymmetry about the Equator. The difference at the surface in Figure 8 reflects the same information (and data) as in Figure 7. The greatest increase (up to 60 ppt) along the section is in the Southern Hemisphere mid-latitude thermocline at a depth of 300–800 m. This concentration change decreases in both depth and magnitude toward the Equator. All of the potential density isolines that pass through this region of significant increase (dashed lines in the bottom panel) outcrop in the Southern Ocean. These outcrops (especially during austral winter) provide the primary pathway by which radiocarbon is entering the South Pacific thermocline. In the North Pacific the surface ocean decrease extends as a blob well into the water column (>200 m). This large change is due to the extremely high surface concentrations measured during GEOSECS and to subsurface mixing and ventilation processes that have diluted or dispersed the peak signal. The values contoured in the bottom panel represent the change in D14C between the two surveys, not the total bomb D14C. WOCE results from the Indian Ocean are not yet available. Once they are, changes since GEOSECS in the South Indian Ocean should be quite similar to those in the South Pacific because the circulation and ventilation pathways are similar. Changes in the North Indian Ocean are difficult to predict owing to water inputs from the Red Sea and the Indonesian

200 400 600 800

• • •• • • • • • •

•• • • • • • • • •

1000

•

1200

•

60°S

(A)

Depth (m)

_100

GEOSECS 1973−1974 WOCE 1992−1995

0 200 400 600 800 1000 1200

0 200

•• • •• •• • •• • • • •• • ••

•• • ••• ••• • •• •• • ••• • • •• • •• • ••• • • • •• • ••• •• • • • • • •• • •• •• • •• • • • • •• • _ 150 • • •

40°S

•• •• •• 100 • •• • • •• • • •• • • • •• • •• • •

• •

• • ••• •• • •••• • • • ••• • • • • • • • • •• • •

• •

• •• • _ 200 •

0°

20°S

•• •• • •• •• • •• •• • •• •• • ••

• •••••• • •••••• 100 • 50•• 0•• 150 • • _•• 50 •••••• • _ 100 •••••• • • •• ••• • •••• •• • •• ••••••• • • • ••• • •• • •• •• • • • •• • ••••

20°N

40°N

••••••••• • • • • •• • •• • • • • ••• •• •• •••• •• •• ••••••••••• •• •• •• ••• •• •• •• ••• •• • • • • ••• •• •• 100 ••• • • • • • •• • • 50• • • • •• • ••••••••••• • • • • ••• •••100 • •••••••••••••••• • ••• ••• •• • ••• •• • •• •• •••• •• •• •••••• • •• ••••••••••••••••• ••• ••• ••• •• •• ••• ••• •••• ••• • •••• •••• • •••••••••• ••• • • • •••••••••••••• •• •• • • •••• • •• •• •• ••• •• ••• •••• ••• •• •••••••••••••• •• •• ••• •• •• •• •••• ••• •• •• •• •• •••• •• •••• • • • • • • • • • • • •• • ••••••• • • • • • • •• ••• • • • •••••• •••• • •• • • •• • •• • •• •• •• •• •• ••••• •• •• •••••••••• • •• •• • • • ••••• ••• • • • • • • • • • • • • • • • • • • • • • • • • ••••••• • •• •• • ••• • •• • •• • • • •• • • ••••• • • • • • • •• •• •• • •• • • ••• • • • • • • ••• •• ••• •••••• • • • • • •• • •• • •• • • • •• • • • ••• • • • • • • _• 100 • •• • • • • • • • • • • • •• • • • ••• ••••• • • • • • •• • • • • • • • • • •• ••• • • • • • • • • • •• • •• •• ••• •• • • •• • • • • • • • • • • • • • • • • • • • ••• •• • • • • •• ••• •••••• • • • • •• • • • • • • • • • • •• • •• • • • • •• • •• • • • • • • • • • •• • •• • • • • • • •• ••••• • • • ••• ••• •• •• • •••• • • • • • • • • _ 150 • • •• • •• • • • •• •_ 150 • • • •_ 200

60°S

(B)

Depth (m)

14

Δ C (ppt)

200

40°S

_ 75 _ 50

0°

20°S _ 25

20°N

••• ••••••• •••0•• _••••50 ••• _•100 • •• •• • •• • •• • •• • • •• • • ••

40°N

60°N

_ 50 _ 25 0

25 50

25

60°N

50

400

50 25

600 800 1000 1200

_ 25

60°S

0

40°S

(C)

0

20°S

0°

20°N

40°N

60°N

Latitude

Figure 8 Panel (C) shows the change in the meridional eastern Pacific thermocline distribution of D14C between the GEOSECS (1973–1974, (A)) and WOCE (1991–1994, (B)) surveys. The change was computed by gridding each section then finding the difference. The dashed lines in (C) indicate constant potential density surfaces. Negative near-surface values indicate maximum concentration surfaces moving down into the thermocline after GEOSECS. The region of greatest increase in the southern hemisphere is ventilated in the Southern Ocean.

throughflow region and to the changing monsoonal circulation patterns. ¨ stlund and Claes Rooth described radioGo¨te O carbon changes in the North Atlantic Ocean using data from GEOSECS (1972) and the TTO North Atlantic Study (1981–1983). The pattern of change they noted is different from that in the Pacific because of the difference in thermohaline circulation mentioned previously. Prior to sinking, the formation waters for NADW are at the ocean surface long enough to pick up significant amounts of bomb radiocarbon from the atmosphere. The circulation pattern coupled with the timing of GEOSECS and TTO sampling resulted in increased D14C levels during the latter program. The significant changes were mostly limited to the deep water region north of 401N latitude. When the WOCE Atlantic samples are analyzed, we expect to see changes extending farther southward.


646

RADIOCARBON

Separating the Natural and Bomb Components

Atlantic Pacific Indian

_ 150

14

Δ C (ppt)

_ 100

_ 200 _ 250 0

50

100

150 _1

Silicate (μmol kg )

(A) _ 50

Atlantic Pacific Indian

_ 100 _ 150

14

Δ C (ppt)

Up to this point the discussion has been limited to changes in radiocarbon distribution due to oceanic uptake of bomb-produced radiocarbon. Many radiocarbon applications, however, require not the change but the distribution of either bomb or natural radiocarbon. Ocean water measurements give the total of natural plus bomb-produced D14C. Since these two are chemically and physically identical, no analytical procedure can differentiate one from the other. Far too few D14C measurements were made in the upper ocean prior to contamination by the bomb component for us to know what the upper ocean natural D14C distribution was. One separation approach derived by Broecker and co-workers at LDEO uses the fact that D14C is linearly anticorrelated with silicate in waters below the depth of bomb-14C penetration. By assuming the same correlation extends to shallow waters, the natural D14C can be estimated for upper thermocline and near surface water. Pre-bomb values for the ocean surface were approximated from the few prebomb surface ocean measurements. The silicate method is limited to temperate and low-latitude waters since the correlation fails at high latitudes, especially for waters of high silicate concentration. More recent work by S. Rubin and R. Key indicates that potential alkalinity (alkalinity þ nitrate normalized to salinity of 35) may be a better co-variable than silicate and can be used at all latitudes. Figure 9 illustrates the silicate and PALK correlations using the GEOSECS data set. Regardless of the co-variable, the correlation is used to estimate pre-bomb D14C in contaminated regions. The difference between the measured and estimated natural D14C is the bomb-produced D14C. In Figure 10 the silicate and potential alkalinity (PALK) methods are illustrated and compared. The upper panel (A) shows the measured D14C and estimates of the natural D14C using both methods. The bomb D14C is then just the difference between the measured value and the estimate of the natural value (B). For this example, taken from the mid-latitude Pacific, the two estimates are quite close; however this is not always true. In Figure 11A the upper 1000 m of the Pacific WOCE D14C section shown in Figure 5C is reproduced. Figure 11B shows the estimated natural D14C using the potential alkalinity method. The shape of the two contour sets is quite similar; however, the contour values and vertical gradients are very different, illustrating the strong influence of bombproduced radiocarbon on the upper ocean. The

_ 50

_ 200 _ 250 2350 (B)

2400

2450

2500 _1

Potential alkalinity, PALK (μmol kg )

Figure 9 Comparison of the correlation of natural D14C with silicate (A) and potential alkalinity (PALK ¼ [alkalinity þ nitrate] 35/salinity) (B) using the GEOSECS global data. Samples from high southern latitudes are excluded from the silicate relation. The presence of tritium was used to surmise the presence of bomb-D14C. The somewhat anomalous high PALK values from the Indian Ocean are from upwelling–high productivity zones and may be influenced by nitrogen fixation and/or particle flux.

integrated difference between these two sections would yield an estimate of the bomb-produced D14C inventory for the section.

Oceanographic Applications As illustrated, the D14C distribution can be used to infer general large-scale circulation patterns. The most valuable applications for radiocarbon derive from the fact that it is radioactive and has a half-life appropriate to the study of deep ocean processes and that the bomb component is transient and is useful as a tracer for upper ocean processes. A few of the more common uses are described below.

Deep Ocean Mixing and Ventilation Rate, and Residence Time Since the first subsurface measurements of radiocarbon, one of the primary applications has been the


RADIOCARBON

0

Depth (m)

500 1000 Measured PALK estimate Silicate estimate

1500 2000 _ 200

_ 100

0

100

14

Δ C (ppt)

(A) 0

Depth (m)

500 1000 PALK estimate Silicate estimate

1500

0

50

100

150

and mass fluxes for the abyssal ocean. In this case the model had only four boxes, one for the deep region (41500 m) of each ocean. New bottom water formation (NADW and Antarctic Bottom Water, AABW) were included as inputs to the Atlantic and Circumpolar boxes. Upwelling was allowed in the Atlantic, Pacific, and Indian boxes and exchange was considered between the Circumpolar box and each of the other three ocean boxes. Results from this calculation gave mean replacement times of 510, 250, 275, and 85 years for the deep Pacific, Indian, Atlantic, and Southern Ocean, respectively, and 500 years for the deep waters of the entire world. Upwelling rates were estimated at 4–5 m y1 and mass transports generally agreed with contemporary geostrophic calculations. Applying the same model to more recent data sets would yield the same results.

Oxygen Utilization Rate

2000

(B)

647

200

14

Bomb Δ C (ppt)

Figure 10 Panel (A) compares measured D14C from a midlatitude Pacific WOCE station with natural D14C estimated using the silicate and potential alkalinity methods. Bomb-D14C, the difference between measured and natural D14C, estimated with both methods is compared in (B). Integration of estimated bombD14C from the surface down to the depth where the estimate approaches zero yields an estimate of the bomb-D14C inventory. Inventory is generally expressed in units of atoms per unit area.

determination of deep ocean ventilation rates. Most of these calculations have used a box model to approximate the ocean system. The first such estimates yielded mean residence times for the various deep and abyssal ocean basins of 350–900 years. Solution of these models generally assumes a steady-state circulation, identifiable source water regions with known D14C, no mixing between water masses, and no significant biological sources or sinks. Another early approach assumed that the vertical distribution of radiocarbon in the deep and abyssal ocean could be described by a vertical advection–diffusion equation. This type of calculation leads to estimates of the effect of biological particle flux and dissolution and to the vertical upwelling and diffusion rates. The 1D vertical advection– diffusion approach has been abandoned for 2D and 3D calculations as the available data and our knowledge of oceanic processes have increased. When the GEOSECS data became available, box models were again used to estimate residence times

Radiocarbon can be used to determine the rate of biological or geochemical processes such as the rate at which oxygen is consumed in deep ocean water. The simplest example of this would be the case of a water mass moving away from a source region at a steady rate, undergoing constant biological oxygen uptake and not subject to mixing. In such a situation the oxygen utilization rate could be obtained from the slope of oxygen versus 14C in appropriate units. The closest approximation to this situation is the northward transport of CDW in the abyssal Pacific, although the mixing requirement is only approximate. Figure 12 shows such a plot for WOCE Pacific Ocean samples from deeper than 4000 m and north of 401S. In this case, apparent oxygen utilization (saturated oxygen concentration at equilibration temperature measured oxygen concentration) rather than oxygen concentration is plotted, to remove the effect of temperature on oxygen solubility. The least-squares slope of 0.83 mmol kg1 per ppt converts to 0.1 mmol kg1 y1 for an oxygen utilization rate. Generally, mixing with other water masses must be accounted for prior to evaluating the gradient. With varied or additional approximations, very similar calculations have been used to estimate the mean formation rates of various deep water masses.

Ocean General Circulation Model Calibration Oceanographic data are seldom of value for prediction. Additionally, the effect of a changing


648

RADIOCARBON

Depth (m)

0 200 400 600 800 1000

.. .. .. .. . . . . .

. .... . . . . . . . . . . . . . . . . . .. ...... . . . . . . . . .. . . . . ... .. . . .. ... ...... ... .. . ... . ... .. 80 . . . . . . . . . . ..... .. .. . .. . .. .. . . . . . . .. . . ... .. ... . . .. 40 . . .. ...... . . . . . . . . . . .. . . . . .. . . ... . .. .. .. . . . . . .. . . . . . . .. .. . . . .. . . . . . . .0. . . . .. . . . . . .. . . . . . . . . . . . . . . . . . .. . . . ... . . . . . . . . . . . . . . . 60°S

40°S

. ... .. .. .. .. .. .. ............ .. .. .. .. .. . . . .. . . .. .. ... .. . .. ... ... .. .. . ... . .. ............... .. . .. ... .. ... ... .. . .. 100 .. 100 . .... .... . ... ... ... ... . . . ............ .. .. ... .. .. .. . . . .. . . .... ... ... .. ... .. .. ......_ .. . .. .. ... .. ..._.. .. .. .. .. .. .. .. .. . . .. .. . .. . ..........40 . . . ... . . . .. .. .. .. .. .. .. .. .......... .. .. .. 80 .. . .. .. . . . . . . . .. . . ... . . .. . .. .. .. .......... .. .. . . . . . . . . . . ... .. . . . .. . ._. . . . ... . . . . . . . . . . . . . . . . 120 . .. . . . . . ... . . . . .. .. . . . . ... . . . . . . . . . .. . . . .. . . . . .. ... . . . . . . . ._ 160 . .. . . _ .. . .. 200 .. . . . . . . . .. . . . .. 0

20°S

20°N

40°N

.. ....... ... .......... ... ........ .. ..... .. ..... . .... . ... . .... . .. . . .. 60°N

Latitude

Depth (m)

0 200 400 600 800 1000

........................... . . .. ........ ........................... .. ............... .................. .. . . . .... . . . . .. .. . .. . . . .. .. ... ... ... ......... ... ..... ... ... . . . . ... .. .. .. ............. .. ... ... . . ... .... .... .. ... ... .. ... .. .. ....... ... ........ .. ... .. . .. . ........ . . . . .. .. . .. .. . . .. .. .. ...... . .. . . .. . . ... . .. .. .. .. .. . .. . .... .. ... . . . ......... .. .. . .. . .. .. .._..80 . . . . . . . . . . . . . . . . .. ...... . .. . . . . .. . . . .. . . . . ... . .. .. . . . .... . ... . . . . . . .. . . . .. . .. . .. .. .. . . . . . ...... .. .. .. . . . . .. .. . . . .. . . ..... . .. .. . . . . . .. . . . . . . . . . . . . . . ... . .. .. . . . . . ..... . . . . . . . . .. . . . . . . . ... . .. . . . . . . . .. . . 60°S

40°S

.. ... .. .. .. .......... .. .. .. ... .. .. .. .. .. .. . .. .. . .. .... .. .. .. .......................... .. ... .. .. .. ............... .. .. ... .. .. .. ... ... ... ... _ . ..... ... .. .... .. ... . ... ............................... . . . . . . . . . . . . . .. 60 . . . . . . .. ... ... ... ... .. ... .............. ... ... ... .. .. .. .. .. .. ... .. . . .. . ... . . .............. ....... . .. .. . .. . ... . .. . .. . .. ... .. ... . .100 .. .. .. .. .. ......... . .. . _ .. . . . . ...... . . . . . . .. . . . .. . . . . . .... . .. ._.. .. .. .. ......... .. .. .. . . . . .. . .. .. .. .. ... . .. .. . .. . . . ... . . . . 120 . .. .. ........_..140 . . . . . . . . . . . . . . . . ... . .. . . . ...... . . . . . . . . . . . .. .... .. .. .. .. ... . .. ... . ... . . . . . .._..160 .. . . . . . . . .. ... ....... . . .. .. . ._.. ..200 . . . . . ... . . . . . . . .._.180 . .._. ... . 220 . . . ... . . . . 0

20°S

20°N

40°N

60°N

Latitude

_1

Apparent oxygen utilization (μmol kg )

Figure 11 Upper thermocline meridional sections along 1521 W in the central Pacific. (A) The same measured data as in Figure 5C (B) An estimate of thermocline D14C values prior to the invasion of bomb-produced radiocarbon.

200

180

160

140

_ 240

_ 220

_ 200

_ 180

_ 160

14

Δ C (ppt) Figure 12 Apparent oxygen utilization plotted against measured D14C for WOCE Pacific Ocean samples taken at depths greater than 4000 m and north of 401S. The slope of the line ( 0.83170.015) can be used to estimate an approximate oxygen utilization rate of 0.1 mmol kg1 y1 if steady state and no mixing with other water masses is assumed.

oceanographic parameter on another parameter can be difficult to discern directly from data. These research questions are better investigated with numerical ocean models. Before an ocean model result can be taken seriously, however, the model must demonstrate reasonable ability to simulate current conditions. This generally requires that various model inputs or variables be ‘tuned’ or calibrated to match measured distributions and rates. Radiocarbon is the only common measurement that can be used to

calibrate the various rates of abyssal processes in general circulation models. M. Fiadeiro carried out the first numerical simulation for the abyssal Pacific and used the GEOSECS 14C data to calibrate the model. J.R. Toggweiler extended this study using a global model. Both the Fiadeiro and Toggweiler models, and all subsequent models that include the deep water D14C, are of coarse resolution owing to current computer limitations. As the much larger WOCE 14C data set becomes available, the failure of these models, especially in detail, becomes more evident. Toggweiler’s model, for example, has advective mixing in the Southern Ocean that is significantly greater than supported by data. Additionally, the coarse resolution of the model prevents the formation of, or at least retards the importance of, deep western boundary currents. Significant model deficiencies appear when the bomb-14C distribution and integrals at the time of GEOSECS and WOCE are compared with data. During the last 10 years the number and variety of numerical ocean models has expanded greatly, in large part because of the availability and speed of modern computers. The Ocean Carbon Model Intercomparison Project (OCMIP) brought ocean modelers together with data experts in the first organized effort to compare model results with data, with the long-term goals of understanding the processes that cause model differences and of improving the prediction capabilities of the models. The unique aspect of this study was that each participating group


RADIOCARBON

649

Figure 13 (Right) Global ocean circulation model results from 12 different coarse-resolution models participating in OCMIP-2 compared to WOCE data for natural D14C on a meridional Pacific section. The model groups are identified in each subpanel and in Table 1 All of the models used the same chemistry and boundary conditions.


650

RADIOCARBON

Table 1

OCMIP-2 participants

AWI CSIRO IGCR/CCSR IPSL LLNL MIT MPIM NCAR PIUB PRINCETON SOC

PMEL PSU PRINCETON

Model groups Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany Commonwealth Science and Industrial Research Organization, Hobart, Australia Institute for Global Change Research, Tokyo, Japan Institut Pierre Simon Laplace, Paris, France Lawrence Livermore National Laboratory, Livermore, CA, USA Massachusetts Institute of Technology, Cambridge, MA, USA Max Planck Institut fur Meteorologie, Hamburg, Germany National Center for Atmospheric Research, Boulder, CO, USA Physics Institute, University of Bern, Switzerland Princeton University AOS, OTL/GFDL, Princeton, NJ, USA Southampton Oceanography Centre/ SUDO/Hadley Center, UK Met. Office Data groups Pacific Marine Environmental Laboratory, NOAA, Seattle, WA, USA Pennsylvania State University, PA, USA Princeton University AOS, OTL/GFDL, Princeton, NJ, USA

essentially ‘froze’ development of the underlying physics in their model and then used the same boundary conditions and forcing in order to eliminate as many potential variables as possible. Radiocarbon, both bomb-derived and natural, were used as tracers in each model to examine air–sea gas exchange and long-term circulation. Figure 13 compares results from 12 global ocean circulation models with WOCE data from section P16. The tag in the top left corner of each panel identifies the institution of the modeling group. All of the model results and the data are colored and scaled identically and the portion of the section containing bomb radiocarbon has been masked. While all of the models get the general shape of the contours, the concentrations vary widely. Detailed comparison is currently under way, but cursory examination points out significant discrepancies in all model results and remarkable model-to-model differences. Similar comparisons can be made focusing on the bomb component. Discussion of model differences is beyond the scope of this work. For information, see publications by the various groups having results in Figure 13 (listed in Table 1). These radiocarbon results are not yet published, but an overview of the OCMIP-2 program can be found in the work of Dutay on chlorofluorocarbon in the same models (see Further Reading).

Air–Sea Gas Exchange and Thermocline Ventilation Rate Radiocarbon has been used to estimate air–sea gas exchange rates for almost as long as it has been measured in the atmosphere and ocean. Generally, these calculations are based on box models, which have both included and excluded the influence of bomb contamination. W. Broecker and T.-H. Peng summarized efforts to estimate air–sea transfer rates up to 1974 and gave examples based on GEOSECS results using both natural and bomb-14C and a stagnant film model. In this, the rate-limiting step for transfer is assumed to be molecular diffusion of the gas across a thin layer separating the mixed layer of the ocean from the atmosphere. In this model, if one assumes steady state for the 14C and 12C distribution and uniform 14C/12C for the atmosphere and surface ocean then the amount of 14C entering the ocean must be balanced by decay. For this model the solution is given by eqn [3]. "

# C=Cocean a14CO2 P 14 C=Cj aCO2 D ½CO2 jocean V mix " # l ½3 ¼ P 14 ½CO2 jmix A z C=Cmix a14CO2 1 14 C=Cjatm aCO2 14

Here D is the molecular diffusivity of CO2, z is the film thickness, ai is the solubility of i, V and A are the volume and surface area of the ocean, and l is the 14 C decay coefficient. Use of pre-industrial mean concentrations gave a global boundary layer thickness of 30 mm (D/zB1800 m y1 ¼ piston velocity). The film thickness is then used to estimate gas residence times either in the atmosphere or in the mixed layer of the ocean. For CO2 special consideration must be made for the chemical speciation in the ocean, and for 14CO2 further modification is necessary for isotopic effects. The equilibration times for CO2 with respect to gas exchange, chemistry, and isotopics are approximately 1 month, 1 year, and 10 years, respectively. Radiocarbon has been used to study thermocline ventilation using tools ranging from simple 3-box models to full 3D ocean circulation models. Many of the 1D and 2D models are based on work by W. Jenkins using tritium in the North Atlantic. In a recent example, R. Sonnerup and co-workers at the University of Washington used chlorofluorocarbon data to calibrate a 1D (meridional) along-isopycnal advection–diffusion model in the North Pacific with WOCE data. [4] is the basic equation for the


RADIOCARBON

651

14

Contour lines = Bomb Δ C (ppt) •

50°N

Latitude

30°N 20°N

140• • • • • • •

10°N 120

• • • • • • • • • • • • • • • • • • •

0° 140°E 160°E

• •• • •

• •

40°N ••• • • • • • • • • • • • • • • • • • • • • •• • ••

•

• • • • • • • • • • • • • •

• • •• •• • •• •• • • •• • • • • • • • • • • • • • • • • • • • • • • • •

• • • • • • • •

••

• •

100 •• 180° 160°W 140°W 120°W 100°W 80°W Longitude

Figure 14 Bomb-D14C on the potential density surface sy ¼ 26.1 in the North Pacific. The blue line is the wintertime outcrop of the surface based on long-term climatology. The Sea of Okhotsk is a known region of thermocline ventilation for the North Pacific.

would be input considerations to an investigation of thermocline ventilation.

14

Bomb Δ C (ppt)

200

150

Conclusions 100 25.75 26.10 26.30 26.50 26.65

50

0 0°

10°N

20°N 30°N Latitude

40°N

50°N

Figure 15 Meridional distribution of bomb-D14C on potential density surfaces in the North Pacific thermocline.

model. dC dC d2 C ¼ v þ K 2 dt dx dx

½4

In eqn [4] C is concentration, K is along-isopycnal eddy diffusivity, v is the southward component of along isopycnal velocity, t is time, and x is the meridional distance. Upper-level isopycnal surfaces outcrop at the surface. Once the model is calibrated, the resulting values are used to investigate the distribution of other parameters. The original work and the references cited there should be read for details, but Figure 14 shows an objective map of the bomb-14C distribution on the potential density surface 26.1 for the North Pacific and Figure 15 summarizes the bomb-14C distribution as a function of latitude. These figures illustrate the type of data that

Since the very earliest measurements, radiocarbon has proven to be an extremely powerful tracer, and sometimes the only available tracer, for the study of many oceanographic processes. Perhaps the most important of these today are large-scale deep ocean mixing and ventilation processes and the calibration of numerical ocean models. The first global survey of the radiocarbon distribution collected on the GEOSECS program resulted in radical changes in the way the abyssal ocean is viewed. The newer and much denser WOCE survey will certainly add significant detail and precision to what is known and will probably result in other, if not so many, totally new discoveries. Progress with this tracer today is due largely to the decrease in required sample size from B250 liters to B250 milliliters and to the availability and application of fast, inexpensive computers.

Glossary dpm Disintegrations per minute: a measure of the activity of a radioactivesubstance frequently used rather than concentration. t1/2 Half-life: time required for one half of the atoms of a radio active species to decay. l Decay constant for a radioactive species ¼ 1n(2)/ t1/2 Mean life, l 1 Average time expected for a given radioactive atom to decay. Abyssal Very deep ocean, often near bottom.


652

RADIOCARBON

Steady state Unchanging situation over long time interval relative to the process under consideration; frequently assumed state for the deep and abyssal ocean with respect to many parameters.

See also Abyssal Currents. Air–Sea Gas Exchange. Atmospheric Input of Pollutants. Bottom Water Formation. Carbon Cycle. Carbon Dioxide (CO2) Cycle. Elemental Distribution: Overview. LongTerm Tracer Changes. Marine Silica Cycle. Ocean Carbon System, Modeling of. Ocean Subduction. Current Systems in the Indian Ocean. Radioactive Wastes. Stable Carbon Isotope Variations in the Ocean. Ocean Circulation: Meridional Overturning Circulation. Tritium–Helium Dating. Water Types and Water Masses.

Further Reading Broecker WS, Gerard R, Ewing M, and Heezen BC (1960) Natural radiocarbon in the Atlantic Ocean. Journal of Geophysical Research 65(a): 2903--2931. Broecker WS and Peng T-H (1974) Gas exchange rates between air and sea. Tellus 26: 21--34. Broecker WS and Peng T-H (1982) Tracers in the Sea. Columbia University, Palisades, NY: Lamont-Doherty Geological Observatory. Broecker WS, Sutherland S, Smethie W, Peng T-H, and ¨ stlund G (1995) Oceanic radiocarbon: separation of O natural and bomb components. Global Biogeochemical Cycles 9(2): 263--288. Craig H (1969) Abyssal carbon radiocarbon in the Pacific. Journal of Geophysical Research 74(23): 5491--5506. Druffel ERM and Griffin S (1999) Variability of surface ocean radiocarbon and stable isotopes in the south-

western Pacific. Journal of Geophysical Research 104(C10): 23607--23614. Dutay JC, Bullister JL, and Doney SC (2001) Evaluation of ocean model ventilation with CFC-11: comparison of 13 global ocean models. Global Biogeochemical Cycles. (in press). Fiadeiro ME (1982) Three-dimensional modeling of tracers in the deep Pacific Ocean, II. Radiocarbon and circulation. Journal of Marine Research 40: 537--550. Key RM, Quay PD, Jones GA, et al. (1996) WOCE AMS radiocarbon I: Pacific Ocean results (P6, P16 and P17). Radiocarbon 38(3): 425--518. Libby WF (1955) Radiocarbon Dating. Chicago: University of Chicago Press. ¨ stlund HG and Rooth CGH (1990) The North Atlantic O tritium and radiocarbon transients 1972–1983. Journal of Geophysical Research 95(C11): 20147--20165. Schlosser P, Bo¨nisch G, and Kromer B (1995) Mid-1980s distribution of tritium, 3He, 14C and 39Ar in the Greenland/Norwegian Seas and the Nansen Basin of the Arctic Ocean. Progress in Oceanography 35: 1--28. Sonnerup RE, Quay PD, and Bullister JL (1999) Thermocline ventilation and oxygen utilization rates in the subtropical North Pacific based on CFC distributions during WOCE. Deep-Sea Research I 46: 777--805. Stuiver M and Polach HA (1977) Discussion: Reporting of 14 C data. Radiocarbon 19(3): 355--363. Stuiver M and Quay P (1983) Abyssal water carbon-14 distribution and the age of the World Ocean. Science 219: 849--851. Taylor RE, Long A, and Kra RS (eds.) (1992) Radiocarbon After Four Decades, An Interdisciplinary Perspective. New York: Springer. Toggweiler JR, Dixon K, and Bryan K (1989) Simulations of radiocarbon in a coarse-resolution World Ocean model. 1. Steady state prebomb distributions. Journal of Geophysical Research 94(C6): 8217--8242.


RARE EARTH ELEMENTS AND THEIR ISOTOPES IN THE OCEAN Y. Nozakiw, University of Tokyo, Tokyo, Japan Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2354–2366, & 2001, Elsevier Ltd.

Introduction The rare earth elements comprise fifteen lanthanide elements (atomic number, Z ¼ 57–71) as well as yttrium (Z ¼ 39) and scandium (Z ¼ 21), although promethium (Z ¼ 61) does not appear in nature due to its radioactive instability (Figure 1). They are an extremely coherent group in terms of chemical behavior and have recently been subjected to intense investigation in the field of marine geochemistry. All rare earth elements occur as a trivalent state with exception of Ce4þ and Eu2þ. In the lanthanide series, the progressive filling of electron in the shielded inner 4f shell with increasing atomic number leads to gradual decrease in the ionic radii from La3þ to

1

Lu3þ, which is called ‘the lanthanide contraction’. Consequently, small but systematic changes in chemical properties allow them to be used for unique tracers in studying fundamental processes that govern the cycling of rare earth elements in the ocean. For instance, the heavier lanthanides were predicted to be more strongly complexed in seawater and hence more resistant to removal by scavenging of particulate matter. Yttrium mimics the heavy lanthanides, particularly holmium, because of similarity in the ionic radii. However, Sc is a substantially smaller cation with a geochemical behavior that differs from other rare earths. In most literature, therefore, ‘rare earth elements (REEs)’ generally include only lanthanides and Y, and Sc is treated separately. Two elements, Ce and Eu can take the other oxidation states. Although Ce is generally well accommodated within the strictly trivalent lanthanide series in igneous rocks, oxidation reaction of Ce3þ to Ce4þ proceeds in oxygenated aqueous systems. In seawater, the resulting Ce4þ hydrolyzes readily and

2

H

3

Li

5

4

Be

11

12

19

20

Na K

37

Rb 55

Cs 87

Fr

Mg Ca

38

Ti

Sc Y

Ra A c

Cr

V

90

Th

59

Pr 91

Pa

Nd

Fe

Co 45

Ru Rh 77

76

Re Os 108

107

60

27

44

Tc

W

58

Ce

Mn

75

74

73

26

43

42

Nb Mo

Zr

57

89

25

24

41

40

39

Ba L a 88

23

22

Ir 109

29

28

Ni 46

Cu 47

Pd 78

Pt 110

Ag 79

Au

30

Zn 48

31

Si

Cd 80

32

Ga Ge 49

50

Hg

P 33

82

Se 52

Sb 83

F

Po

Ne 18

17

Cl 35

Ar 36

Br

Kr 54

53

Te 84

Bi

Pb

Tl

S 34

As 51

Sn

In 81

16

He

10

9

O

N 15

14

Al 21

Sr 56

C

B 13

8

7

6

I 85

At

Xe 86

Rn

111

61

92

Pm

U

93

Sm

Np

94

62

Pu

63

Eu

64

Am

96

95

Gd Cm

65

Tb 97

Bk

66

Dy 98

Cf

67

Ho 99

Es

68

Er

10

0

Fm

69

Tm 10

1

Md

70

Yb 10

2

No

71

Lu 10

3

Lr

Figure 1 Periodic table showing the position of rare earth elements.

w

Deceased.


653

654

RARE EARTH ELEMENTS AND THEIR ISOTOPES IN THE OCEAN

tends to be removed by scavenging. For this reason, seawater is typically depleted in Ce relative to that expected from neighboring La and Pr, whereas Ce is often enriched in some authigenic minerals such as manganese nodules and phosphorites. Europium can be reduced to Eu2þ, a larger cation that can be segregated from other REEs, during magmatic processes. Anomalous concentrations of Eu are not uncommon in various igneous and sedimentary rocks. However reduction of Eu does not normally take place within the ocean, although an Eu enrichment has often be encountered in hydrothermal fluids venting at midoceanic ridges. The anomalous behavior of Ce due to oxidation–reduction reactions can be best and quantitatively evaluated relative to the trivalent neighbors (La and Pr) in the lanthanides series without significant influence of the other processes affecting their oceanic distributions. This is a notable advantage of the element over the other transition metals, such as Mn and Fe, which behave individually affected by the oxidation states. In addition, there are two geochemically important isotopes of the lanthanides, 143Nd and 138Ce. The 143 Nd is produced by decay of 147Sm with a half-life of 10.6 1011 years. Natural variation of 143Nd/144Nd in terrestrial materials occurs depending on mantle/ crust segregation of Sm and Nd, and the age of the rocks. Thus, the 143Nd/144Nd ratio may be used to constrain the sources of the REEs and mixing within the ocean. Likewise, the 138Ce/142Ce ratio may also be used to constrain homogenization of the element by oceanic mixing since the 138Ce is produced by 138La decay (half-life, 2.97 1011 years). However, the 138 Ce/142Ce ratio has not been well exploited in marine geochemistry yet, because of its smaller natural variation as compared to that of Nd isotopes and analytical difficulty.

History and REE Normalization The analysis of picomolar REE concentrations in seawater has been difficult due to lack of sensitivity in the conventional methods. Earlier attempts to measure REEs relied almost entirely on high-sensitivity instrumental neutron activation analysis. In 1963 reliable REE concentrations were reported in a few waters from the Eastern Pacific as well as those in a manganese nodule and a phosphorite and their significance was recognized. Basic features of the REEs in seawater were found: a progressive increase across the lanthanide series from the light Pr to the heaviest Lu when the seawater concentrations were divided by those of sediments, and that Ce is markedly depleted in the seawater but enriched in the manganese nodule relative to neighboring La and Pr, as expected from its 4 þ valency state. Europium was normal relative to

other trivalent REEs in all the sample. It was also noted that the concentrations of the heavy REEs (Ho, Yb and Lu) in the Pacific deep water were considerably higher than those in the surface water. Although these earlier findings had to be refined and confirmed by subsequent workers with more precise modern techniques, the fundamental aspects of the REE marine geochemistry were developed for that time. Prior to 1980, reliable data on the distribution of REEs in seawater were few. Since the early 1980s, a growing number of reports on the subject have become available. Now, several laboratories in the world are capable of determining REEs in seawater with precision between one and a few percent by use of isotope dilution thermal ionization mass spectrometry (ID-TIMS) or inductively coupled plasma mass spectrometry (ICPMS). Therefore, more detailed arguments are possible regarding geochemical processes controlling the concentration, distribution, fractionation and anomalous behaviors in the oceans. When REE fractionation is discussed, it is common to normalize the data to the values in shale which are thought to be representative of the REEs in the upper continental crust. The shale-normalization not only helps to eliminate the well-known distinctive even– odd variation in natural abundance (the Oddo–Harkins effect) of REEs but also visualizes, to a first approximation, fractionation relative to the continental source. It should be noted, however, that different shale values in the literature have been employed for normalization, together with the ones of the PostArchean Australian Sedimentary rocks (PAAS) adopted here (Table 1). Thus, caution must be paid on the choice of the shale values if one ought to interpret small anomalies at the strictly trivalent lanthanides such as Gd and Tb. Alternatively, for detailed arguments concerning fractionation between different water masses in the ocean, it has been recommended that the data are normalized relative to the REE values of a distinctive reference water mass, for example, the North Pacific Deep Water (NPDW, Table 1). The NPDW-normalization eliminates the common features of seawater that appeared in the shale-normalized REE pattern and can single out fractionation relative to the REEs in the dissolved end product in the route of the global ocean circulation.

The Oceanic Distributions In 1982, the first ‘oceanographically consistent’ vertical profiles were reported for nine out of ten lanthanides that could be measured by the ID-TIMS method in the North Atlantic. Since then, a significant amount of data on the distribution of REEs have accumulated from various oceanic regions. For



Table 1 Element

Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Tb Lu a b

655

Ionic radii and the average REE concentrations in shale and seawater used for normalization Atomic number Z

39 57 58 59 60 62 63 64 65 66 67 68 69 70 71

˚ )a Ionic Radius (A CN ¼ 6

CN ¼ 8

0.900 1.032 1.010 0.990 0.983 0.958 0.947 0.938 0.923 0.912 0.901 0.890 0.880 0.868 0.861

1.019 1.160 1.143 1.126 1.109 1.079 1.066 1.053 1.040 1.027 1.015 1.004 0.994 0.985 0.977

PAAS a (mmol/kg)

NPDW b (pmol/kg)

304 275 568 62.7 235 36.9 7.1 29.6 4.9 28.8 6.0 17.0 2.4 16.3 2.5

236.3 38.7 3.98 5.10 23.8 4.51 1.24 6.83 1.13 8.38 2.34 7.94 1.23 8.37 1.46

After Taylor and McLennan (1985) for trivalent cations; CN, coordination number. Dissolved (o0.04 mm) REE in the North Pacific Deep Water at 25007100 m (after Alibo and Nozaki, 1999).

example, Figure 2 shows the station locations where the REEs were measured in seawater, together with the Nd concentrations in the surface water. Some of them are based on filtered samples, using generally a 0.4–1.0 mm membrane, of which results are called ‘dissolved’ concentrations. Many others, however, were measured on unfiltered and acidified seawaters, and hence can be ascribed to ‘acid-soluble total’ concentrations. The difference between the two is generally small, less than 5% for all trivalent REEs in the open oceans, even if the finer 0.04 mm membrane is used for filtration (Table 2), and gross features of their vertical profiles remain unchanged. Nevertheless, when fine structures of the REE patterns are discussed, filtration becomes critical in changing the pattern since the particulate fraction decreases from B5% at the light and middle to less than 1% at the heavy REEs. Furthermore, there is an obvious exception for Ce of which more than 35% is associated with particles, being consistent with its 4 þ oxidation state. Also, the REEs in unfiltered samples close to the bottom often show anomalously high concentrations due to resuspension of underlying sediments. The vertical profiles of dissolved (o0.04 mm) REEs for different oceanic basins are shown in Figure 3. Except for Ce, all REEs show ‘nutrient-like’ gradual increase with depth. There are small but systematic differences in the profiles across the series, although the North Atlantic profiles show somewhat complex features dominated by horizontal advection of different water masses (from the above, the Sargasso Sea Surface Water, the North Atlantic Deep Water, and the Antarctic Bottom Water). For instance, in the Southern Ocean and the North Pacific, the light and

middle REEs (e.g. Pr-Gd) almost linearly increase with depth, whereas the heavy REEs (e.g. Ho-Lu) and Y show convex features. The concentrations of the heavy REEs below 1500 m are in the order of North AtlanticoSouthern OceanoNorth Pacific much like those of nutrients, whereas those of the light REEs are Southern OceanoNorth AtlanticoNorth Pacific, presumably reflected by scavenging intensities in those regions. Although the vertical profiles of the REEs are similar to those of nutrients, e.g. dissolved silica, there is a difference in that the REE concentrations never approach zero in the surface water like nutrients in the temperate oligotrophic zone. Analysis of interelement correlations indicate that the heavy REEs and Y better correlate with dissolved silica and alkalinity than with nitrate and phosphate. Within the REEs, the best correlations (R240.99) can be found between neighboring trivalent REEs, and between Y and the heavy REEs. As a trivalent REE pair is apart in their atomic number, the correlation between the two becomes worse. Lanthanum often deviates in these general trends from the light REEs toward the heavy REEs. The vertical profiles of Ce is unique among the REEs showing a decrease from the high concentrations in the surface waters to the low and nearly constant values in the deep waters (Figure 3). Such a distribution pattern can be seen for other leastsoluble-elements, such as Al, Co, and Bi which also hydrolyze readily. The different profiles among the basins may be governed by the balance in the strength of external inputs (eolian þ riverine) to the surface ocean and particle scavenging throughout the water column.


656


90˚N 35 38 32

40 39

96-236 63

60˚N

7.9

30˚N

6.8 7.0

24

28

35 32

21 25 7.5 28 DBB-#86/1 9.8 5.1 9.3 13 14 14 7.9 9.8 14 CM-22 4.3 4.2 8.2 8_15 3.9 TPG-7/8 34 19 10 10 31 4.7 4.7 5.2 4.6 51 5.4 4.6 14 8.3 6.3 4.4 8.7 SS-#8 5.4 6.6 4.8 73 4.6 4.2 11 4.9 5.8 6.4 3.4 27 25 7.6 5.1 13 5.1 20 4.7

36

(See Fig. 2B)

32 13 10 11

0˚

8.3 9.7 8.1

6.9 6.9 7.7

8.2

30˚S

9.1 3.4 7.0

2.9 9.1

2.7

8.5

4.1 9.2

7.9

PA-4 8.7

60˚S

13

Nd (pmol/kg), 90˚S (A)

0˚

Unfiltered Filtered (0.04 mm), after Alibo and Nozaki (1999). Based on analysis on particulate matter collected on 0.4 mm membrane filters (Bertram and Elderfield, 1994). c Sum of analyses on successive chemical treatments on 0.4 mm-filtered particles (Sholkovitz et al. 1994). d Estimated value from neighboring trivalent REEs. b

0 Depth (m)

1000

La

Ce

Nd

Pr

2000 3000 Y

4000 5000

0 100 200 300 400

0

25

50

75 100

0

5

10

15

20

5

0

10

15

0

20

60

40

0 Depth (m)

1000

Eu

Sm

Dy

Tb

Gd

2000 3000 4000 5000 4

0

8

12

0

1

2

3

0

5

10

15

0

1

2

3

4

0

5

10

15

0 Ho

Depth (m)

1000

Er

Tm

Yb

Lu

2000 3000 4000 5000 0

1

2

3

4

5

0

5

10

15 0 0 5 1 2 3 _ Dissolved rare earth elements (pmol kg 1)

10

15

0

1

2

3

Figure 3 The vertical profiles of dissolved (o0.04 mm) REEs in the western North Pacific (CM-22; &), and the Southern Ocean (PA4; J) from Alibo and Nozaki, unpublished. The North Atlantic profile data based on dissolved (o0.4 mm) REEs (SS-#8; ~) from Sholkovitz and Schneider (1991) and acid-soluble total concentrations for monoisotopic REEs (DBB-#86/1; ’) from De Baar et al. (1983) and yttrium (TPG-7/8; m) from Alibo et al. (1999).



0.8

1.0 5m 100 m 200 m 400 m 600 m 1000 m 1200 m 1750 m 2250 m

0.8

Seawater/NPDW

0.6 (Seawater/PAAS) × 106

659

0.4

0.6

0.4

0.2 0.2

0

0 La

(A)

Y

Pm

Pr Ce

Nd

Tb Ho Tm Lu Eu Dy Sm Gd Er Yb

La (B)

Pm

Pr Ce

Nd

Tb Ho Tm Lu Eu Dy Sm Gd Er Yb

Figure 4 The PAAS-normalized (A) and the NPDW-normalized (B) patterns of dissolved REEs in the western North Pacific (Alibo and Nozaki, 1999).

than 0.1 at the oxic/aoxic boundary of B100 m depth to B1.0 (no anomaly) in reducing waters below 200 m (Figure 8). Even positive Ce anomaly can be found in the deep Cariaco Trench. Cerium reduction also takes place in coastal and hemipelagic sediment pore waters and can affect the distribution of Ce in the overlying waters.

Inputs of REEs to the Ocean The most obvious source of REEs in the ocean is riverine input. The concentrations of dissolved REEs in river waters are significantly higher than those of seawaters. The behaviors of the dissolved REEs in rivers and estuaries have been intensively studied as a link of REE geochemistry between the crust and the ocean. Shale-normalized REE patterns in river and estuarine waters often show nonflat patterns indicating that fractionation takes place with respect to continental crust. One of the prominent features derived from those works is that large-scale removal of dissolved REEs (in particular, light REEs) occurs in the estuarine mixing zone. Planktonic uptake, coprecipitation with iron hydroxides, and salt-induced coagulation of colloids have been suggested as the removal mechanism. This reduces the effective fluvial flux of dissolved REEs to the ocean considerably and affects the budget calculation in the ocean. For example, the river flux of dissolved Nd is

estimated to be 4.6 106 mol year1 (Table 3), being lowered by a factor of 3–4 in the estuaries, and the mean oceanic residence time with respect to the river input is about 104 years. This Nd mean residence time appears too long in the light of heterogeneous distribution of its isotopic composition observed in the different oceanic basins. The 143Nd/144Nd isotopic composition is generally expressed as, 2

143

eNd ¼ 4

Nd=144 Nd

3 measured

0:512638

15 104

where the value of 0.512638 is referred to the 143 Nd/144Nd ratio in the chondritic uniform reservoir (CHUR). Rivers entering into the Atlantic Ocean are influenced by old continental crust and have a relatively nonradiogenic eNd value of 12, whereas values for rivers entering into the Pacific which is surrounded by young oceanic islands and island arcs are more radiogenic, 3 to 4. Rivers draining into the Indian Ocean have intermediate values around 9. The Nd isotopic compositions in the riverine input are almost faithfully reflected in seawater. The measured eNd values for seawater also average 1272.5 for the Atlantic, 872 for the Indian Ocean, and 372 for the Pacific Ocean. Thus, the mean residence time of Nd must be short


660


Coatings (organic or oxide)

Ce (IV)

Ce (III)

Ce/Ce* 0.0

0.1

0.2

0.3

0.4

0.5

0 Detrital materials

1000

Oceanic Particles

.... Lu - nCO2_ Seawater 3 Complexes

Oceanic removal via Particle Settling

Figure 5 A conceptual model of REE fractionation between particles and seawater. Main features include (1) the systematic variation in the relative affinity of trivalent REEs for complexation to solution carbonates and binding to particles, (2) the enhanced formation of particulate Ce due to the oxidation of Ce(III) to Ce(IV), and presence of surface active coatings on detrital particles. These features lead to fractionation of REE between seawater and particles and to fractionation via the settling of large particles. After Sholkovitz et al. (1994).

1.6 1.4

South China Sea (4200 m)

Bay of Bengal (3600 m)

Seawater/NPDW

1.2 1.0

Sulu Sea (4950 m)

0.8 Andaman Sea (3780 m)

0.6 AABW (5500 m)

0.4

AAIW (990 m)

0.2 0.0 La

Ce

Pr

Nd

Ho Lu Pm Eu Tb Tm Dy Er Sm Gd Yb

Figure 6 The characteristic features of NPDW-normalized patterns of dissolved REEs in different oceanic basins. Data from Alibo and Nozaki (unpublished) and Nozaki et al. (1999).

compared with the oceanic mixing time of B103 years. This is in contrast to the Sr isotopes which show a constant 87Sr/86Sr ratio in seawater because of homogenization by oceanic mixing during its residence time of B106 years. Some additional sources of REEs appear to exist in the ocean to maintain geochemical consistency for Nd isotopes.

2000 Depth (m)

_ nCO23 - La .... Ce .... Nd .... Eu .... Er

3000

N.W. Pacific Southern Ocean N. Atlantic

4000

5000 Figure 7 The vertical profiles of Ce-anomaly calculated from the profile data in Figure 3.

The concentrations of REEs in hydrothermal fluids venting from hot springs at mid oceanic ridges have recently been investigated. They are generally 1–2 orders of magnitude higher than those of ambient seawater with a distinctly positive Eu anomaly, but are also intensively removed by scavenging in the vicinity of hydrothermal vent fields. Consequently, the effective hydrothermal flux of REEs to the ocean is negligibly small compared to the fluvial flux. Atmospheric input due to fallout of terrestrial aerosols and subsequent solubilization into seawater has been thought to be important, or even predominant in the open ocean for some reactive heavy metals such as iron and aluminum. However, the eolian fluxes are poorly quantified for the REEs as yet. Available estimates (Table 3), though highly uncertain, suggest that the eolian fluxes are 30– 130% of the fluvial input of REEs. Terrestrial aerosols are transported through the atmosphere for relatively long distances from the sources by longitudinally prevailing winds such as the westerlies and the trade wind. This is in contrast to the fluvial input which enters the ocean across the land–sea interface. Thus, the relative importance of these REE sources may be reflected in the geographical distribution of REEs in the surface waters. The distribution of dissolved Nd in the surface waters (Figure 2) shows the higher concentrations in regions of strong coastal influence, such as Baffin Bay, the Bay of Bengal, and the South China Sea. In the open ocean, there is a general tendency of Atlantic4Indian Ocean4Pacific for the surface dissolved Nd. The surface waters of



_

Ce/Ce*

REE concentration (pmol kg 1) 0

100

150

200

0.0

0

0

50

50

100

100

150

Sulfide interface

150

200

200

250

250

0

0 Depth (m)

Depth (m)

50

500

0.2

La Ce Nd

0.6

0.8

1.0

1.2

500

No anomaly

1500

2000

2000

2500

2500

(A)

0.4

Sulfide interface

1000

1000

1500

661

(B)

Figure 8 The vertical profiles of La, Ce, and Nd (A) and calculated Ce-anomaly (B) in the Black Sea. After German et al. (1991).

the different oceanic province have different NPDWnormalized REE patterns due to the input of REEs from local terrigneous sources (Figure 9). The Nd isotopic composition (Figure 10) that can serve as a ‘fingerprint’ of REE source also varies significantly with location. The geographical distributions of Nd concentration and its isotopic composition, and the NPDW-normalized REE pattern are hardly explained by the pattern of prevailing winds. All suggest that the REEs are supplied to the ocean from local sources and their mean residence times in the surface waters are relatively short as compared to homogenization by lateral mixing. The surface Nd distribution somewhat resembles that of fluvially and coastally derived 228Ra with a half-life of 5.7 years which is predominantly supplied to the coastal and shelf waters from underlying sediments. Since the direct riverine flux of dissolved Nd is not sufficient to yield its short mean residence time required for the oceanic Nd isotopic heterogeneity, some additional mechanism must be sought for the

coastal and shelf sources of the REEs. Remineralization of REEs from river-transported sediments may occur near the coast and on the shelf. For instance, if only 1–2% of detrital Nd is released to the coastal water, then the mean oceanic residence time of Nd is shortened to as less than 1000 years. In reality, there is some evidence that the REEs are released from sediments to the overlying water in the outer high-salinity regions of the Amazon and Fly River estuaries. Fractionation can also take place in the sediment diagenesis, and hence, makes the REE patterns of the surface waters different from those of river waters, depending upon the oceanic province. The highly radiogenic eNd values observed near the Indonesian Archipelago (Figure 10) also suggest that erosion of volcanic islands can be an important source of REEs to the ocean. Although quantification of these local sources of REEs is difficult at present, the coastal and shelf mechanism appears to be significant in the balance of dissolved REEs in the ocean (Table 3).


662


Table 3 Mean oceanic residence times derived from the particle reactivities and estimated remineralization fluxes of the REEs Element

Mean dissolved concentration (Cd)(pmol kg1)

Atmospheric flux a (106mol y1)

Riverine flux a (106mol y1)

Remineralization flux b (106mol y1)

Mean residence time (tREE)c (y)

Y La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

220 30 4.5 3.8 20 4.0 1.0 4.7 1.1 6.5 2.0 6.5 1.0 6.5 1.3

8.3d 4.1 8.5 1.1e 4.4 0.91 0.20 0.85 0.13e 0.72 0.15e 0.45 0.055e 0.32 0.058

15d 5.0 6.3 1.2e 4.6 1.1 0.3 1.4 0.22e 1.2 0.27e 0.8 0.14e 1.1 0.2

155 72 107 19 59 11 2.8 10 2.3 9.9 1.1 2.4 0.36 2.2 0.35

1670 500 50 240 400 400 410 520 570 740 1820 2420 2430 2440 2890

a

After Greaves and Elderfield (1994). The values may be uncertain by a factor of 2–3. Remineralization flux ¼ (1.35 109 Cd)/tREE Atmospheric flux Riverine flux. c Based on tREE ¼ tp/Fp for tp ¼ 10 years, g ¼ 0.870.2 and Fp values given in Table 2. The values may be uncertain by a factor of 2–3. d Assumed to be 55 times of Ho flux from the crustal ratio. e Estimated from neighboring trivalent REEs on the basis of shale (PAAS) composition. b

Particle Reactivity and Mean Oceanic important role in this respect. For instance, the REE concentrations of shells are extremely low. Also, it is Residence Times of REEs The major process that removes dissolved REEs from the water column is considered to be adsorptive scavenging by sinking particulate matter (Figure 5) which controls the distributions of not only REEs but also a number of reactive elements in the ocean. Some elements can also be removed by lateral transport along with currents and subsequent intensified particle scavenging and/or uptake at the sediment–water interface of the oceanic margins. This, ‘boundary scavenging’ is known to occur for 210Pb and 231Pa, but this mechanism is probably insignificant for dissolved REEs. The vertical profiles of dissolved REEs (Figure 3) strongly suggest that they are involved in the oceanic biogeochemical cycle. Marine particulate matter comprises biogenic organic matter, opaline silica, and carbonates as well as inorganic oxides and terrestrial detritus. The biogenic particles are produced in the surface water and transported by settling to the deep water and the seafloor where they are largely remineralized to the water. Despite a rough similarity in the vertical profile between the REEs and dissolved Si or alkalinity, however, active processes such as biological uptake and incorporation of REEs into calcareous and opaline skeletons do not seem to play any

unlikely that the dissolved REEs are actively taken up into organic tissues. Adsorptive scavenging of reactive metals has been clearly demonstrated, since some short-lived radionuclides such as 234Th, 210Po and 210Pb are highly enriched in particulate matter collected by filtration and sediment traps. The scavenging processes of reactive metals may be governed by competitive reactions between binding affinity onto the particle surfaces and complex formation with ligands in seawater (Figure 5). Although marine particles are composed of various materials, their surfaces are thought to be coated with organic matter and oxides of manganese and iron. The fine suspended particles, that can be conventionally collected by filtration, predominate in the surface area onto which reactive metals are adsorbed and the standing stock of mass. In contrast, the large, fast sinking particles, that can be effectively collected by a sediment trap, virtually govern the vertical flux of materials. Those particles however, are frequently exchanged through aggregation and disaggregation during sinking through the water column. Since the vertical transport of particulate matter to the sedimentary sink is a process common to all associated elements, the overall rate of removal from the ocean for metals such as the



0.8

1.2

0.4 Philippine Sea

1.0

West Australian Coast

Indonesian Seas 0.6

Seawater/′NPDW′

0.3 0.8

0.4

0.6

0.2

0.4 SSW-1 SSW-2 SSW-3 SSW-4

0.1

La Pr Pm Eu Tb Ho Tm Lu Ce Nd Sm Gd Dy Er Yb

(A)

PA-1 SSW-6 SSW-7 SSW-8 SSW-9

0.2

0.2

(B) 0.8

0.4 Southern Ocean


(C)

0.6

Seawater/′NPDW′

Bay of Bengal + Andaman Sea

0.8 0.4

0.2

0.1

SSW-14 SSW-15 PA-4 La

(D)

0.6

PA-5 SSW-17 SSW-18 PA-7 SSW-19

0.2

0.0

0.0 Ce

Pr

Pm Eu Tb Ho Tm Lu Nd Sm Gd Dy Er Yb

2.0

(E)

0.4

(F)

0.8

1.2

0.6

0.8

0.4


0.4 South China Sea

1.6

SSW-21 PA-9 PA-10 SSW-22 SSW-23

0.2 0.0


1.0 Strait of Malacca

Seawater/′NPDW′


1.2

Eastern Indian Ocean

1.0 0.3

SSW-10 SSW-11 SSW-12 SSW-13

0.0

0.0

0.0

0.3

East China Sea + Kuroshio near Japan

0.2

0.4

SSW-24 SSW-25 SSW-26

0.0

(G)

663

SSW-28 SSW-29 SSW-30 SSW-31 SSW-32 PA-11 SSW-33

0.2

0.0 La Pr Pm Eu Tb Ho Tm Lu Ce Nd Sm Gd Dy Er Yb

(H)


SSW-36 SSW-37 SSW-38 SSW-39 SSW-40

0.1

0.0

(I)

La Ce

Pr Pm Eu Tb Ho Tm Lu Nd Sm Gd Dy Er Yb

Figure 9 Comparison of NPDW-normalized REE patterns in the surface waters in the eastern Indian Ocean and its adjacent seas. See Figure 2(B) for the locations. Adapted from Amakawa et al. (2000).

REEs must depend on their partitioning between dissolved and particulate phases and a mean sinking velocity of the particles. Available data on the REEs associated with suspended particles are summarized in Table 2. With the assumptions that a large portion of particulate REEs are exchangeable with dissolved REEs and that the suspended fine particles are incorporated into the large, fast-sinking particles through repeated aggregation and disaggregation and eventually removed from the ocean with a certain mean residence time (tp), the mean residence time of REEs (tREE) is given by the following equation. tREE ¼ tp =gFp where Fp is the fraction of particulate REEs in seawater (Table 2) and g is the portion that is effectively exchangeable with the dissolved REEs. Based on the

studies of Th isotopic systematics in the water column, the fine particles, on which most reactive metals reside, settle with a mean speed of B1 m day1 corresponding to the mean oceanic residence time of about 10 years. From the chemical leaching experiments of particulate matter and acidification treatment of both filtered and unfiltered seawaters to pHB1.5 to determine acid-soluble particulate fraction of REEs, the value of g is deduced to be between 0.6 and 1.0. The mean residence times of REEs estimated using these values are given in Table 3. They range from 50 years for Ce to 2900 years for Lu. The short mean residence time of Ce is approximately equal to that of Th, being consistent with their predicted dominance of hydrolysis species in seawater. The 400 year residence time is well suited for Nd with its heterogeneous distribution of the isotopic ratios in the oceans.


664


90°N

_ 10.0

60°N

_ 17.7 _ 23.4 _ 9.1 _ 17.0 _ 23.1 _ 17.1 _ 17.9 _ 26.1 _ 15.4

_ 19.8 _ 14.5 _ 0.1

_ 9.7 _ 9.3

30°N

_ 5.2

_ 11.2

0°

_ 6.0 _ 7.0

30°S

_ 11.4 _ 10.6

_ 9.9 _ 9.7

_ 9.5

_ 3.2

_ 11.8 _ 12.0

_ 4.8

_ 10.4

_ 10.2 _ 1.5 _ 10.9 _ 4.1 _ 4.0 _ 3.9 _ _ 4.1 4.6 _ 5.6 _ 5.5 _ 7.0 _ 7.5 _ 10.4 _ 11.4 _ 9.9

_ 11.2 _ 10.5

_ 9.9

_ 9.6

_ 6.4 _ _ 9.58.7

_ 10.3

_ _ 11.4 10.7 _ 12.2

+0.3

_ 8.7

60°S

Nd value in the surface water

90°S 0°

60°E

120°E

180°

120°W

60°W

0°

Figure 10 The distribution of 143Nd/144Nd isotopic composition (expressed as eNd value) in the surface waters. Compiled from Byrne and Sholkovitz (1996) and Amakawa et al. (2000).

It is also clear that eolian and riverine sources alone are insufficient to yield such short residence times and there must be additional sources of REEs in the ocean. The most likely candidate of the potential REE source is remineralization of nearshore, coastal and shelf sediments as described earlier. The magnitude of this remineralization flux required to balance each REE in the ocean is given in Table 3. Those fluxes are quite large particularly for the light REEs, whereas the relative importance of remineralization decreases and becomes somewhat comparable with the sum of eolian and riverine inputs for the heavy REEs. This REE fractionation during remineralization on the shelf may contribute the different NPDW-normalized REE patterns observed in the surface waters (Figure 9), although the mechanism is not well understood.

of other heavy metals, is best understood by the competitive complexation reactions of dissolved REEs with ligands on the surface of particles and in solution. This yields the progressive enrichment from the light to the heavy REEs in the shale-normalized pattern as commonly observed in seawater. Remineralization of REEs from coastal and shelf sediments is thought to play a significant role in the global budget of REEs and the geochemical and isotopic consistency of Nd in the ocean. The REEs and Nd isotopes provide novel and unique oceanographic tracers or chemical problems in studying (1) particle scavenging processes, (2) redox sensitive geochemical processes with Ce, and (3) identification and modification of water masses. Applications to the wider problems are being explored.

See also

Summary and Conclusion Although the oceanic distributions of rare earth elements somewhat resemble those of nutrients, their behaviors are clearly different in that they are not actively taken up by photoplankton but passively scavenged by particles. Elemental reactivity with suspended particles, which controls, together with sinking of particle aggregates, the mean oceanic residence times of not only REEs but also a number

Estuarine Circulation. River Inputs. Tracers of Ocean Productivity. Water Types and Water Masses.

Further Reading Alibo DS and Nozaki Y (1999) Rare earth elements in seawater: particle association, shale-normalization, and



Ce oxidation. Geochimica et Cosmochimica Acta 63: 363--372. Alibo DS, Nozaki Y, and Jeandel C (1999) Indium and yttrium in North Atlantic and Mediterranean waters: comparison to the Pacific data. Geochimica et Cosmochimica Acta 63: 1991--1999. Bertram C and Elderfield H (1993) The geochemical balance of the rare earth elements and neodymium isotopes in the oceans. Geochimica et Cosmochimica Acta 57: 1957--1986. Byrne RH and Sholkovitz ER (1996) Marine chemistry and geochemistry of the lanthanides. In: Gschneider KA Jr and Eyring L (eds.) Handbook on the Physics and Chemistry of Rare Earths, vol. 23, pp. 497--593. Amsterdam: Elsevier Science. DeBaar HJW, Bacon MP, and Brewer PG (1983) Rareearth distributions with a positive Ce anomaly in the western North Atlantic Ocean. Nature 301: 324--327. Elderfield H and Greaves MJ (1982) The rare earth elements in seawater. Nature 296: 214--219. German CR, Holliday BP, and Elderfield H (1991) Redox cycling of rare earth elements in suboxic zone of the Black Sea. Geochimica et Cosmochimica Acta 55: 3553--3558. Goldberg ED, Koide M, Schmitt RA, and Smith RH (1963) Rare-earth distributions in the marine environment. Journal of Geophysical Research 68: 4209--4217.

665

Greaves MJ, Statham PJ, and Elderfield H (1994) Rare earth element mobilization from marine atmospheric dust into seawater. Marine Chemistry 46: 255--260. Lipin BR and McKay GH (eds.) (1989) Geochemistry and mineralogy of rare earth elements. Reviews in Mineralogy 21. Washington, DC: Mineralogical Society of America. Nozaki Y (1997) A fresh look at elemental distribution in the North Pacific Ocean. EOS Transactions 78: 221. Nozaki Y, Alibo DS, Amakawa H, Gamo T, and Hasumoto H (1999) Dissolved rare earth elements and hydrography in the Sulu Sea. Geochimica et Cosmochimica Acta 63: 2171--2181. Sholkovitz ER, Landing WM, and Lewis BL (1994) Ocean particle chemistry: The fractionation of rare earth elements between suspended particles and seawater. Geochimica et Cosmochimica Acta 58: 1567--1579. Sholkovitz ER and Schneider DL (1992) Cerium redox cycles and rare earth elements in the Sargasso Sea. Geochimica et Cosmochimica Acta 55: 2737--2743. Taylor SR and McLennan SM (1985) The Continental Crust: Its Composition and Evolution. Oxford: Blackwell.


RED SEA CIRCULATION D. Quadfasel, Niels Bohr Institute, Copenhagen, Denmark Copyright & 2001 Elsevier Ltd. This article is reproduced from the 1st edition of Encyclopedia of Ocean Sciences, volume 4, pp 2366–2376, & 2001, Elsevier Ltd.

Introduction

Gulf of Aqaba

N

25°

20° 200

666

Gulf of Suez

30°

1000

The Red Sea is a fiord type marginal sea in the northwest of the Indian Ocean, exposed to the arid climate of northern Africa and Arabia. Strong evaporation leads to a buoyancy-driven overturning circulation and to the production of a dense and highly saline water mass, the Red Sea Water. Near the surface the circulation is substantially modified through the forcing by the monsoon winds. Red Sea water spills at a rate of 0.37 106 m3 s1 through the Strait of Bab El Mandeb into the Gulf of Aden and spreads at intermediate depths in the Indian Ocean, where it can be traced as far south as the southern tip of Africa. The circulation of the Red Sea appears not to be controlled by the hydraulics in the strait. This article summarizes the circulation internal to the Red Sea, the formation and transformation of the deep water masses and the exchange of water between the Red Sea and the Gulf of Aden. The Red Sea was created as part of the spreading rift system between the African and the Arabian plates. It stretches between 121 and 201N from tropical to subtropical latitudes (Figure 1), is about 2000 km long, on average 220 km wide and reaches depths of up to 2900 m, but its mean depth is only 560 m. The only connection to the oceans is in the south to the Gulf of Aden–Indian Ocean via the narrow Strait of Bab El Mandeb. The shallow Hanish sill of 163 m depth in the north of the strait limits the free communication to the upper part of the water column. Northern appendices to the Red Sea are the Gulf of Suez, a shallow shelf basin, and the 1800 m deep Gulf of Aqaba, a smaller version of the Red Sea itself, bounded by a 260 m deep sill in the Strait of Tiran. Both regions are important for the production of the deep water of the Red Sea. The circulation of the Red Sea is driven by local atmospheric forces, through buoyancy and momentum fluxes. The net evaporative freshwater flux in the Red Sea is among the highest in the world, reaching 2 m year1 with some seasonal variability associated with the monsoon wind system. During summer, values are highest in the north, but during

winter in the south. Although the strong evaporation leads to sea surface salinities of up to 42 PSU (practical salinity units) in the Gulf of Suez, the buoyancy flux associated with the latent and sensible heat losses is an order of magnitude larger than that caused by the freshwater flux. In the north, where convection and water mass formation take place, the total buoyancy flux amounts to more than 20 108 m2 s3. Here surface temperatures may drop below 201C in winter. The surface wind field over the Red Sea is guided by the high mountains and plateaux on either side of the basin, and its dominant component is in the along-axis direction. In the southern part winds reverse twice per year. During the south-west monsoon from June to September they blow towards south south-east, whereas during the north-east monsoon and the transition times between the monsoon periods (October to May) the airflow is directed towards north north-west. North of about 201N the wind is blowing towards south south-east throughout the year, being strongest during summer and weakest during winter. Laterally the mean wind stress

Hanish Sill

15° Strait of Bab EI Mandeb

Perim Narrows

12° 32°

35°

40°

E 44°

Figure 1 Bathymetry of the Red Sea with 200 m and 1000 m depth contours. Geographical names mentioned in the text are indicated.


RED SEA CIRCULATION

decreases from the central axis towards the boundaries on either side, due to lateral friction effects. The geographical setting of the Red Sea with a length to width ratio of about ten has frequently led to the use of a two-dimensional channel approach to model the circulation. Also most of the observational data have been collected along the main axis of the Red Sea, limiting the interpretation of physical features to the two channel dimensions. The more recent three-dimensional surveys and modeling studies have, however, shown that at least in the upper layers significant cross-axis flows exist. Mesoscale gyres and eddies are an important part of the Red Sea circulation.

Hydrographic Structure In terms of the classical hydrographic parameters the Red Sea can be viewed as a two-layer system consisting of a deep layer below sill depth, which is very homogeneous in temperature and salinity, and an upper layer, comprising the surface mixed layer and the strongly stratified thermocline. This upper layer is about 200–300 m deep. It is composed of warm and relatively fresh surface water that enters from the Indian Ocean and is then modified due to air–sea interaction. Evaporation increases the salinity, from 36 PSU in the Strait of Bab El Mandeb to above 40 PSU in the north, and cools the surface water, although locally some heating may also take place due to radiative heat gain. Convection and vertical mixing renews the subsurface water which then returns southward in the lower part of the thermocline. The temperature and salinity distribution of the upper layer changes substantially between the two monsoon seasons, due to the varying atmospheric forcing (Figure 2). In winter the vertical temperature stratification is stable throughout the Red Sea. Highest surface temperatures are found around 181N, in the zone of wind convergence and thus low wind speeds. Toward the north, surface temperatures decrease to less than 201C in the Gulf of Suez. In the thermocline below the mixed layer, temperatures decrease to the deep water value. This pattern is consistent with the above described overturning circulation, typical for concentration basins, with an inflow at the surface and an outflow of colder and more saline, and thus denser water in the lower thermocline. It is this water that constitutes the major part of the outflow at the bottom of the strait of Bab El Mandeb. In contrast during the summer the temperature stratification in the southern part of the Red Sea is unstable. A cold wedge penetrates below the shallow mixed layer from the Gulf of Aden into the Red Sea and outflows of warmer water exist

667

above and below. This cold wedge has been observed as far north as 201N. This three-layer structure in the southern Red Sea is caused by the northerly monsoon winds, which during the summer season between June and August drive the surface water out of the Red Sea. The full depth hydrographic section shown in Figure 3 was taken during October 1982. In the upper layer the structure is still that of the south-west monsoon, with an intrusion of cold and low salinity water at intermediate depths. Surface temperatures are highest just north of Bab El Mandeb (B321C) decreasing to 271C in the north, and surface salinities increase from 37.5 PSU in the strait to more than 40 PSU in the north. Compared to these changes in the upper layer the deep water temperatures and salinities are almost homogeneous with values of 21.61C for temperature and 40.6 PSU for salinity. Horizontal differences between the northern and the southern part of the Red Sea do not exceed 0.2 K and 0.02 PSU, respectively. In contrast, the distribution of dissolved oxygen (and other chemical parameters) does show some structure in the deep water. There is a pronounced oxygen minimum around 300 m depth in the southern part of the Red Sea, whereas values are relatively high near the bottom and in the northern part. The distributions of chemical parameters have been used to evaluate the pattern and magnitude of the deep circulation. New deep water is formed in the north, it sinks to the bottom, spreads southward and upwells slowly. The intermediate depth oxygen minimum is thus a result of the balance between upwelling of Red Sea Deep Water and downward diffusion of thermocline water.

Circulation in the Upper Layer The annual mean circulation above the pycnocline is largely driven by thermohaline forces. The excess of evaporation over precipitation results in a loss of 0. 03 106 m3 s1 and forces a near surface inflow through the Strait of Bab El Mandeb. Applying Knudsen’s formula and balancing mass and salt fluxes through the strait and the sea surface results in an overall strength of the circulation and thus an outflow into the Gulf of Aden of 0.33 106 m3 s1. Direct measurements of the exchanges in the strait have confirmed this number. These mean conditions have been successfully modeled as a turbulent convective flow driven by a constant buoyancy flux B0 at the sea surface. The surface currents scale with ðB20 xÞ1=3 where x is the distance from the inner boundary of the estuary, and the overturning circulation is confined to the layer above sill depth. Superimposed on this pattern is the


668

RED SEA CIRCULATION 7−10 Aug 1985 25°N UB 06 0

20°

> 30

15° 32

31

32

30

m

28

18

25 24

16 23

100

21

22.5

22

200

300 500

(A)

16 _18 Dec 1985 25°N UB 10

20°

0 m

24 25

km

1000

2000

15°

27

25

26

25

27

28 28

Encyclopedia of Ocean Sciences,

Encyclopedia of Ocean Sciences, Second Edition, Volume 3: G-M

Encyclopedia of Ocean Sciences, Second Edition, Volume 1-6